SERVICING TELEVISION VOLUME 2 G. N. PATCHETT LONDON: NORMAN PRICE (PUBLISHERS) LTD. The Cathode Ray Tube. Sawtooth Current Generators

Transcription

1 m 3 TELEVISION SERVICING VOLUME 2 The Cathode Ray Tube Synchronizing Separators Timebases Field Output Stage Line Output Stage Sawtooth Current Generators G. N. PATCHETT B.Sc. (Eng.)., Ph.D., C. Eng., F.I.E.E., F.I.E.R.E., M.I.E.E.E. Fellow of the Royal Television Society The theory of television and the servicing of receivers (monochrome) are covered by this series of four volumes. The series is concerned mainly with modern receivers but also deals with television systems and studio equipment as used in this country. This new edition of the second volume has been enlarged to include information on the expanding use of transistors in receivers. Students of television servicing and those requiring a general knowledge of television will find this a valuable series. LONDON: NORMAN PRICE (PUBLISHERS) LTD 15s (75p) net

2 TELEVISION % SERVICING Volume 2 G. N. PATCHETT B.Sc. (Eng.), Ph.D., CEng., F.I.E.E., F.I.E.R.E., M.I.E.E.E. Fellow of the Royal Television Society LONDON NORMAN PRICE (PUBLISHERS) LTD

3 NORMAN PRICE (PUBLISHERS) LTD 17 TOTTENHAM COURT ROAD, LONDON, W.l NORMAN PRICE PUBLISHERS LTD., th edition nd impression 1971 Printed in Great Britain by A. Brown & Sons Ltd., Hull

4 CONTENTS 1 THE CATHODE RAY TUBE page 1 2 SYNCHRONIZING SEPARATORS 18 3 TIMEBASES 46 4 FIELD OUTPUT STAGE 65 5 LINE OUTPUT STAGE 75 6 SAWTOOTH CURRENT GENERATORS 117 APPENDIX AVOIDING SWITCHING-OFF BURNS page 120 INDEX 122

5

6 CHAPTER 1 THE CATHODE RAY TUBE In Volume 1 we considered the r.f./i.f. section of the receiver, the vision and sound demodulators, video amplifier and sound section. At this point it is appropriate to deal with the cathode ray tube. Later we shall deal with synchronizing separators, timebases and deflecting circuits, which are all closely allied to the cathode ray tube. We can divide the cathode ray tube into three sections: (1) The beam-forming section or electron gun. (2) The beam-deflecting section. (3) The viewing or screen section. (1) THE BEAM-FORMING SECTION OR ELECTRON GUN Before we can produce an electron beam we need a source of electrons. In a cathode ray tube the electrons are obtained from a heated cathode, essentially similar in principle to that used in all thermionic valves. In a valve we need a large number of electrons and they are liberated in all directions from a cylindrical cathode. In a cathode ray tube we do not need as many electrons and, ideally, these should all come from a point source. In practice this means that we must make the source of electrons as small as possible, consistent with obtaining the required emission current. The cathode is usually constructed in one of two ways shown in figure 1.1. At (a) the cathode OXIDE COATING pi ^ /COATING OXIDE HEATER CATHODE (2) Q FIG CONSTRUCTION OF THE CATHODE OF A CATHODE RAY TUBE consists of a small tube about &' in diameter, with a heater down the centre. Only the end of the tube is oxide coated. At (b) the cathode consists of a short length of flat tube say i" to &* wide and i" to -fe' long, with a heater in the centre. A small portion of one side is oxide coated. The method of manufacture and the materials used are similar to those for the cathode of a thermionic valve. The next thing that we must do is to accelerate the electrons to a high velocity and confine them to a relatively narrow beam. The electrons are accelerated to the required velocity by an anode, which is maintained at a high positive potential relative to the cathode. It can be shown that the velocity is given by the expression: Velocity v / 2 - ^ metres/second where e/m is the ratio of the electric charge to the mass of the electron and is equal to x 10" coulomb/kg; and Kis the accelerating voltage in volts. When the accelerating voltage is lokv the velocity is enormous approximately 60,000,000 metres/second or 110,000,000 miles per hour. The simplest 1

7 2 TELEVISION SERVICING way of producing an approximate beam is shown in figure 1.2 where the electrons are attracted towards an anode having a hole in it. Some of the electrons will pass through the hole and form an electron beam, but the *;- CATHODE ELECTRONS ANODE (Po.,v<) FIG 1.2. PRINCIPLE OF ELECTRON GUN majority will be collected by the anode. The beam can be made smaller only by reducing the size of the hole which decreases the number of electrons in the beam, the efficiency then being extremely poor. In order to improve the efficiency we want to converge the electrons on to the hole, and this may be done by the use of a negative cylinder around the cathode, as in figure 1.3. zm^ CATHODE / ^Negative) ELEC ANODE (Positive) FIG 1.3. PRODUCTION OF ELECTRON BEAM USING A WEHNELT CYLINDER The cylinder was originally called a Wehnelt cylinder but, in the form that it is used in the modern cathode ray tube, it is called the grid. By making the cylinder negative, with respect to the cathode, the electrons are repelled from it and converge on the hole in the anode. Although a narrow beam may be formed at the anode in this way the beam would not remain narrow up to the screen, owing to the mutual repulsion of the electrons. The force tending to repel the electrons from each other (since they have the same polarity of charge) causes the electrons to move sideways as they travel towards the screen and, as a result, the spot on the screen would be large and quite useless. To prevent this we must use some form of focusing. This is rather similar to the focusing of a light beam by a lens. One may consider the source of electrons as being like the filament of a lamp and, if we wish to obtain a narrow beam of light as in a spotlamp, we use a lens in front of the lamp. The light from the filament diverges outwards as it leaves it and the purpose of the lens is to collect the rays of light and form them into a parallel or converging beam. We wish to do a similar thing with the beam of electrons. Instead of allowing them to spread out we want them to converge so that they will strike a single spot on the screen. To do this we use an electron lens, which may take the form of a magnetic lens produced by a magnetic field or an electrostatic lens produced by suitable electrostatic fields. The exact operation of these lenses is extremely complicated and only a brief explanation can be given here. Magnetic lenses A beam of electrons is equivalent to the flow of an electric current (flowing in the opposite direction to that of the electrons). Thus, if a beam of electrons travels through a magnetic field a force will be exerted on it in a similar manner to the force exerted on a conductor carrying

8 THE CATHODE RAY TUBE 3 a current in a magnetic field, e.g. in a motor or a loudspeaker. Due to this force, an electron beam entering a magnetic field is bent in a direction at right angles to the magnetic field. This is shown in figure 1.4, the path of the X MACNETIC FIELO (into paper) FIG 1.4. FORCE ON ELECTRONS (alwayt at riant Onal«i to motion) DEFLECTION OF AN ELECTRON BEAM DUE TO MAGNETIC FIELD electrons during their travel in the magnetic field (which is assumed uniform) being a circle. When an electron leaves the magnetic field it will travel in a straight line, along the direction in which it left the field. This principle is made use of for the focusing of the electron beam. A simple electromagnetic lens is shown in figure 1.5. This consists of a NECK OF CATHODE RAY TUBE FIG 1.5. ELECTROMAGNETIC FOCUSING UNIT coil C around which is an iron shroud 7, completely surrounding the coil apart from an annular gap G. When a current is passed through the coil a magnetic field is set up which passes through the iron portions I and across the gap G. The field across the gap G spreads out into the neck of the tube and into the space occupied by the electron beam. The effect of the magnetic field is to converge the beam, as shown, rather similar to a convex lens used in light. The action is rather complex as the field is not uniform and the path of the electron is not as simple as shown, but the electrons travel in a helical path as well as being converged. Thus, the beam is actually twisted as it passes through the lens but, since the cathode is symmetrical, this is of no consequence and we can consider the lens just converging the beam, so that all the electrons will arrive at a single point on the screen. The strength of the magnetic field can be varied by varying the current in the coil and this varies the amount of convergence and hence the focus. This is equivalent to varying the focal length of a lens. At one time this type of magnetic focus unit was in general use, but it was later replaced by a permanent magnet type. There are two general types of permanent magnet focus units. One is basically the same as that shown in figure 1.5, except that the coil is now replaced by a permanent magnet. The arrangement is shown in figure 1.6. The permanent magnet is in the form of a ring P. At the sides are attached iron discs D, and Z> 2 and in the centre is a sleeve S which slides in the centre hole of D 2. As before, a magnetic field is produced across the gap G and extends into the

9 TELEVISION SERVICING DISC- CAP - c ^; PERMANENT MAGNET, p ywmn ^ryri( -DISC, 2 ~SLEEVE, s FIG 1.6. PERMANENT MAGNET FOCUSING UNIT neck of the cathode ray tube. As the strength of the permanent magnet cannot be varied directly for focusing, it is so arranged that the gap G can be varied thus varying the amount the fringing field of gap G extends into the tube. The gap is varied by sliding the sleeve S in and out. A mechanical device (not shown) is arranged to slide the sleeve S when a lever is moved which is attached to the disc D 2. The range of adjustment is rather limited but it is sufficient for variations likely to be required for any particular type of tube. Later the arrangement shown in figure 1.7 became common. This FIG PERMANENT MAGNET FOCUSING UNIT type makes use of two Ferroxdure rings (sintered oxides of iron and barium), which are magnetized as shown and arranged to oppose each other. The effective strength of the focusing system is now conveniently varied by varying the spacing between the two rings. A mechanical arrangement is designed so that by movement of the focusing lever this spacing is varied sufficiently to cover the normal variations required in practice. Magnetic focusing was used almost exclusively for television cathode ray tubes, but now electrostatically focused tubes are universally used. Let us now return to the electrode system of the magnetically focused tube. The simplest type of tube consists of a cathode with heater, a grid and anode and is called a triode tube, since it has the same number of electrodes as the normal triode valve. A usual arrangement is shown in figure 1.8. The grid which is similar to the Wehnelt cylinder already described, is now a cylinder with a disc at one end having a small hole in it, through which the electrons pass. The anode commonly consists of a tube, often with a disc (in which there is a small hole) inside it. The anode is also continued as a conducting coating (of graphite) on the inside of the neck and part of the way up the

10 THE CATHODE RAY TUBE m 75 rrtt VT '1 E \ *ANODE ' ' CAWODE \ 'ANODE GRID *fhite COATINC ^POSITION OF MAGNETIC LENS FIG TRIODE ELECTRON GUN flared portion of the tube. Connection to this anode (which operates at a high voltage of lokv and upwards) and coating is made about half way up the flared portion of the tube by means of a connector. The action of the grid is to concentrate the electrons towards a point P which is known as the crossover. The way in which the crossover is formed is very complicated. If the cathode were focused on to the screen of the tube we should obtain a large spot; but by focusing the cross over point (which has a much smaller cross-section, say 100th of that of the cathode), a satisfactory small spot may be obtained. By varying the voltage on the grid we can vary the number of without making electrons in the beam, and so the brilliance of the screen, appreciable changes in the focus. The purpose of the disc D with a hole in it is to prevent any stray electrons, which are widely diverging, reaching the screen. It acts somewhat like the stop in a camera lens. As well as the triode tube a tetrode type of cathode ray tube was used. This has another anode placed between the grid and the final anode and forms an electrostatic lens (see later) although the tube is still mainly focused by the magnetic lens. The advantage of the tetrode is that the width of the electron beam can be reduced, for a given beam current, which tends to improve the focus, particularly with permanent magnet focusing units. The narrow beam also reduces the defocusing which often occurs when the beam The potential required on this first anode is deflected by the deflecting coils. is low, usually V, and the current is negligible. In some cases two additional anodes were used compared with a triode, the first being maintained at about V and the second at zero voltage. Electrostatic Focusing When an electron enters an electrostatic field the electron tends to travel in the direction of the electrostatic lines (or, to be more correct, in the opposite direction since, by definition, an electrostatic line is the direction in which a positive charge will move). Suppose that we have an electrostatic field between two surfaces A and B, as shown in figure 1.9, which could be produced by applying a voltage between two fine + 4) A B FIG DEFLECTION OF ELECTRON BEAM DUE TO ELECTROSTATIC FIELD

11 6 TELEVISION SERVICING wire meshes. If an electron travels in the direction as shown at (a), then the electrostatic field will accelerate it to a higher velocity, but will not alter its direction of travel as it was initially travelling in the direction of the electrostatic lines. On the other hand, if an electron enters the electrostatic field at an angle as shown at (b), the force acting in the direction A to B will accelerate it in this direction and, therefore, alter its direction of travel. When it leaves the electrostatic field it will again travel in a straight line, not in the same direction that it entered the field, but in a direction more nearly that of the electrostatic lines. This principle is made use of in an electrostatic lens. A simple electrostatic lens consisting of two cylinders is shown in figure Here, due to the shape of the electrostatic lines between the two cylinders, ELECTROSTATIC FIELD ELECTRON BEAM FIRST ANODE + IOOV SECOND ANODE + SOOV FIG SIMPLE ELECTROSTATIC LENS the electrons are converged as shown and hence they meet at a single point on the screen. The amount of convergence, or focal length of the lens, is varied by varying the potential between the two cylinders, or anodes as they are called. The shape of the two anodes may vary considerably but the principle remains the same. Another arrangement is shown in figure 1.11 which uses three anodes. + IOOOV + IOOOV -KM I + 200V I ELECTROSTATIC FIELD GRID t FIRST ANODE SECOND I. ANODE FIG THREE-ANODE ELECTROSTATIC LENS The potentials shown are typical of those which might be used on a cathode ray tube in an oscilloscope. Focusing is by variation of the potential of the second anode. Electrostatically focused cathode ray tubes are always used in oscilloscopes and the three-anode type of figure is now the most common. The potential of the focusing anode (anode 2) is commonly about l/5th of that of the final anode (anode 3). The first and third anodes are usually at the same potential but this is not essential. In a cathode ray tube used for television a potential of about l/5th of the final anode voltage would be rather difficult to supply. The arrangement is therefore modified so that the focusing anode has a much lower potential, as shown in figure The first anode is fed with a fixed voltage of V ANODE ANODE FIG FOUR-ANODE ELECTROSTATIC LENS

12 THE CATHODE RAY TUBE 7 while the second anode is fed with the same potential as the fourth or final anode. The third anode is the focusing anode and has a potential of between 50 and +450 volts. These lower voltages are easily provided. The final anode is continued as a graphite coating on the neck and part way up the flared portion of the tube. In some designs another anode is used between the first and second anodes of figure 1.12 and this electrode is connected to the third anode, i.e. the focusing anode. This type of electrostatic focusing is now common and dispenses with the need for a focusing magnet but, of course, makes the cathode ray tube more complicated. Ion Trap Negative and positive ions are produced in a cathode ray tube as well as electrons. It does not appear to be understood from where these negative ions originate but they are formed and cause what is known as ion burn on the screen. A negative ion is an atom which has an additional electron attached to it, so that it exhibits a negative charge and travels in the same direction as an electron. An electron is an extremely light body and most of the weight of an atom is on the centre nucleus. Hence, the negative ion is much heavier (may be several thousands times) than an electron. When electrostatic focusing is used these ions are focused on a spot exactly similar to the electrons. When magnetic focusing is used they are not focused appreciably owing to their large mass. When they reach the screen the effect of the ions is to damage it and reduced its fluorescent properties. As a result the portion of the screen bombarded with these ions appears darker than the rest of the screen and is referred to as an ion burn. With magnetic focus this corresponds to the diameter of the unfocused spot 2" to 3" diameter but, with electrostatic focusing (and magnetic deflection) it would correspond to a much smaller spot. The positive ions are atoms which have lost an electron and these are produced by bombardment of molecules of gas (which remain owing to the fact that the vacuum is never perfect), the process being known as ionization. The positive ions are attracted to the negative end of the cathode ray tube, i.e. the cathode. More will be said later about positive ions. It is thought that the negative ions are produced in the vicinity of the cathode and the purpose of the ion trap is to remove them so that they cannot travel to the screen and burn it. It has already been mentioned that the magnetic field causes the electron beam to be bent but it has little effect on the heavy negative ions. This fact is utilized in the ion trap, one design being shown in figure 1.13(a). This is shown applied to a tetrode with magnetic NEGATIVE SECOND FIRST IONS ANODE *NODE \ J CATHODE FIG. 1.13(a). TILTED ION TRAP focusing but exactly similar arrangements may be used with electrostatically focused tubes. The electron gun is now bent so that the electrons start off at an angle to the neck of the tube but they are brought into line by the magnetic field produced by the ion trap magnet. The ions travel almost straight on and hit the side of the final anode, are absorbed and so do not reach the screen. The magnetic field is produced by an ion trap magnet shown in figure 1.13(b).

13 TELEVISION SERVICING (Clomping arrangement not shown) FIG. 1.13(b). ION TRAP MAGNET This clamps on to the neck of the tube. Obviously, the magnetic field must be of the correct polarity and strength and must also be in the correct position. The strength is settled by the magnet which should be the correct one for the particular tube. It is important that the magnet be placed in the correct position because if it is not (apart from the screen not being at full brilliance) some of the electron beam will fall on the final anode and may cause overheating of the anode and possible liberation of gas. The recommended method of adjustment is as follows (/) The ion trap has an arrow marked on it and there is also a mark on the neck of the tube. Place the magnet on the neck of the tube so that the arrow points away from the screen and so that the arrow is diametrically opposite the line marked on the neck of the tube. (If preferred, the arrow may point to the screen, in which case the arrow should be in line with the mark on the neck). (if) With the tube supplies on, adjust the brilliance control to obtain a raster. If this cannot be obtained slide the magnet along the neck until a raster appears. Now carefully slide the magnet backwards and forwards until a point of maximum brilliance is obtained. Also, rotate the magnet slightly to give maximum brilliance. (Hi) Lock in position on the neck. Another method of reducing ion burn is to aluminize the fluorescent screen. This means covering the screen with a thin layer of aluminium which will stop the heavy negative ions but not the lighter electrons. There are other effects of aluminizing which will be considered later in more detail when dealing with screens. Another type of ion trap is shown in figure The gun itself is now 2k*v*JAAA354 FIRST ij»r.mttii- \ SECOND *... FIG ION TRAP straight but inclined at a small angle to the tube axis. The gap between the first and second anodes is now inclined so that an inclined electrostatic field is produced. The electron beam is also inclined by this field but brought back into line with the tube axis by the magnetic field, produced by the ion trap magnet. As before, negative ions will not be affected by the magnetic field and will hit the side of the second anode. The main feature of this design is that it also prevents the positive ions from bombarding the cathode where they tend to damage the oxide coating. It will be seen that the positive

14 THE CATHODE RAY TUBE 9 ions are attracted across the inclined gap and hit the first anode and so do no harm. Ion traps are not used in modern tubes using electrostatic focusing. There is little need for an ion trap in a tube with electrostatic deflection since the ions are also deflected by the electrostatic field and hence are spread out over the whole of the tube and so burn the screen almost evenly. (2) BEAM-DEFLECTING SECTION There are two methods of deflecting the electron beam: by (0 an electrostatic field; («') a magnetic field. (i) Electrostatic deflection In figure 1.15 is shown an electron beam passing ELECTROSTATIC FIELD FIG ELECTROSTATIC DEFLECTION OF CATHODE RAY TUBE between two deflecting plates. The electrons will be repelled from the negative plate and attracted towards the positive plate. Hence, the electrons will travel in a curved path (actually a parabola) between the plates, but will continue in a straight line in a new direction after leaving the plates. If the potential applied to the plates is reversed the beam will be deflected in an upward direction, and the amount of deflection can be shown to be proportional to the voltage applied across the deflecting plates. As we normally wish to be able to deflect the beam in two directions at right angles, these plates are followed by another pair of plates at right angles to those shown. This type of deflection is used in all cathode ray tubes used in oscillographs (together with electrostatic focusing). The main reason is that we are normally interested in observing the waveforms of voltage and, since the plates take no current, they can be connected across most circuits without upsetting the operation of the circuit. This type of tube is not used for television since it is difficult to avoid defocusing of the spot at the edges of the screen, particularly if the angle of deflection is large. The tube is also much longer than a magnetically deflected tube, and it is difficult to supply the high deflecting voltages that would be necessary with a tube operating at a high e.h.t. voltage, as is required in television. The electrostatically deflected tube does not suffer to any extent from ion burn since the electrons and ions are both deflected equally, hence (when operated with a raster as in television) the burning takes place evenly over the whole screen. This means that it is neither noticeable nor serious because it is spread out over such a large area. (ii) Magnetic deflection The effect of a magnetic field on a beam of electrons is shown in figure When the beam enters the magnetic field a force acts FIG ELECTROMAGNETIC DEFLECTION OF CATHODE RAY TUBE on it at right angles to the field and causes it to travel in a circular path through the field and to emerge travelling in a straight line but in another direction. The direction of the deflection so produced can be reversed by

15 10 TELEVISION SERVICING reversing the direction of the magnetic field; the amount of deflection depends on the strength of the magnetic field. To obtain deflection in two directions at right angles one could follow the magnetic field by another at right angles to it, but in order to reduce the tube length, the two magnetic fields act in the same space but, of course, are at right angles to each other. This type of deflection is universally used in television receivers. Deflecting coils are necessary to produce the magnetic field. Before dealing with these let us consider the shape and size of the tube, as this is closely connected with the deflecting coils and deflecting power required. In small tubes a fairly small deflecting angle of about 50 was used because it was easier to deflect the beam over a small angle and less distortion of the spot was likely to occur. With a 9" tube using magnetic focus and deflection the length was about 15" and the general shape is shown in figure A 12" tube, which was popular at one time, followed from this by retaining the same deflecting angle and this is also shown in figure "TUBE 13" TUBE FIG AND 12" CATHODE RAY TUBES USING NARROW ANGLE DEFLECTION The length was now about 18". The length of the tube is most important as it must be possible to fit the tube into a cabinet of reasonable depth. If an electrostatically focused and deflected tube had been used the length would have been much greater say about twenty-five inches for a 12" tube. The next step was a 15" tube which had the same deflecting angle but, of course, was longer than the corresponding 12" tube. Provided the e.h.t. voltage is kept the same, the same deflecting system can be used for all these tubes, since the angle of deflection is the same, which is all that matters as regards the deflecting system. In practice the e.h.t. voltage would be increased for the larger tubes because as the size of screen is increased the area of the picture increases; a 15" tube has an area of over 2\ times that of a 9" tube. If the e.h.t. voltage is kept the same the total light of the screen would be the same (for a given beam current) but, of course, it is spread over a larger area so it will not appear as bright. Hence, the larger tubes are normally operated with a greater e.h.t. voltage. It can be shown that the magnetic field required for a given deflecting angle is proportional to -\/F where. V is the final anode voltage of the tube. Thus, as the e.h.t. voltage is increased the strength of the magnetic field must be increased and this means more power in the deflecting coils. As the demand grew for larger tubes it was necessary to alter the deflecting angle as the length of the 15" tube was becoming excessive. The angle was now increased to 70 and, at the same time, the rectangular tube was introduced. The difference in shape of a narrow angle (i.e. 50 and a 70 tube

16 THE CATHODE RAY TUBE 11 FIG. 1.18(a). 12" TUBE, 50 FIG. 1.18(b). 14" RECTANGULAR, 70 (shown across diagonal) is shown in figure 1.18(a) and (b) when it will be seen that the length is considerably reduced. Using a rectangular tube, although it does not alter the length required, does save considerable space as can be seen in figure 1.18(g), enabling a smaller cabinet (in width and height) to be used and yet obtain the same size of picture. The first rectangular tube had a 14* diagonal. A 17* diagonal tube (figure 1.18(c)) was commonly used and 21* diagonal FIG. 1.18(c) 17" RECTANGULAR, 70 (shown across diagonal) rectangular tubes were used. A 21* 70 tube is rather long (see figure 1.18(d) ) so tubes were made with a 90 deflecting angle, the reduction in length being shown in figure 1.18(e). A 21* 70 electrostatically focused tube is about FIG. 1.18(d). 21" RECTANGULAR, 70 MAGNETIC FOCUS (shown across diagonal) FIG. 1.18(e). 21" RECTANGULAR, 90 MAGNETIC FOCUS (shown across diagonal) 1* shorter than the corresponding magnetically focused tube, due to the reduction of length of neck. Increasing the angle to 70 means (for the same voltage) a greater magnetic field for deflection and hence more power e.h.t.

17 12 TELEVISION SERVICING into the deflecting coils. The 90 angle means still more power. Electrostatically focused tubes with 110 deflecting angle are now used and these require still more deflecting power. A 17* tube with this angle and electrostatic focus is shown in figure 1.18(f) where it will be seen that an important reduction in length results. The deflection over this large angle does, of course, cause some difficulties. - 12/2 FIG. 1.18(f). 17* RECTANGULAR, 110" (shown across diagonal) SPACE SAVED RECTANGULAR FIG. 1.18(g) SAVING OF SPACE BY RECTANGULAR TUBE FIG. 1.18(h). 19* RECTANGULAR TUBE FIG. 1.18(i). 23* RECTANGULAR TUBE Shortly after the introduction of the 110" 17* tube the shape of the screen was changed. The original rectangular tubes were, in fact, far from rectangular, the corners being considerably rounded. The new shape is much more rectangular so that although the new tubes are approximately the same size (as regards height and width) as the 17* and 21* tubes the diagonal was increased (by the use of sharper corners) to 19* and 23*. The tube size is, of course, given as the size of the diagonal. The effect of the change in shape is shown in figure 1.19 where the old 21* tube is compared with the new 23* tube. The reduction in length of the new tubes compared with the older ones is illustrated in figure 1.18(h) and (i). It will be seen that the new tubes have an even shorter neck. The newer tubes also have a smaller diameter of neck as seen in the figure. This reduces the size of deflecting coils required. It should be noted that the ratio of width to height of a modern rectangular cathode ray tube is approximately 5/4 and not the aspect ratio of the picture transmitted, i.e. 4/3 (see Volume 1). This means that if the width of the picture is adjusted to be correct (i.e. just to fill the screen) there will be black spaces at the top and bottom of the picture. In general it is preferable to adjust the height of the picture so that it just fills the screen and then adjust the width to give a 4/3 aspect ratio (i.e. a circle on a test card) and under these conditions the horizontal scan will exceed the width of the tube and a small

18 THE CATHODE RAY TUBE 13 NEW RECTANGULAR TUBE FIG MODERN RECTANGULAR TUBE portion of the picture will be missing. This is not important and has some advantages in dual-standard receivers (see page 108). As well as increasing the size of picture, the e.h.t. has risen from 6kV with a 9* tube to 14kV with a 17" diagonal tube and to 20kV with a 23* tube. This increase in e.h.t. means considerable increase in deflecting power. Thus, with the combination of increased e.h.t. and increased angle, the power requirements for the deflecting coils has increased greatly and, if more efficient circuits had not been developed, compared with the early circuits, it would appear that we should not have been able to provide this greatly More will be said about this when increased power with a practical circuit. dealing with line deflecting circuits in Chapter 5. The simplest deflecting system is shown in figure 1.20(a) which shows one set of coils only. The two coils are placed at the side of the tube and f ) FIG DEFLECTING COILS (a) Basic arrangement; (b) Practical arrangement. produce an approximately uniform field across the neck of the tube. The only portion of the magnetic field which is useful, as regards deflection, is obviously that which crosses the neck of the tube and which, in the arrangement shown, is only a small fraction of the total. Thus, the efficiency of the arrangement is poor and a large stray magnetic field is produced. As a result, this arrangement is not used and in practice the coils are bent round the neck of the tube as shown at (b) (one set only). By suitable spacing of the sides of the coils an approximately uniform field can be produced. If the field is not uniform the deflecting angle will not be proportional to the current in the coils. Although it might appear that a uniform field is required if we are not to get any distortion or non-linearity, in practice, since the centre of curvature of the screen does not correspond to the position of the deflecting coils, it may be necessary to have a non-uniform field. In some cases the coil may be split into a number of coils and the turns on the various coils graded so as to obtain the required field distribution. In older deflecting

19 14 TELEVISION SERVICING coil assemblies a core of laminated iron was arranged round the outside of the deflecting coils, so as to concentrate the flux, reduce the stray field and also reduce the reluctance of the magnetic circuit. The deflecting coils may be wound as shown in figure The coils FRAME COIL CASTELLATED FERROXCUBE CORE FIG DEFLECTING COIL YOKE are arranged in a castellated Ferroxcube core, the two sets of coils being at right angles as shown. The line coils produce a vertical field, so deflecting the beam horizontally and these are placed nearest to the neck as more power is required for the line coils than the field. The field coils produce a horizontal magnetic field, so deflecting the beam vertically. The use of a Ferroxcube core reduces the losses in the deflecting system and reduces the reluctance of the magnetic circuit and so produces a stronger field for a given current. The ends of the coils have a considerable influence on the distribution of the field produced and care is necessary in maintaining the correct shape of coil. This is particularly important in 90 and 110 deflecting coils, where the coils are taken part way up the flared portion of the tube. In modern deflecting coil assemblies the castellations of the Ferroxcube cores are done away with and the coils tend to be much flatter and more spread round the neck of the tube. In tubes using 110 deflection the coils are taken quite a long way up the flared portion of the tube and their shape is most important if defocusing and other troubles are to be avoided. An alternative arrangement that is used is shown in figure The coils are wound as shown and produce a vertical magnetic field which gives the MAGNETIC FIELO FIG TOROIDAL DEFLECTING COIL YOKE line deflection. Field coils (not shown) are wound on limbs A and B so producing a horizontal magnetic field for the field deflection. It is claimed that this arrangement is more efficient, but the coils are difficult to wind. This arrangement for field deflection is sometimes used together with coils as in figure 1.20(b) for line deflection.

20 THE CATHODE RAY TUBE 15 FIG DIAGRAM SHOWING SHADOWING DUE TO DEFLECTING COILS BEING TOO LONG The longer the length of the coil the less the strength of magnetic field required. The length must not be too great or, as is seen in figure 1.23, the beam will strike the shoulder of the tube before it is deflected to the edge of the screen. This is particularly so with 90 and 110 deflecting systems and the coils have to be shorter than with 50 and 70 systems. To reduce this effect the end of the coils are actually taken well up the flared portion of the tube. The shorter length of coil increases still further the power required in the deflecting coils; the power required for 90 deflection is about twice that of 70 deflection. The actual turns required on the coils, which settle the impedance, will be considered when dealing with deflecting systems. (3) VIEWING OR SCREEN SECTION The beam of electrons is invisible and must be made visible by the use of a fluorescent screen. The electrons have energy by virtue of their velocity and mass. The extremely high velocity makes up for their very small mass. This energy is converted into light energy by the fluorescent coating resulting in a glow wherever the electrons strike the screen. The brilliance of the spot depends on: (0 the energy of each electron; and (//') the number of electrons reaching the screen per second. The accelerating voltage is determined by (0 and the beam current by (i'»). A large current results in a large beam and tends to give a large spot, hence the anode voltage is increased as the tube size increases to give increased spot brilliance. Since this is spread over a larger area the actual picture brightness may not be altered much. There is a limit to the high voltage which may conveniently be used, owing to insulation problems, but the e.h.t. rises continually as the maximum tube size increases from year to year. When the screen is not aluminized the electrons return from the screen to the anode by secondary emission from the screen material, which is a material chosen to give a picture that is substantially white and has a fairly short afterglow. (Afterglow is the time that the screen continues to emit light after the electrons have ceased to bombard the screen). The material is commonly coated on the face of the tube by settling a suspension of the phosphor in water and decanting the water. Many tubes have aluminized screens. This means that a thin layer of aluminium is coated on the back of the phosphor by evaporating aluminium from a filament during the manufacture of the tube. The aluminium coating does three things: (0 It reduces the possibility of ion burn because the penetration of the electrons through the aluminium is much greater than that of heavy negative ions. Some loss of electron energy occurs but this is not important, particularly above about lokv anode voltage. (if) It increases the light output by reflecting in the forward direction the light given out by the phosphor in the backward direction. The aluminium layer is put on so that a highly reflecting surface is produced. This results in

21 : 16 TELEVISION SERVICING quite an important increase, as at least 50 per cent, of the light output from a normal screen goes to the back of the tube. (hi) It forms a conducting layer on the fluorescent coating which enables the electrons to return to the anode (the coating being in contact with the anode) without the process of secondary emission. The exact behaviour of secondary emission is too involved to be dealt with here* but, in certain cases (particularly at high anode voltages and beam currents) the process may not act successfully with the result that the light output does not increase with anode voltage and beam current as it should. Most cathode ray tubes have a graphite coating on the outside which is earthed. The purpose of this, together with the inside coating connected to the anode, is to form a smoothing capacitor for the e.h.t. and save an additional high voltage capacitor. In older tubes it is necessary to have some form of implosion protection so that should the tube crack and implode {implode rather than explode since the pieces of the tube are initially sucked violently towards the centre) the viewer will not suffer injury. The effects of an imploding tube can be very serious the force applied to the front face of the tube due to atmospheric pressure is about li tons. Originally, special armour plate-glass was used to protect the viewer. This was later replaced by moulded plastic protective covers, moulded to approximately the shape of the tube and so avoided the large gap between the tube and the protective cover at the edges of the screen. The drawback with this arrangement is that dust collects between the tube and the protective screen however good the seal may be between tube and protective cover. Periodic cleaning is necessary or the light output of the tube is greatly reduced. This is a job which requires the services of a television service engineer and so is often neglected by the viewer. More recently, tubes have been produced with bonded safety or protective screens. The plastic screen is moulded to fit the tube and is bonded to the glass with a special resin. This has the advantages of making a stronger tube and reducing the possibility of violent implosions in fact it is claimed that the tube just breaks or cracks rather than implodes, owing to the greater strength of the front face. When a separate protective screen is used there are four partially reflecting surfaces: two on the glass; and two on the safety screen. These surfaces reflect objects in the viewing room and may interfere with the picture. Further, they cause part of the light given out by the phosphor to be reflected back to the screen, so reducing the brightness and contrast of the picture. They also cause halos around bright objects in the picture. By bonding the safety screen and glass together with a resin of reflective index approximating to that of glass the reflecting surfaces are reduced to two, reflection of objects is reduced and more light is available from the screen. Lugs are moulded on to the plastic screen to assist in mounting the tube in the cabinet. Obviously, no dirt can now penetrate between the tube and the safety screen, the front of which can easily be kept clean by the viewer. Only soap and water should be used: solvents must be avoided as they may attack the plastic of the screen. In many tubes the plastic safety screen is tinted so that the picture contrast is increased and the black of the picture made to look blacker. On a tube with no tinted screen (since the phosphor is white) any ambient lighting causes the screen to look white and the black portions of the picture only look black because they are not as white as the bright portions of the picture. If a tinted screen is used some light is lost from the phosphor screen since this light has to pass through the screen, but only once. However, the ambient light has to pass through the screen twice: to illuminate the phosphor; and when it is reflected from the phosphor. Thus, the light from the screen due to ambient * Further details of secondary emission are given in Volume 3.

22 THE CATHODE RAY TUBE 17 lighting is reduced more than that from the phosphor due to the electron beam. The blacks of the picture appear more black and a more contrasty picture is obtained when the tubes are used under conditions of high ambient lighting. Obviously, a compromise must be found between loss of light from the phosphor and increase of contrast. A more recent improvement to cathode ray tubes is the P-tube which has a metal band around its edge. The purpose of this band is to prevent an implosion and obviously is much better than previous precautions intended to contain the effects of implosion. The band prevents expansion or deformation of the tube should a crack occur, hence preventing the crack from opening or spreading. It has been proved that this device is a perfectly safe precaution against implosion; therefore no protective screen is required in front of the tube. The tube often protrudes beyond the front of the cabinet so that a slim cabinet can be used. The band comprises a cadmium plated ring of mild steel and is sealed by a polyester resin to the periphery of the face plate of the tube, which is the region having the highest stress. Four lugs are on the ring to facilitate fixing of the tube. The ring should be connected to chassis by a resistor of about 2MQ as a safeguard against the ring acquiring a high potential.

23 CHAPTER 2 SYNCHRONIZING SEPARATORS purpose of the synchronizing separator is to separate the synchronizing The pulses from the signal and also to separate them into line and field pulses. We can divide the separator section into two parts: (1) The picture synchronizing separator. The purpose of this is to separate the pulses from the vision component of the composite signal. (2) The line and field synchronizing separators. The purpose of these sections is to separate the line and field pulses so that they may be fed to the appropriate timebases. We shall deal with these parts separately. (1) THE PICTURE SYNCHRONIZING SEPARATOR In this section of the television receiver we have to separate the synchronizing pulses from the picture signal and it is important that the separation be as complete as possible. If the picture component is not completely removed then synchronizing will be poor and the synchronizing will depend on the content of the picture, particularly of subjects on the right hand side of the screen. The synchronizing separator is basically a clipper or limiter, which prevents the passage of any signal above (say) 25 per cent, of the peak white signal. The principle is shown in figure 2.1. Suppose that we have some FIG PRINCIPLE OF PICTURE SYNCHRONIZING SEPARATOR device which conducts when a positive voltage is applied to it, but not when a negative voltage is applied (such as a rectifier) and that it has a characteristic as in figure 2.1. When the composite signal is applied to this device only that portion to the right-hand side of line OX will cause current to flow, and so give an output. If we position the waveform correctly, then we can arrange that the synchronizing signals are passed (as in figure 2.1) but that the picture portion is suppressed. The important factor is the positioning of the waveform relative to the characteristic of this one-way device. To position the waveform correctly we must maintain, or restore, the d.c. component of the signal. It is, therefore, necessary to say more about the d.c. component. 18

24 I SYNCHRONIZING SEPARATORS 19 This has already been mentioned briefly in Volume 1 but it is necessary to go now into more detail. Unlike the sound signal, which is purely alternating in nature, a television signal contains a steady, or d.c. component, which represents the average brightness of the scene being televised. The d.c. component is the value we should get if we removed the variations of voltage from the signal, i.e. it is the voltage we should get if we passed the signal through a suitable smoothing circuit. Suppose that the scene being televised consisted of a white background with a narrow vertical black bar. The resultant line waveform would be as shown in figure 2.2(a). The average D. ;. CO rfponent u rjjj i-l J-l ilfj-/ l-lj-> ris. component o-l (0) (b)!t u \f ^r- FIG IMPORTANCE OF D.C. COMPONENT <a> & (a) and (c). White picture with narrow black bar; (b) and (d). Black picture with narrow white bar. D.C. component retained in (a) and (b); d.c. component missing in (c) and (d). light would be large resulting in a large mean value, or d.c. component, of the waveform as shown. If, on the other hand, the background had been black with a narrow vertical white bar the waveform would have been as shown at (b). The average light is now small and, therefore the mean or d.c. component of the waveform is small, as seen in the figure. Waveforms (a) and (b) are typical of those produced at the demodulator where the d.c. component is present. It is common practice to feed the synchronizing separator from the anode of the video amplifier in order to obtain sufficient signal amplitude. To remove the steady positive voltage which is present on the anode of the video amplifier, the synchronizing separator is usually fed through a capacitor. This, of course, also removes the d.c. component, since d.c. cannot flow through a capacitor. Let us now see what difference this would make to the waveforms. Since there is no d.c. component, the d.c. component lines of figure 2.2(a) and (b) must now lie on the zero line, as shown in the figure at (c) and (d). Put in another way, the absence of the d.c. component means that the average value of the waveform over a line must be zero, or that the area enclosed above the zero line must equal that which is enclosed below the zero line. Thus, if we passed the waveform through a smoothing circuit we should finish up with nothing. If we examine waveforms (a) and (b) it will be seen that the synchronizing pulses occupy the same voltage variation in both cases, hence the picture component may be removed by clipping off

25 20 TELEVISION SERVICING everything above a certain value, say that corresponding to the d.c. component line shown at (b). In the case of waveforms (c) and (d) the synchronizing pulses do not correspond to the same voltage. Suppose that we clipped everything above a voltage corresponding to line XY. At (d) we should separate the pulses from the picture, but if we clipped at the same voltage level in case (c) we should not separate the picture from the synchronizing pulses. Thus, if the waveform is allowed to float up and down as at (c) and (d), due to the absence of the d.c. component, it will be quite impossible to find a voltage level for clipping or limiting which will remove the picture signal for all pictures. This means that the d.c. component must be present and, if we cannot maintain it by direct connection from the anode of the video amplifier, it must be restored by what is called a d.c. restorer. The circuit of a d.c. restorer is simple but its operation is not so easy to explain. Let us consider the circuit shown in figure 2.3. C is the coupling INPUT OUTPUT R <> (b) FIG D.C. RESTORER, (a) Circuit; (b) Waveform capacitor and R is often the grid resistor of the synchronizing separator valve, i.e. C and R form the normal CR circuit used in resistance-capacitance coupling. The d.c. restoring action is produced by the diode valve V. Imagine that the waveform across A and B has no d.c. component, i.e. it is similar to figure 2.2(c) and (d). When the signal goes in a positive direction (i.e. B is positive with respect to A) a current will flow from B to A, through R, tending to charge C. No current will flow through the valve since this is the non-conducting direction. C and R are arranged to have a large time constant (i.e. C = O-l^F and R = 1MQ), hence C will not charge appreciably and all the voltage appears across R. If V were not present the same thing would happen on the negative portion of the waveform but, with V present, the action is different. When the input waveform goes negative (i.e. B is negative with respect to A) a current will flow in the direction shown tending to charge C. Since the valve conducts in this direction the current will flow through V rather than through R. The resistance of V is, of course, much less than R (say 1,000 ohms instead of lmo), hence the time constant is short This means that C charges up rapidly with the polarity shown to almost the largest negative input voltage. When the input voltage decreases, C cannot follow it as the valve V prevents C discharging and the high value of R makes the discharge time very long. If figure 2.2(a) and (c) are examined it will be seen that the largest negative voltage of the waveform at (c) corresponds to the d.c. component of the signal. Thus, C is charged to the d.c. component of the input voltage during the first line after switching on and remains charged to this voltage. The output voltage is the input (as 2.2(c) ) plus the d.c. component voltage across C, which raises the waveform up so that it

26 SYNCHRONIZING SEPARATORS 21 appears as at 2.2(a). This means that the waveform always sits on the synchronizing pulses as shown in figure 2.3(b). It is seen that there are two time constants in this circuit: a large one for current flowing from B to A and a short one for current flowing from A to B. During the period between pulses C discharges slightly through R, but each time a synchronizing pulse arrives a current flows to recharge C. The operation of the circuit is really quite involved but the above explanation is sufficiently accurate for our purpose. It may be interesting to note that this circuit is identical to the diode demodulator which is described in Radio Servicing, Volume 2. If the waveform is of opposite polarity the d.c. component can be restored by reversing the diode as shown in figure 2.4. The action is essentially similar but will now be seen that the output waveforms hangs from the zero line INPUT + AV- c OUTPUT (<0 «0 FIG D.C. RESTORER, (a) Circuit; (b) Waveform by the tips of the synchronizing pulses. This connection of the d.c. restorer is the more usual, because it is the polarity of signal which is normally available at the anode of the video amplifier. Having seen how we may restore the d.c. component of the signal let us now return to the synchronizing separator itself. The simplest synchronizing separator to understand is probably that of figure 2.5. The potential divider P is adjusted so as to produce a voltage i i FIG. 2.S. DIODE SYNCHRONIZING SEPARATOR USING MANUAL CONTROL OF BIAS somewhat less than the magnitude of the synchronizing pulses, i.e. just below the black level of the signal. Since the diode will conduct only when the cathode of the diode is negative with respect to the anode, only the synchronizing pulses will be passed to the output, owing to the fact that the cathode is driven positive by the picture portion of the signal. To be more exact, when the input is zero a current flows in i?i in the direction shown and, neglecting the drop in the diode, the anode and therefore the output

27 22 TELEVISION SERVICING terminal, are at zero potential, a voltage drop V occurring across R x. As the input voltage increases, at the end of the pulse, this current is reduced until it becomes zero at a point when the input equals the voltage V. As there is now no current there is no drop in R x and the output terminal is at the potential V. When the input increases above V the diode is non-conducting and still no current flows. Accordingly, there is no change in the voltage of the output terminal whatever the variation of the input, provided it is greater than V. The circuit may be fed directly off the anode of the video amplifier valve, so long as the voltage V is increased to overcome the effect of the steady voltage on the anode of the valve. Although the circuit is simple it has the disadvantage that P must be adjusted to suit the magnitude of the input signal and, what is more important, some picture signal is fed through the anode-cathode capacitance C^.. The anode-cathode capacitance forms a potential divider with the stray capacitance from the anode to earth C?e and is independent of the conduction of the an appreciable picture diode. Since C ac is appreciable compared with CM signal is fed to the output, and it is often of sufficient magnitude to upset the synchronizing of the timebase. Owing to these faults this circuit is not used in modern television receivers, and it has been described because it demonstrates the principle in a simple manner. Instead of using a fixed bias from a potential divider an automatic bias version may be used as shown in figure 2.6. Suppose that the circuit is fed ^H&, Rl («> >R2 _n_ -JL_ w (O ojt JT Jl_ OUTPUT (voltage acros$ R2). FIG DIODE SYNCHRONIZING SEPARATOR WITH AUTOMATIC BIAS (a) Circuit; (b) Waveform of input voltage and voltage across Ci; (c) Output voltage waveform from the anode of the video amplifier valve so that the waveform is lifted due to the steady positive anode voltage. A current will flow through V x and R z charging up C x with the polarity shown. C x will become charged to the peak positive input, i.e. the top of the synchronizing pulses. During the period between pulses the diode is not conducting (since the anode is negative) and C x discharges slowly through R x. Thus, when each pulse arrives a sudden surge of current flows in V x and R 2 to recharge C x. This causes a corresponding voltage drop across R 2 and is shown at (c). Provided the component

28 SYNCHRONIZING SEPARATORS 23 values are suitable and we neglect the anode-cathode capacitance, the picture signal will be removed and the output will consist of pulses. The effect of the anode-cathode capacitance causes some picture signal in the output and, for this reason, the circuit is not generally used. It has been used in a slightly modified form in a receiver in conjunction with the pentode synchronizing separator described later. It is important to note that the circuit will only operate with the input of correct polarity, i.e. that shown. There are other diode circuits but as they are not used they will not be described. The main fault with any diode circuit lies in the fact that the separation is not complete. The arrangement almost universally used in modern receivers consists of a pentode, and the basic circuit is shown in figure 2.7. This circuit will FIG CIRCUIT OF PENTODE SYNCHRONIZING SEPARATOR only operate with a negative video signal (i.e. positive synchronizing pulses). This is the polarity available at the anode of the video amplifier, using cathode modulation of the cathode ray tube, hence this fits in with the modern practice of cathode modulation. The capacitance C t will remove the d.c. component (which is, of course, present at the anode of the video amplifier) but the grid and cathode of the pentode act like a diode (grid current flows when the grid is positive in the same way as when the anode is positive in a diode). The circuit is therefore identical to that of figure 2.4, with the grid and cathode replacing the diode. Hence, the waveform hangs from the zero line as shown in figure 2.4(b). The pentode must be a sharp cut-off valve and the grid base is made small by operating the valve at a low screen grid voltage. This is achieved by either a high screen dropping resistor (say lmci) or a suitable potential divider as in figure 2.7. The action of the pentode is shown in figure 2.8. The input must be sufficiently large so that the magnitude of the synchronizing pulses is greater than the grid base of the valve. This means that all the picture component must be to the left-hand side of line AB. Between pulses, the valve is cut off and so no picture signal is fed to the output. The capacitance between input and output circuits (i.e. C ga ) is now extremely small due to the screening action of the screen grid and so there is no feed to the output, as occurs with a diode. During the synchronizing pulses (either line or field) the grid is at zero potential or slightly positive (see later) and anode current flows. The pulse of anode current, flowing in the anode load R 4, causes a drop in anode potential and hence a negative output pulse. Note that this circuit causes a reversal of the pulse from input to output, unlike the diode circuits. The pulse is, of course, also amplified which is not so with the diode circuits.

29 24 TELEVISION SERVICING Output anode current PRINCIPLE OF OPERATION OF PENTODE SYNCHRONIZING SEPARATOR The above method of operation would be satisfactory if there were no "noise" with the signal but, unfortunately, noise is always present to some extent and is particularly severe in fringe areas. As it is essential that the timebase should not be triggered by this noise it is desirable to use only a portion of the synchronizing pulse such as between lines AB and CD of figure 2.9. In other words, we want to use a slice out of the centre of the FIG LINE SYNCHRONIZING PULSE WITH "NOISE" synchronizing pulse. This can be done quite easily by the circuit of figure 2.7 and by choosing suitable component values. Earlier it was stated that the waveform hung from the zero line. In practice this is not quite true, as this assumed zero internal resistance of the device feeding the synchronizing separator and zero resistance of the diode (or grid and cathode circuit in this case). Since the diode has an appreciable resistance the grid must be driven somewhat positive to pass a current through the valve, in order to charge capacitor Cy If a resistor is also placed in series with the capacitor C it will make little difference when the grid is negative, since no current flows in the circuit. On the other hand, when the input goes positive, current flows, so reducing the voltage actually fed to the grid. In other words, when the input goes positive little change takes place in the voltage across the gridcathode of the valve, due to the flow of current and this, therefore, clips the top of the synchronizing pulses corresponding to line AB of figure 2.9. Much

30 . SYNCHRONIZING SEPARATORS 25 of the noise is thus removed. The amount of clipping may amount to 2 to 3 volts, depending on the value of the resistor used. The resistor must not be made too high because, together with the input capacitance of the valve, it forms a filter reducing the high frequency response. This causes distortion of the pulse and the effect known as pulling on whites occurs, i.e. the lines of the picture are moved to the left whenever a white object appears on the right-hand side of the picture. The reason for this is that if the picture has a black right-hand edge as in figure 2.10(a) the start of the (a) Block object on extreme right picture. (b) White object on extreme right of picture FIG EFFECT OF TOO LARGE A RESISTOR IN THE GRID CIRCUIT OF THE SYNCHRONIZING SEPARATOR pulse occurs at the same instant whether the resistor is present or not, even when the resistor appreciably upsets the frequency response and shape of the pulse. If the picture has a white right-hand edge, as in figure 2.10(b), and a high resistor is used, the voltage does not drop from the peak white to black level before the start of the pulse occurs, and so the front edge of the pulse is delayed from its normal position. This causes the timebase to be triggered late with the result that the start of the next line is late, and the picture appears to be shifted to the left. The purpose of the front porch is to allow the voltage to drop to black level before the start of the pulse, and it will do this so long as the frequency response is adequate. Since the front porch is only lj/xs long on 625 lines, to prevent pulling on whites the time constant, i.e. the product of the series resistor (plus the internal resistance of the source which has the same effect) and the input capacitance of the valve should not exceed about 0-2/xS. In some cases the resistor is not a physical component, the internal resistance of the source {i.e. video amplifier) is sufficient. When a resistor is used it is usually about 10kO. Limiting in the other direction (corresponding to line CD of figure 2.9) is settled by the magnitude of the grid base of the valve, relative to the signal amplitude. Therefore, if the circuit is designed correctly, any reasonable noise occurring on the front and back porches will be removed by the synchronizing separator, provided the cut-off point of the valve corresponds to a line such as CD. As well as the limiting action on the grid the performance may be further improved by a suitable choice of the anode resistor and the valve. The anode current can be limited in the maximum direction as in figure 2.11(a). This figure shows typical characteristics of a modern high slope pentode valve which, of course, has a short grid base. The load line is plotted from the point

31 26 TELEVISION SERVICING ANODE CURRENT DYNAMIC CHARACTERISTIC INPUT WAVEFORM ON GRID FIG LIMITING ACTION OF SYNCHRONIZING SEPARATOR VALVE on the horizontal axis corresponding to the h.t. supply voltage and the slope of this load line is made the reciprocal of the value of the anode resistor. The intersection of this load line with the valve characteristic then gives the operating point for any grid voltage. It will be seen that the operating point is almost identical for grid voltages between zero and 2 volts and hence the dynamic characteristic {i.e. the valve characteristic including the effect of the anode resistor) as plotted at (b) is almost horizontal from to about 2 volts. As the grid of the valve is made more negative than 2 volts the points of intersection on the load line move to the right (figure 2.11(a) ) and the anode current decreases rapidly to zero at about -4i volts. It will now be seen that the noise at the tip of the pulses corresponds to the flat portion of the K~ v g\ characteristic [figure 2.11(b)] and hence the noise is not reproduced in the output. Noise on the front and back porches is removed because this is more negative than 4i volts and is beyond the cut-off point of the valve.

32 SYNCHRONIZING SEPARATORS 27 FIG TYPICAL PENTODE SYNCHRONIZING SEPARATOR A typical circuit is shown in figure In this the low screen voltage is obtained by the potential divider. In one modification of this type of circuit a choke of approximately 1 henry inductance is connected in the anode circuit, This results in a very sharp leading edge in place of the resistor i? 4 (figure 2.7). to the pulse because, when the valve conducts, current cannot suddenly rise in the inductance and so the current must be fed from the stray capacitance across the anode. Since this is very small and the anode current is large, a very rapid drop in anode voltage occurs until the current in the choke can increase. A small damped oscillation is produced at the end of the pulse, but is of no consequence. In some receivers additional limiting and clipping may be performed by additional valves, operating in a similar manner, but these circuits will not be considered here. Although other circuits are possible as a picture synchronizing separator they are not worth mentioning, as the pentode synchronizing separator now appears to be universally used in valve circuits because it is simple and extremely satisfactory. Although the pentode synchronizing separator just described is satisfactory for 405-line receivers and is used on most dual-standard receivers at the present time (1969), a new type of synchronizing separator known as a noise gated synchronizing separator may be used on 625 lines. On 405 lines where positive modulation is used, interference causes large pulses in the peak white direction, causing white spots on the screen of the cathode ray tube, but the interference has little effect on the synchronizing pulse region. On 625 lines, where negative modulation is used, the effect is reversed. Interference pulses not only cause black spots but they extend into the synchronizing pulses region. The carrier for a 625-line system is shown in figure 2.13(a) and the demodulated waveform (as at the anode of the video amplifier) is show at (b). It will now be seen that the interference pulses extend into the synchronizing pulse region and are usually of a much larger amplitude than the synchronizing pulses. The waveform of figure 2.13(b) is that which is fed to the synchronizing separator where d.c. restoration takes place at the grid. In the absence of interference pulses the waveform is d.c. restored so that the tips of the synchronizing pulses lie on the zero line (or approximately so). This is shown in figure 2.14 at (A). When large interference pulses arrive they drive the grid into grid

33 28 TELEVISION SERVICING FIG EFFECT OF INTERFERENCE ON NEGATIVELY MODULATED CARRIER SYSTEM (a) CARRIER (b) DEMODULATED WAVEFORM FIG BLOCKING OF SYNCHRONIZING SEPARATOR VALVE DUE TO INTERFERENCE PULSES current and tend to d.c. restore the circuit so that the tips of the interference pulses are now on the zero line. These pulses are obviously reproduced in the output and may cause false triggering of the timebase. If the interference pulses are large enough then the increased bias (due to the flow of grid current as above) will cause the waveform to be displaced as at (B) so that the synchronizing pulses no longer come into the grid base of the valve and hence there are no output pulses irom the synchronizing separator. The increased bias due to the d.c. restoring of the interference pulses will gradually decrease (as the grid capacitor discharges through the grid resistor) and the output pulses will be restored. The result is that a few line pulses (or field pulses) are

34 SYNCHRONIZING SEPARATORS 29 missing as shown in the anode current waveform. In other words the synchronizing separator may be "blocked" for a period after a large interference pulse (or a number of interference pulses considerably greater than the synchronizing pulses). There is, of course, nothing that can be done after this stage and some arrangement must be used to prevent this blocking. One method of reducing the effect is shown in figure This is a conventional pentode synchronizing separator apart from the introduction V\A FIG CIRCUIT TO REDUCE THE POSSIBILITY OF BLOCKING OF SYNCHRONIZING SEPARATOR VALVE of CxRi. The time constant of C^ is much smaller (about 40/iS) than the time constant of C 2 R 2 which has a value of about 50,000 to 100,000/xS. When a pulse of interference occurs the valve takes grid current and C t becomes charged in such a direction as to make the grid of the valve negative and so reduce the grid current, i.e. tending to d.c. restore the interference pulse. When the pulse has passed C l discharges rapidly through R t (in a period normally less than a line) and the bias produced by the grid current is rapidly removed so that the circuit is not blocked for any appreciable time. The purpose of the noise gated synchronizing separator circuit is to prevent the synchronizing separator operating when a noise pulse occurs, i.e. it is gated out of action by the noise pulse itself. A special heptode valve is normally used and the basic circuit (Mullard) is shown in figure The basic operation of the circuit is the same as the pentode synchronizing To anode of video amplifier FTT To qrid of *7k!l - W IMSl video amplifier FIG NOISE GATED SYNCHRONIZING SEPARATOR CIRCUIT separator, except that the video signal from the anode of the video amplifier is now fed through d and C 2 to grid 3 of the heptode valve (in place of grid 1 of the pentode). So long as the first grid allows electrons to pass then grid 3 takes grid current and the d.c. restoring action takes place as in the

35 30 TELEVISION SERVICING pentode. The grid 3 base is short and only synchronizing pulses appear in the anode circuit. The principle of the gating operation is to feed noise pulses only to the first grid in opposite phase to those on grid 3. When a positive noise pulse occurs on G 3 of the valve a negative pulse is fed to G x and the valve is cut off. Hence G 3 cannot pass grid current (since there are no electrons for the grid to collect) and blocking cannot occur. Also, since the valve is cut off, no interference pulse can occur in the anode circuit. The valve is usually cut off between synchronizing pulses and although the interference pulse causes the voltage on G 3 to come within the grid base, the pulse fed to G t which is of opposite phase (i.e. negative) will simultaneously cut off the valve. Thus, as far as the anode circuit is concerned the valve is cut off all the time and no pulses occur. In practice some small pulses will occur in the anode circuit but they will be of negligible amplitude. For the above operation to take place we have to feed an opposite signal to Gi and this is done by taking the signal from the grid of the video amplifier which is, of course, of opposite phase to that in the anode circuit which feeds G 3. It is necessary to introduce a limiting action because if the complete video waveform were fed to G t then the synchronizing pulses would also be cancelled by the opposite pulses fed to G u This limiting action is obtained by G t itself. The grid is fed through C 3 (which is just a d.c. blocking capacitor) and resistor R 3. The grid is fed with a positive voltage from the potential divider through resistor R t. With no signal input applied to R 3 current will flow in the Gi circuit through R t and from the grid to the cathode, the grid and cathode acting as a diode. Since this diode is conducting, its impedance will be low and when a small voltage is applied to R 3 little voltage will appear on G, (since the grid and cathode act as a short circuit) and most of the voltage will be dropped across R 3. At a certain voltage level fed to R 3 the grid will be driven negative (i.e. the current flowing in R 3 will be greater than that originally flowing in R+ and the current tries to reverse in the grid-cathode circuit. As it cannot reverse, the grid cathode circuit appears as an open circuit). The voltage now applied to R 3 will appear on Gi and the grid will be driven negative. The action of the grid is similar to the delay diode used in vision a.g.c. circuits and described in more detail in Volume 3. By suitable choice of voltage from the potential divider the grid can be made to remove the video and synchronizing signals and only noise signals (greater than the synchronizing pulses) will appear on the grid. The action is shown in figure The setting of potential divider P is rather critical. If the positive voltage from P is too low then correct clipping will not occur and some _Voltage on C3 from anode of video amplifier ^/\ / \ /r \fv~-ytsssrx "T CUPPING LEVEL (clipping above this line) (from video» amplifier grid J FIG Voltage on C t Anode voltage (output) ^nnrr OPERATION OF NOISE GATED SYNCHRONIZING SEPARATOR CIRCUIT

36 i, 1 t 1 1 SYNCHRONIZING SEPARATORS 31 synchronizing pulses will be fed to G 1 and some cancellation of line and field pulses will occur. If the voltage is too high, the interference pulses (unless they are very large) will not cut off the valve and interference pulses will occur in the output and the valve may be blocked. The heptode may be followed by a further stage of clipping, reducing the capacitance loading on the heptode anode circuit as well as cutting out still further any interference pulses. One arrangement of clipper is given in figure Grid current flows through the anode resistor of the heptode and through resistor R 6. A certain negative input must be applied before the valve innr Ji_rt_ or 470kft i WNt HI FIG LIMITED STAGE FOLLOWING THE NOISE GATED SYNCHRONIZING SEPARATOR CIRCUIT grid goes in a negative direction, the action being similar to that already explained for G t of the heptode. The signal is, of course, fed through capacitor C 6. When the input exceeds a certain negative value the valve will be cut off. The limiting action, therefore, takes place in both directions. Various improvements are possible with this synchronizing separator circuit. First, the contrast control can be arranged to vary the screen voltage of the heptode, which will vary the grid base of the valve as in figure With a small signal input the screen voltage is made low and the grid base is I. / [ / 1 v, - jijpjg A FIG s RELATIONSHIP BETWEEN PULSE AMPLITUDE AND VALVE CHARACTERISTIC B small as at (A). When the signal is as shown at (A) correct operation will take place and most of the noise at the tip of the pulses will be removed because the noise corresponds to the flat portion of the characteristic. If a larger signal is applied as at (B) the noise on the tips of the synchronizing pulses now come into the curved portion of the valve characteristic and will appear in the anode circuit. This is overcome by increasing the valve grid base (by increase of screen grid voltage) to that shown at (B). Obviously, the limiting action on G t of the heptode valve should be varied according to the signal input, i.e. a larger positive voltage should be fed from

37 32 TELEVISION SERVICING the potential divider P (figure 2.16) with a larger signal. This may be obtained by connecting the potential divider from the screen grid to chassis rather than from h.t. positive to chassis. (The voltage will increase with signal input due to movement of the contrast control as explained above). The greater contrast increases the screen grid voltage (increasing the grid 3 base) and increasing the limiting current fed to G x to increase the limiting action. An alternative arrangement is to use a separate noise detector to feed G x of the heptode, which results in better noise elimination and less critical adjustment. Instead of just using amplitude limiting to separate the pulses the idea is to separate them on a frequency basis. Noise pulses are reasonably distributed over the full video band but the synchronizing pulses are limited to a frequency band about 1MHz wide centred on the vision carrier. The required signal output may be obtained by the elimination of signals up to about 0-5MHz on each side of the carrier and amplifying the high frequencies of the vision signal which contain the noise information and the high frequency vision signals. This may be done by a bandpass filter feeding a demodulator (such as a triode anode bend type), the bandpass filter being displaced from the vision carrier so that it picks up the high frequency sidebands (either upper or lower), and not the sidebands corresponding to modulation frequencies below about 0-5 MHz. An alternative is a diode demodulator with a trap circuit designed to reject the carrier and lower frequency sidebands. Since the output of the demodulator does not now contain synchronizing pulses the adjustment of the positive bias on G t is not critical. As well as the noise pulses there will be high frequency vision signals, but these will have no effect because when they occur (between synchronizing pulses) the heptode valve is already cut off on G 3. Valves are still generally used for picture synchronizing separators except in portable television receivers where transistors are essential. The general arrangement is similar to the valve circuit and a typical transistor circuit is shown in figure FIG TRANSISTOR PICTURE SYNCHRONIZING SEPARATOR Bias is provided for the transistor by the potential divider formed by R t and R 2. Resistor R 3 is the normal collector load. The base of the transistor is fed with a positive-going signal through R* and the blocking capacitor C u of large value relative to valve circuits. The signal is d.c. restored by the baseemitter diode in a similar way to the grid-cathode in the valve circuit, as shown If the input signal is sufficient then only the synchronizing pulses in figure will be amplified, the picture portion of the signal driving the base positive,

38 SYNCHRONIZING SEPARATORS 33 so cutting off the transistor. Pulses of collector current then flow corresponding to the synchronizing pulse, resulting in positive-going output pulses. FIG OPERATION OF TRANSISTOR PICTURE SYNCHRONIZING SEPARATOR It should be noted that it is necessary to apply a standing bias to a transistor but not to a valve. This is because increase of grid voltage (negative) on a valve reduces the anode current, whereas increase in base voltage (negative) on a p-n-p transistor causes increase of collector current, no appreciable collector current flowing when the base voltage is zero. A better synchronizing separator may be made using two direct coupled transistors so that clipping takes place in both directions. (2) THE LINE AND FIELD SYNCHRONIZING SEPARATORS Before considering the separation of the pulses it is worth while saying a little about the type of pulse required to correctly synchronize a timebase. For good synchronization the pulse should have a sharp leading edge, the remainder of the pulse not being really important. The reason is shown in figure If the leading edge is curved as at (a) then variations in the pulse Instant of Instant at triggering with triggering with large pulse small pulse._' Triggering^ voltage Instant of triggering with f large and )mol pulses Triggering/ voltage FIG NEED FOR SHARP LEADING EDGE TO THE SYNCHRONIZING PULSE height cause the instant of triggering of the timebase to vary, since the timebase is triggered when the pulse voltage reaches a certain critical value. If the

39 34 TELEVISION SERVICING pulse has a sharp leading edge as at (b) then quite large variations in the magnitude of the pulse can occur, without alteration of the instant of triggering. The line synchronizing separator does not really separate the pulses but makes use of both line and field pulses. Its action is to produce suitably shaped pulses to trigger the line timebase. In the 405-line system the line pulses have a duration of 9fiS while the field pulses have a duration of 4tyiS. In the 625-line system the line pulses are 4-7/iS long and the field pulses 27^S long. In order to make the effect of the two pulses similar, they are fed through what is called a differentiating circuit. A typical circuit is shown in figure It consists of a small series capacitor and a shunt resistor. The c 5-SOpF FIG SIMPLE DIFFERENTIATING CIRCUIT values of the components used vary considerably but typical values are given in the figure. The effect of such a circuit is to give an output when the input voltage changes rapidly, but little or no output when the input is nearly constant. The reason is that, when a sudden change of input voltage occurs, a current flows in R to charge or discharge the capacitor C, and this current causes a drop across R and hence an output. The output soon decreases to zero, after a rapid input voltage change, as the time constant (i.e. C x R) of the circuit is small, which means that C is soon charged. When the input is almost constant little current flows in the circuit and so there is little, or no, output. Actually, the output is approximately proportional to the rate of change of input voltage, hence the name "differentiating circuit". The waveform around the field pulse period of a 405-line transmission is shown in figure 2.24, both before and after the differentiating circuit. It will be seen that at the start of each pulse, a sharp negative pulse is produced in the output (note that the input pulses are negative) and, at the end of each pulse, a sharp positive pulse is produced. The sharp negative pulses are used to synchronize the timebase and it will be seen that there is one corresponding to the end of each line, not only during the period of line pulses, but also during field pulses. Thus, the line timebase is maintained in synchronism during the field pulse period. The negative pulses which occur half way along a line during the period of field pulses, have no effect on the timebase as it is nowhere near ready for triggering (see later, when dealing with timebases). It should be noted that some timebases are synchronized by positive pulses and some by negative pulses. In the above it is assumed that the timebase is synchronized by negative pulses, and that the positive pulses will have no effect. The simple circuit just described is all that is necessary to provide the required pulses for synchronizing the line timebase, so long as the signal is reasonably free from interference. In fringe areas and on 625-line receivers flywheel synchronizing may be used and is considered in Volume 3. In the 625-line system the waveform shown in figure 2.24 will be modified by the presence of equalizing pulses. There is an equalizing pulse corresponding to each line pulse (and one in between) so when these pulses are differentiated, suitable pulses are available to maintain the line timebase in synchronism. The next problem is that of separating the field pulses from the line pulses, or obtaining a suitable pulse from the field pulses to operate the field timebase.

40 'OS,- 1 SYNCHRONIZING SEPARATORS 35,, LINES.. *.? ««' rnjuuuijjjjirnrt -11" U j PULSES \ / ' LINE PULSES FIELD PULSES KWrWHH-f EVEN FIELDS DIFFERENTIATED PULSES iruijjjjljljlnririrt" ODD FIELDS DIFFERENTIATED PULSES FIG Z24. EFFECT OF DIFFERENTIATING THE LINE AND FIELD SYNCHRONIZING PULSES 005-LINE SYSTEM) This is much more difficult and there are also many variations. Although synchronizing the line timebase is important, the accuracy of synchronizing the field timebase is more so if correct interlace is to be obtained, i.e. so that the lines of one field fit centrally between those of the other field. Space does not permit further details but incorrect interlace causes the lines of the picture to be much more visible and decreases the definition. If we assume that the interlace must be correct to 10 per cent. (i.e. that the lines are equally spaced to an accuracy of 10 per cent.) then the maximum error in synchronizing the timebase is l/10th of half a line, late or early. Considering a 405-line system, since a line is IOOjuS long, this means a timing error of 5/u.S. The time of one l/50th second or 20,000/xS and, therefore, the allowable error is: field is 5 20,000 x 100 = 0-025% This is a very small error and shows the high accuracy required in synchronizing the field timebase if correct interlace is to be obtained. In the 625-line system the line time is frtyis and the allowable timing error is 3-2juS. The allowable error is: -^- x 100 = 0016% 20,000 which is even a smaller error. To obtain correct synchronizing it is essential that the timebase be synchronized to the leading edge of the first field pulse, or with a fixed time delay from this leading edge. This time delay must be identical on odd and even fields. It is not possible to synchronize the timebase to the leading edge, since there is no means of distinguishing the field pulses from the line

41 ' 36 TELEVISION SERVICING pulses until a time greater than the length of a line pulse has elapsed. Accordingly, synchronizing must take place with a delay from the start of the first field pulse, but it is extremely important that the time delay be fixed. The purpose of the field synchronizing separator is, therefore, to produce a triggering pulse, which is related to the train of field pulses with this fixed time delay. To do this it is necessary to modify the field pulses, so that they stand out as regards magnitude relative to the line pulses; or to use some method which will detect the difference in the length of the pulses. There are many different ways of separating the pulses and some of the more important will now be described. THE INTEGRATOR CIRCUIT This is the most common circuit used in modern television receivers. The basic circuit is shown in figure 2.25 and the arrangement is the opposite of the differentiating circuit used to feed the line timebase. The time constant {i.e. C x R) is now larger, being (say) 40(iS. Consider that positive pulses on a 405-line system are fed to this circuit as shown in figure 2.26(a). During the INPUT R C g OUTPUT «jinnnnnnituljl FIG FIG EFFECT OF INTEGRATING SIMPLE INTEGRATOR CIRCUIT LINE AND FIELD PULSES period of the pulse a current will flow through R, charging C; between the pulses, C will discharge through R. During the line pulse, C will start to charge but, since the period of the pulse is only 9/x.S, the capacitor will not be charged appreciably and only a small rise of voltage will occur across it. During the long period (91/xS) between pulses, the capacitor discharges through R. During the period of the field pulses, C will charge and, since the duration of the pulse is now 40/xS, it will charge to a much higher voltage. The period between pulses is only lofis and, therefore, C will not discharge appreciably between the field pulses. Thus, the charge on C will accumulate during the eight field pulses and the voltage will rise as shown in figure 2.26(b). At the end of the eighth pulse, C will slowly discharge again. It will be seen that only a small pulse occurs, corresponding to the line pulses but a large pulse occurs corresponding to the eight field pulses. Thus we have raised the field pulse section above that of the line pulses. Although the circuit will operate without a limiter it is desirable to use one, to remove at least the small pulses caused by the line pulses. A circuit is shown in figure Here V t is the picture synchronizing separator which produces negative pulses at the anode. These are integrated by R t and C x, resulting in a large negative pulse during the field pulse period. Limiting is achieved by MR t. The right-hand side of MR t is connected to the potential divider formed by,r 2 and J? 3, which has such values that MR t is maintained non-conducting until the voltage across C x drops below a certain value, such as shown dotted in the figure. Thus, the small pulses due to the line pulses are removed. After passing through MR X the field pulse is fed to the field timebase, through the blocking capacitor C 2. The pulses are taken from the anode of V u through the 6pF capacitor which, together with the input resistance of the timebase form a differentiating circuit to the line timebase. It is interesting to note the waveforms produced by the integrator circuit are not identical on odd and even fields in the 405-line system. Figure 2.28(a)

42 SYNCHRONIZING SEPARATORS 37 \ T~lTTJUL' h "l 47 K.LimKcr < / tting < R 2 SlSOK ^ Ol 47 K To field timebase SOOpF."a 3SOK 6pF FIG ^ T<\ line tfmebote INTEGRATOR AND LIMITER (a) Even fields (b) Odd fields (c) Odd and even fields (50Hz timebase) FIG PHOTOGRAPHS SHOWING THE EFFECT OF INTEGRATING LINE AND FIELD PULSES (405-LINE SYSTEM) shows the result on even fields and (b) that on odd fields, while (c) shows the two superimposed, by running the timebase of the cathode ray oscillograph at 50Hz. A difference occurs at the start due to the fact that on even fields there is a period of 90/x.S between the last line pulse and the first field pulse, while on odd fields there is a period of only 40ju.S. Thus, the initial charge on C is different in the two cases, which may cause errors in triggering. If the triggering voltage of the timebase is CD, then the difference will be small but, if it corresponds to line AB, then the difference is appreciable. The latter will result in poor interlace, as the time delay from the start of the first field pulse to the instant of triggering is different on odd and even fields. There is also a difference at the end of the pulses which may cause poor interlace. Although the simplest type of limiter shown in figure 2.27 is satisfactory, better results can be obtained by a double limiter which takes a slice out of the waveform. This largely overcomes the differences at the start of the pulse, particularly if the slice is taken at the right place, and also reduces the differences at the end of the pulse. By using a narrow slice and amplifying the result, a pulse is obtained with a sharp leading edge. It will be seen that

43 38 TELEVISION SERVICING the waveforms of figure 2.28 do not have a sharp leading edge, and this reduces the accuracy of synchronizing. A circuit using this principle of taking a slice out of the waveform is shown in figure It shows the picture synchronizing separator and field inruwj FIG INTEGRATOR AND DOUBLE LIMITER (Milliard) synchronizing separator. Valve Vi acts as the picture synchronizing separator, using a choke, L lt in the anode circuit as described earlier. The low screen voltage is obtained by the high value of screen resistor. The output from V x is fed to the line timebase through C 3, which differentiates the pulses. The other end of L x is connected to R 4, R 5 and C 4, which form the integrating circuit, and during the field period a negative pulse is produced across C 4 (since the pulses from V t are of negative polarity). The valve V 2 acts as a double limiter. By returning the grid to h.t.+ through R t and R 9 a steady current flows through R 9 (and, of course, R 7) causing a drop across it. Unless the drop in voltage across C 4 exceeds the voltage across R 9 it will make little difference to the voltage on the grid of the valve V 2 since, provided grid current is flowing, the grid will remain about zero potential. This is the condition which applies during the line pulses. During the field pulses the decrease in voltage across C 4 exceeds the drop across Rg, and the grid of V 2 is driven negative to a value sufficient to cut the valve off. Thus, limiting occurs in the opposite direction due to valve cut-off. Hence, V 2 is cut off for a period corresponding approximately to the eight field pulses and a positive pulse is, therefore, produced at the anode and used to synchronize the field timebase. The pulse obtained is almost square and has a reasonably sharp leading edge. Several modifications of the simple integrator are possible and one arrangement is shown in figure 2.30 where the line pulses are developed across R 2 and fed to the line timebase, while the integrated field pulses are developed across the integrating circuit iji-cj. The valve would generally also act as the picture synchronizing separator valve. The arrangement is simple but has the disadvantage that it may cause interaction between the two timebases. Many line timebases produce a large pulse on flyback, and this "back-wash" may be fed to the field timebase and upset the interlace. One method of reducing the effect is shown in figure Here, the line synchronizing pulses are obtained from the anode circuit and the pulses are integrated by the capacitor on the screen circuit (which is, of course, smaller in value than normal) and the pulse thus obtained is fed to the field timebase.

44 SYNCHRONIZING SEPARATORS 39 FIG SIMPLE INTEGRATOR CIRCUIT IN ANODE OF PENTODE FIG INTEGRATOR IN THE SCREEN GRID CIRCUIT OF PENTODE The integrator circuit is a commonly used circuit because it is simple and is very free from trouble due to noise. This is due to the fact that any sharp noise pulses which may occur are integrated by the circuit and, like the short line pulses, are largely removed. Its disadvantages are that it does not produce a very sharp leading edge, and that the pulses on a 405-line system are not quite identical on odd and even fields, which may result in poor interlace. In a 625-line system the general operation of the circuit is similar, but because equalizing pulses are used the integrated field pulses are identical on odd and even fields. If figure 2.32 is examined carefully it will be seen that the pulse waveform input to the integrator is identical on odd and even fields from the start of the first equalizing pulse to the end of the last one. Thus when the waveforms are integrated they produce identical pulses. It will also be seen that at the end of even fields there is a half line between the start of the last line synchronizing pulse and the start of the first equalizing pulse, whereas at the end of odd fields there is a whole line between the start of the last line pulse and start of the first equalizing pulse. This difference causes a slight variation in the integrated waveform at the start of the equalizing pulses, but by the time the field pulses start it becomes negligible. The slight differences at the start (and at the end) of the equalizing pulses have no effect on the synchronizing of either the line or field timebases. The effect of integration is shown in figure 2.33, and it will be seen that the pulses are identical during the critical period of the field pulses, which is the object of equalizing pulses. This circuit is commonly used.

45 TELEVISION SERVICING LINE PULSES FIELD PULSES EQUALIZING PULSES, I 2 r t tt r t UJ END OF EVEN FIELD START OF ODD FIELD 3 < T,, 5 6,,. 7 COMPOSITE PULSES LINE PULSES FIELD PULSES TT1 EQUALIZING ITTTT PULSES h---+ 3IO, Tt 1, , 319 I" 1 COMPOSITE PULSES END OF ODD FIELD START OF EVEN FIELD FIG LINE WAVEFORMS

46 SYNCHRONIZING SEPARATORS 41 (a) Even fields (b) Odd fields (c) FIG Combined odd and even fields (negative-going pulses) INTEGRATED WAVEFORM AROUND THE FIELD PULSE REGION ON A 625-LINE SYSTEM THE PARTIAL DIFFERENTIATING CIRCUIT This type of synchronizing separator makes use of the differentiating circuit shown in figure 2.34(a) which consists of a series capacitor and shunt 7 LINE PULSES DCy.i component JUUUULLi FIELD PULSE S co psi«nt FFT LINE PULSES (1) UUUUL FIELD PULSES (e) FIG EFFECT OF D.C. COMPONENT ON LINE AND FIELD PULSES resistor the opposite of the integrator circuit used in the last type of synchronizing separator. The time constant is usually made about 40/aS (i.e. a C.R product of 40/xS. Note that if C is in microfarads and R is in ohms the result is in microseconds). During the period of line pulses (negative pulses) as shown at (b) the d.c. component of the signal is large but, during the field pulses as shown at (c), the d.c. component is small. Since the capacitor C cannot pass direct current there cannot be any d.c. component in the output from the circuit, hence there is a change in the position of the pulses relative to the zero. During the line pulses the waveform will settle down as

47 42 TELEVISION SERVICING shown at (d) and, during the field pulses, it will tend to settle down to that shown at (e). The change will not be instantaneous, as C takes an appreciable time to charge. Oscillograms (using a 405-line system) taken around the field pulse period are shown in figure 2.35, where the change of level of the pulses can be seen during the period of the longer field pulses. The result on even fields is shown at (a) and that on odd fields at (b); it will be seen that there are slight differences. This is shown more clearly at (c) where the two waveforms are superimposed, by running the timebase at 50Hz. The difference is caused (as in the integrating circuit) by the differences in the time between the last line pulse and the first field pulse on odd and even fields. It is useful to note that the waveform from the differentiating circuit is complementary to that from the integrating circuit, the two added together being the input waveform. From figure 2.35 it will be seen that there are large pulses produced during the line pulse period and it is necessary to remove these by a suitable limiter, (a) Even fields (b) Odd fields (c) Odd and even fields (50Hz timebase) FIG EFFECT OF PARTIAL DIFFERENTIATION OF LINE AND FIELD PULSES ( 405-LINE SYSTEM) i.e. to remove all below a line such as AB. A simple limiter circuit is shown in figure 2.36, where d and R l form the differentiating circuit. Suitable bias is provided by the potential divider R 2, R 3 and R 4. During line pulses the grid is not driven in a positive direction sufficiently to cause the valve to conduct, but during field pulses the waveform becomes more positive and anode current flows. This results in a series of negative pulses (corresponding to the period between field pulses) during the field pulse period. A picture synchronizing separator, with differentiator circuit and diode limiter are shown in figure Valve V l acts as the picture synchronizing separator, low screen voltage being produced by the high screen resistor. Ci-J?i form the differentiator circuit. V 2 is biased by R 2 and R 3 so that it only conducts during the field pulse period. It is important to note that negative-going pulses fed to the differentiator circuit produce positive pulses

48 SYNCHRONIZING SEPARATORS 43 IT-ITTJULL / INPUT -» It ~innnr inr-i^- FIG PARTIAL DIFFERENTIATING CIRCUIT AND LIMITER woopf To field timebase T-t-lJ^1 FIG PICTURE SYNCHRONIZING SEPARATOR, PARTIAL DIFFERENTIATING CIRCUIT AND DIODE LIMITER for the field synchronizing (actually these pulses are the spaces between the field pulses). It will be seen that a differentiating circuit type of synchronizing separator has two desirable features: (i) the triggering pulse has a sharp leading edge and 07) the triggering pulses have a fixed time delay from the leading edge of the first field pulse, since they correspond directly to the back edges of the field synchronizing pulses. From these two features it would be thought that interlace should be excellent, but this is not necessarily the case, because the waveforms on odd and even fields are not identical on a 405-line system. On odd fields the pulses are slightly higher at the start, and the first field pulse passing the limiter is slightly greater. The trailing ends of the train of pulses are also quite different. This may be thought unimportant but in fact it may be vitally important and upset the interlace. Space does not permit a detailed explanation of these effects and the reader should consult the author's publication Television Society Journal 7, 1955, pp. 350 and 428, /. Brit, I.R.E., 14, 1954 p. 191, if he wishes to obtain more detailed information on this subject. With a 625-line system the field pulses are identical on odd and even fields. Slight differences do occur at the start and end of the equalizing pulses, but

49 44 TELEVISION SERVICING (a) Even fields (b Odd fields (c) FIG Combination of odd and even fields (negative-going pulses) DIFFERENTIATED WAVEFORM AROUND THE FIELD PULSE REGION ON A 625-LINE SYSTEM they are of no importance, as is shown in figure An interesting differentiator circuit and limiter is shown in figure o-tirtma V TinJUUL i -a s^:,orirt ujyu1' FIG PARTIAL DIFFERENTIATING CIRCUIT WITH AUTOMATIC LIMITER In place of the fixed bias, used earlier, the bias is obtained automatically. C t and Ri form the normal differentiating circuit and bias is provided by C 2 and R 2. The time constant of C 2 -R 2 is large, say 0-2 second. C 2 is charged in the direction shown by the positive pulses occurring during the field pulse period (after differentiating) in a similar way to the self-bias of an oscillator. The whole waveform hangs below the zero line by the tips of these pulses. So long as the applied signal is sufficiently large the waveform corresponding to the line pulses will be at a negative potential greater than the cut-off of the valve, and so will not appear in the output. The time constant of C 2 -R 2 is sufficiently long so that no appreciable discharge takes place between the trains of field pulses, which occur every l/50th second. INTERLACE FILTER The circuit of this type of field synchronizing separator is shown in figure The circuit is fed with negative-going pulses. Between pulses C t is charged through K 2 and, during the period of the pulses, C x discharges through R t. During the short line pulses (9/xS) (405-line system) the discharge

50 SYNCHRONIZING SEPARATORS 45 OUTPUT FIG INTERLACE FILTER is small and only a small drop in voltage occurs across C ± but, on the longer field pulses (40/iS), C x discharges further and larger pulses are produced. Valve V 2 acts as a limiter cutting off the small pulses occurring during line pulses. The anode voltage of V 2 is settled by R 2 and R 3 and it is arranged that V 2 only conducts when the voltage across C x drops, on field pulses, below the anode voltage of V 2. The circuit is simple and produces identical pulses on odd and even fields and in general the interlace is good, even on a 405-line system. PULSE WIDTH DISCRIMINATOR Another type of circuit is shown in figure 2.41 where V y is the normal synchronizing separator. The resulting negative pulses are fed to the grid of valve V 2a so that this valve is cut off during pulses but is heavily conducting between pulses, due to the grid resistor being returned to h.t. +. Between pulses Q is discharged through V 2a but, on pulses, C x charges through R t. The voltage rise is small on the short line pulses but approximately four times as great on the longer field pulses. This produces the waveform shown at the anode of V 2a. V 2b acts as a limiter removing the small pulses, bias being provided by R 2 and R 3. The result is negative pulses on the anode of V 2b. The sawtooth waveform applied to the grid of V^ is clipped by the action of grid current and so produces pulses, as shown, in the anode circuit. This circuit also produces identical pulses on odd and even fields on both 405 and 625-line systems and gives satisfactory interlace. Although there are many other possible arrangements they are not commonly used in modern receivers. For further details see references on page 43. X7ULAJL FIG FIELD SYNCHRONIZING SEPARATOR

51 We CHAPTER 3 TIMEBASES have seen the need to deflect the beam of the cathode ray tube in two directions at right angles, this being done by the use of two electromagnetic fields produced by two sets of coils arranged at right angles to each other on the neck of the tube. In order to produce these magnetic fields we have to pass sawtooth currents through the coils, and the device for producing these currents is called the timebase. There are two general ways of producing these currents as shown in figure 3.1. At (a) is a device designed to produce a suitable sawtooth or pulse VOLTAGE TIME- BASE CURRENT TIME- BASi CURRENT OUTPUT STAGE TO TO DEFLECTING "" *" COILS DEFLECTING COIlS FIG ARRANGEMENTS FOR FEEDING THE DEFLECTING COILS (a) Voltage timebase and output stage; (b) Current generator or timebase voltage output and this will be termed a voltage timebase. The output is of high impedance and is used to feed a current output stage which converts its voltage input into a corresponding current in the deflecting coils. At (b) the sawtooth current waveform is produced directly by a sawtooth current timebase or current generator. Both systems are used in modern television receivers and the voltage timebase will be described in this chapter. The current output stage, which is fed from the voltage timebase, is described in Chapters 4 and 5 and the current generator in Chapter 6. VOLTAGE TIMEBASE All voltage timebases depend on the charge or discharge of a capacitor to produce the required voltage output, and the principle is shown in figure 3.2(a). When the switch S is open the capacitor C will charge through the resistor R, at a rate which depends on the value of C and R. If allowed to, C would eventually charge to the full h.t. voltage. The way in which it would charge is shown in figure 3.2(b) and is known as an exponential rise of voltage. If, at some instant X, the switch S is closed then, assuming R t much R,n Mr i t FIG PRINCIPLE OF TIMEBASE (a) Circuit (b) Waveform 46

52 TIMEBASES 47 smaller than R, C would be rapidly discharged to zero as shown. If the switch is opened again then C would recharge relatively slowly, as at the start. Thus, by closing 5 for short periods (at X and Y, etc.) we would obtain the waveform shown in figure 3.2(b) which is sawtooth in shape, as required. Ideally, the rise of voltage from A to X should be linear {i.e. a straight line) but, as can be seen, it is curved. By arranging that the voltage at is only a small proportion of the total voltage the curve between A and X is almost straight. This, of course, means a smaller sawtooth voltage output. With this simple system of R and C there is nothing we can do about this we may obtain a large output voltage but always at the expense of nonlinearity in the scan. The switch 5 must, of course, be replaced by some electronic device. For the sawtooth waveform to be synchronized, the flyback must correspond with the synchronizing pulses, i.e. the switch S must be closed at the start of each synchronizing pulse. Although timebases can easily be constructed which are driven directly by the synchronizing pulses this type is not used because the timebase output and the deflection would fail if the synchronizing pulses failed due to the end of the programme or failure of the signal for any reason. The result would be a stationary spot which would burn the screen. For this reason a self-running timebase is used which always produces an output of approximately the correct frequency but which is kept in step or synchronism by the synchronizing pulses. This will be explained later in more detail. The simplest timebase to understand (but no longer used) is the thyratron timebase but we must first say a little about the properties of a thyratron valve. This valve is a triode into the bulb of which has been introduced a small quantity (0-2mm of mercury pressure) of inert gas. This gas alters the characteristics from that of a high vacuum triode because the gas becomes ionized, i.e. split up into positive ions and electrons. This is due to the molecules of gas being bombarded with electrons emitted from the cathode. Suppose that we maintain the grid at a fixed negative voltage and increase the anode voltage from zero. Up to a certain anode voltage, known as the critical anode voltage, we should find that the anode current was negligible. When the critical anode voltage is reached it is found that the anode current (due to ionization occurring) suddenly increases to a large value (which must be limited by a resistor in the anode circuit) and the voltage between anode and cathode would drop to about 10V. This is shown in figure 3.3. The reason Io \. CRITICAL ANODE VOLTAGE _/ FIG CHARACTERISTIC OF THYRATRON VALVE why the current must be limited is that if the anode current increased to the saturation value of the cathode, the voltage drop would increase. (Note that the valve will not limit the current as would a high vacuum valve because the I a -V a characteristic is approximately vertical and the anode voltage drop does not increase with anode current until saturation of the cathode occurs). This causes the positive ions (which are attracted towards the negative cathode) to bombard the cathode with sufficient energy to remove the oxide coating from

53 48 TELEVISION SERVICING the cathode. This is known as stripping of the cathode and care must be taken with all gas-filled valves to avoid it. Once anode current flows the grid has no control over the flow of anode current and cannot cut off the anode current as in the case of a normal high vacuum triode. The anode current can only be stopped by removal of the anode voltage. On reapplication of the anode voltage the grid regains control, so long as it is sufficiently negative. If the grid voltage is increased (in a negative direction) the effect is to increase the critical anode voltage and the critical anode voltage is approximately proportional to the grid voltage. Suppose that a thyratron is connected as shown in figure 3.4(a). This is H.T. VOLTAGE VOLTAGE across c HW CRITICAL ANODE VOLTAGE FIG PRINCIPLE OF THYRATRON TIMEBASE (a) Circuit; (b) Waveform identical to that of figure 3.2(a) except that the thyratron valve Know replaces the switch S. Assume that V is supplied with a negative voltage on its grid such that its critical anode voltage is about l/3rd of the high tension voltage. On switching on the supply, C will charge through R as shown in figure 3.4(b). When the voltage across C reaches the critical anode voltage of V (at X) the valve will pass a large current (mainly limited by the resistor R^> and rapidly discharge C. The valve V will go out at Y as there is not sufficient current to maintain the discharge. Thus, C will start to charge again and the process is repeated. The timebase will continue to run in this manner so long as the supply is maintained. Let us now consider what factors affect the frequency and amplitude of the output voltage of the circuit. The rate at which C charges through R is determined by the values of C and R. The time to charge C (to a particular voltage) is proportional to the C.R product. It will charge to 63-3 per cent, of the h.t. voltage in a time equal to C.R seconds (C being in farads and R in ohms). The effect of increasing the product C.R is shown in figure 3.5(a). It will be seen that the effect is to alter the frequency but (assuming the grid voltage to be constant) not to alter the amplitude or linearity. If the grid voltage is increased the critical anode voltage is increased and C must charge to a higher voltage before it is discharged by V. This makes no difference to the voltage at which the valve ceases to conduct. There are three effects as shown in figure 3.5(b): (i) The amplitude of the sawtooth waveform is increased. (if) The frequency is reduced since it takes longer to charge C to the higher voltage. The frequency can easily be brought back to the original value by decrease of C or R. (in) The linearity is not as good since a larger proportion of the exponential curve is now being used. The values of C and R are adjusted so that the timebase runs at approximately the correct frequency, but arrangements must be made to lock the timebase in step as it is called. with the synchronizing pulses or synchronize the timebase Let us now see how this is achieved in the case of the thyratron

54 TIMEBASES 49 INCREASED NEGATIVE GRID VOLTAGE FIG. 3.5(a). EFFECT OF INCREASING THE TIME CONSTANT (C.R.) OF THE TIMEBASE CIRCUIT FIG. 3.5(b). EFFECT OF INCREASING THE BIAS OF THE THYRATRON VALVE timebase, although the principle is the same whatever the type of timebase. The idea is to inject positive synchronizing pulses into the grid circuit of the thyratron so as to trigger or fire the thyratron just before this would normally occur. If the grid voltage were fixed the critical anode voltage could be represented by the straight line PQ of figure 3.6. The timebase would then FIG PRINCIPLE OF SYNCHRONIZING run as shown dotted, the flyback occurring at A, B, etc. By injecting small positive pulses into the grid circuit the critical anode voltage is reduced during the period of the pulse (since the grid voltage is reduced during this period) as shown in the figure. The flyback will now occur where the charging curve cuts the critical anode voltage line, i.e. at points X and Y, etc. These points correspond to the leading edges of the synchronizing pulses and so the timebase is synchronized to the pulses. It will be noted that this will only occur if the timebase frequency is approximately correct. For correct operation the frequency of the timebase should be slightly low, so that the pulses trigger the timebase just before the instant at which it would trigger in the absence of pulses. In other words, the timebase is almost ready to trigger and the effect of the synchronizing pulses is to start the triggering action. This makes the start of each flyback correspond to the leading edge of the synchronizing pulse. It will be noted that the remainder of the pulse has no effect on the synchronizing and it is only the leading edge which is important. (In certain cases this may not be true but space prevents giving further details). The polarity of the synchronizing pulse is important and, in the above case, it will only synchronize with a positive pulse. The polarity of the pulse must be such that it will start the triggering a little earlier than it would do without the pulse. Some timebases are synchronized with positive pulses and some with negative pulses and it is important that the correct polarity of pulse is used. A typical thyratron timebase circuit is shown in figure 3.7, the values being given for both field and line timebases. As previously, C is charged through R and discharged by the valve V, through the limiting resistor R^.

55 50 TELEVISION SERVICING POSITIVE SYNCHRONIZING PULSES / _n_n_ R = kQ; Ri=100Q; R2 = 1-10kn; R3 =47kQ; R4 = 10kn. LINE: C=5,000pF; Ci=0-5/iF. FIELD: C=0-5/iF; Ci=50jiF. FIG PRACTICAL THYRATRON TIMEBASE CIRCUIT This discharge current also flows through R 2 and produces the required bias for V (similar to the normal cathode bias used in amplifiers). The voltage across R 2 is maintained approximately constant over the cycle by the large capacitor C x. R 3 is the grid resistor, and ij 4 is a grid limiting resistor which is necessary to prevent excessive grid current should the grid be driven positive. Part of R is made variable to act as an amplitude control. It would be expected that this would act as a frequency control in the same way as the previous circuits, but in this circuit a complication occurs, owing to the fact that the current through R also provides the bias, since the same average current must flow in R 2. Thus, when R is reduced the current is increased, which charges C more rapidly but the larger current in R 2 increases the bias. The two effects approximately cancel out as regards frequency and only a change of amplitude occurs. In other words, if R is halved it will charge up C twice as quickly but, since it has to charge it up to twice the voltages before flyback occurs, the total time will be the same. The frequency control (or hold control as it is often called since it is normally adjusted until the timebase locks or synchronizes) is R 2. This varies the bias, hence the triggering voltage and, therefore, the frequency. It will also vary the amplitude but this can be corrected by variation of R if necessary. In practice only slight changes of frequency are required. A thyratron timebase is not used to-day in television receivers but, in the author's opinion, it is a very useful and simple timebase. It is often considered to be unreliable but, if properly designed, it is most reliable; the author has used thyratrons in television receivers and oscilloscopes with no trouble whatever. It is important that Ri and R t are not omitted and it is desirable not to apply the h.t. until the valve cathode has reached its correct operating temperature. Having described the operation of the thyratron timebase we must consider other timebase circuits which use hard (i.e. high vacuum) valves and transistors. The general principle remains the same but some arrangement must be introduced to bring about the cumulative action, so that once anode current or collector current starts to flow in the valve or transistor it rapidly builds up to a large value. This occurs in a thyratron due to the action of ionization, since more anode current causes more ionization of the gas and this, in turn, produces more electrons and so more anode current. When a high vacuum valve or transistor is used no such action occurs unless some positive feedback is introduced between the anode and grid circuits or collector and base (or emitter) circuits. Thus, in the case of a valve, when anode current starts to flow and the anode voltage therefore falls, we must arrange that the grid is driven in a positive direction so as to still further increase the anode current. Valve circuits will first be considered.

56 TIMEBASES 51 FEEOMCK no' FIG PRINCIPLE OF HARD VALVE TIMEBASE The basic circuit is shown in figure 3.8 when again C and R are the charging capacitor and resistor. V is now suitably biased so that current does not flow until the anode voltage reaches a certain value (corresponding to the critical anode voltage in the case of the thyratron). As soon as current flows in the anode circuit, the anode voltage falls and this drop in voltage is fed to the grid in the opposite direction (i.e. opposite phase or with a phase shift of 180 ) so driving the grid in a positive direction, and increasing the anode current. A cumulative action occurs so that a heavy anode current flows and C is rapidly discharged. V then becomes non-conducting and C recharges again through R. A feature of all hard valve timebases is the feedback with 180 phase shift between the anode and the grid circuits. This may be achieved by using a transformer, as in the blocking oscillator timebase, or by using another valve, as in the multivibrator timebase. The feedback can also be achieved by transitron action between anode and screen circuits of a pentode valve and used in the transitron and transitron Miller integrator timebases. These are not often used in modern television receivers and they, also, will not be considered in this book. The two most common timebases for television are the blocking oscillator and multivibrator types and these will be considered in some detail. THE BLOCKING OSCILLATOR TYPE OF TIMEBASE This is most easily explained by considering the action of a "squegging" oscillator. The circuit of a tuned anode oscillator is shown in figure 3.9, FIG TUNED ANODE OSCILLATOR CIRCUIT L 2 -C 2 forming the tuned circuit. L x is the grid coupling coil and R r C t provide the self-bias. If Ri-C^ are of suitable value for the oscillator a steady bias is provided across C lt almost equal to the peak voltage across Z, t. The valve acts as a diode (grid and cathode) and charges C t when the grid is driven in a positive direction. Should the h.t. voltage change, or any other change occur in the oscillator, then the bias automaticelly readjusts itself to an equilibrium condition. The equilibrium condition is such that the power fed into the anode circuit just equals the losses in the circuit. If excessive feedback (large mutual inductance between grid and anode coils) and/or large time constant is used for C r R t "squegging" will occur. In this case the amplitude of oscillation builds up rapidly as shown in figure 3.10, hence

57 I 52 TELEVISION SERVICING VALVE FIG OPERATION OF SQUEGGING OSCILLATOR the bias increases rapidly, in fact so rapidly that the bias becomes too large. The result is that the energy fed back to the tuned circuit is less than the losses in this circuit, and so the amplitude of oscillation decreases. Owing to the large time constant in the grid circuit (C^Rj) the bias does not decrease as rapidly and so less energy is fed back into the anode circuit since the grid is not fully driven (as at X). Thus, the oscillations rapidly diminish and the valve is left cut off by the charge on C u as at Y. Nothing will now happen until the charge on C, (i.e. the bias) decreases to the cut-off point of the valve as at Z. At this point oscillations rapidly build up again and the same action is repeated. Thus, the oscillator only operates in short bursts or "squeggs". The frequency of oscillation during the short bursts is settled by the L.C value of the anode circuit as in a normal oscillator. The frequency of the bursts or "squeggs" is determined mainly by the time constant C,..^ in the grid circuit. The voltage across C t is, of course, the bias voltage and it is seen from figure 3.10 that this is approximately sawtooth in shape as is required for a timebase. In the circuit of figure 3.9 both sides of Q are alive but the circuit is soon modified (as shown in figure 3.11) so that one side of C x is at chassis potential. The only change is in the order of L x and C u which makes no difference to the operation of the circuit. The circuit may be used in this way as a timebase known as a squegging oscillator timebase, the oscillation usually occurring at a radio frequency, say 1MHz. This method of operation FIG BLOCKING OSCILLATOR TIMEBASE

58 TIMEBASES 53 is not used to-day, doubtless due to the difficulties of preventing radio interference. Instead, a modified form known as a blocking oscillator timebase is used. This operates fundamentally in the same manner except that extremely excessive coupling is now used and the damping of the tuned circuit is such that only about one cycle of oscillation occurs at each "squegg". The operation is shown in figure The coils L x and L 2 now take the FIG OPERATION OF BLOCKING OSCILLATOR TIMEBASE form of a transformer with an iron core (usually Mumetal) and C 2 consists on only the winding and stray capacities of the circuit. Damping is usually provided by the core losses in the iron core but an additional damping resistor may be placed across the primary or the secondary. The natural frequency of oscillation should be much higher than the timebase frequency so that the flyback time is short. One may also consider the operation in the following way. As soon as anode current starts to flow as at Z (figure 3.12) a voltage is induced in L x tending to drive the grid positive, so increasing the anode current still further. Grid current flows and charges d- Eventually, the anode current cannot increase further (due to the limiting current of the valve) and the grid voltage induced in L x decreases (remember that the voltage induced in L t is proportional to the rate of change of anode current in L 2 and does not depend on the actual value of the anode current). The anode current diminishes and a cumulative action takes place, the decreasing anode current reversing the voltage in L t and so cutting off the valve. A very brief oscillation occurs (due to the heavy damping) and the valve remains cut off due to the charge on d. The capacitor d now discharges through R t until anode current flows again when the action is repeated. This differs from the thyratron is charged rapidly during flyback and discharged slowly circuit in that d (through Ri) during the scan.

59 54 TELEVISION SERVICING It will be seen from figure 3.12 that the resulting voltage is rather nonlinear since a large fraction of the exponential curve is being used. C, would, of course, discharge to zero voltage if no oscillations took place. The output voltage may be made much more linear by returning /J, to h.t. positive instead of to the chassis. The effect is shown in figure Suppose C, is charged to -100V and that the valve cut-off voltage is -20V. In the case R, returned to chassis FIG EFFECT OF CONNECTING GRID RESISTOR TO H.T. POSITIVE where R x is returned to chassis then 80 out of 100 volts, or 80 per cent, of the exponential curve, is being used and the output is far from linear. If if, is returned to h.t. positive (say +200V) then C x would charge to +200V if no oscillations occurred. Hence, the exponential curve now goes from 100V to +200V, or is 300V. We are only using 80 volts of it, or 80/300 = 27 per cent, of the curve, and so the output is much more linear. Since d will now discharge to 20V much more quickly, the value of R x or C x will have to be altered to maintain the same timebase frequency, but this is of no consequence. The frequency of the timebase is varied by varying either R x or by taking Ri to a potential divider across the h.t. supply arranged to vary the voltage fed to R lm The amplitude can be varied, within limits, by varying the h.t. voltage to the timebase by means of a variable series dropping resistor suitably by-passed to chassis. To synchronize a blocking oscillator timebase it is necessary to feed a signal to it such that it will start the flyback. This may be done by feeding a positive pulse to the grid so as to take the grid voltage over the cut-off line (see figure 3.12) at a point just before Z. Alternatively, the timebase may be synchronized by a negative pulse applied to the anode. This is converted to a positive pulse on the grid by the action of the transformer. It should be noted that, unlike a thyratron timebase, a large voltage exists on both the grid and anode during the flyback and that this voltage will be fed to the synchronizing separator circuit. Also, the connection of the synchronizing separator to the grid, or anode, must not prevent the timebase operating. A typical example of a field blocking oscillator timebase is shown in figure The timebase is fed through R x (47kO) and R 2 (looko), the latter being variable and forming the height control. These are by-passed to chassis by C 3 (8/xF). The charging capacitor is C 2 (0-ljixF) which is charged through R 5 (470k 2) from the potential divider formed by R 3 (22kfi) and R 4 (looko), the latter forming the hold control. It will be seen that variations of the height control cause variations of the voltage fed to the potential divider R3-R t. In this way an increase in height causes an increased voltage to be fed to the charging resistor R 5. Thus, the capacitor charges up more rapidly but over a larger voltage range and the frequency therefore remains approximately

60 TIMEBASES 55 /l»- FLYBACK " BLACKOUT FIG TYPICAL BLOCKING OSCILLATOR TIMEBASE CIRCUIT constant. The transformer T is damped by R 7 (looko). The timebase is synchronized by feeding negative pulses to the anode through C x (l.ooopf). By placing a resistor R 6 (4-7k i) in series with C 2 a negative pulse is obtained during flyback, and this is used to black out the screen during the flyback and so prevent the appearance of field flyback lines. An alternative method of operating a blocking oscillator timebase is shown in figure The circuit operates as previously, the frequency being FIG ALTERNATIVE BLOCKING OSCILLATOR TIMEBASE CIRCUIT mainly determined by the time constant Cj.-Ri in the grid circuit. The voltage across C, is no longer used for the output. Instead, the circuit is fed from across the charging capacitor C, which is fed from the resistor R. During the period when the timebase valve V is cut off, C charges through R and, when the valve conducts, a short pulse of anode current flows which discharges C. Hence, a sawtooth voltage is obtained across C in a similar way to the thyratron timebase, the valve acting as a switch and momentarily shorting C when it conducts. The amplitude can now be controlled by variation of R. If R is reduced, C will charge to a higher voltage between the "squeggs", Variation of R has little effect on the and so a larger output will be obtained. frequency, this being controlled by C t -Ru R t usually being variable to act as a frequency or hold control. If a large amplitude of output is obtained the linearity will, of course, be poor, as in the thyratron circuit. Many variations of this type of circuit are possible. In one arrangement a pentode is used in which the operation of the blocking oscillator and the discharge circuit are separated by placing L 2 in the screen circuit and connecting the anode to C (figure 3.15). When screen current flows so does anode current, hence C is rapidly discharged. Another arrangement is the use of a double triode with the grids joined together. L 2 is placed in one anode circuit and the other anode is used to discharge C. Blocking oscillators are used in television receivers for both line and field timebases. It is a simple circuit and uses a single triode valve or often the

61 56 TELEVISION SERVICING triode section of a multiple valve. The most likely cause of failure to operate is a fault in the blocking oscillator transformer. Large voltages are induced in this transformer which, unfortunately, is not as reliable as it should be. The usual fault is shorted turns and the only remedy is a replacement. Different designs of transformers are used for line and field, the line transformer having a smaller core and fewer turns, so as to maintain a short flyback time. MULTIVIBRATOR TIMEBASE This is a two-valve type of timebase in which the second valve serves the same purpose as the transformer of the blocking oscillator, i.e. to produce a 180 phase shift, for the required positive feedback during flyback. The circuit is given in figure 3.16 where R is the charging resistor and C is the FIG BASIC MULTIVIBRATOR TIMEBASE CIRCUIT charging capacitor. Like all timebases it is difficult to know where to start, but let us assume that V2 is cut off, due to a charge on the capacitor C 2. C will charge through R during this period and the voltage across C will rise, At the same time the capacitor C 2 will discharge through R 2 and R 3, and eventually valve V 2 will start to conduct. Immediately this happens its anode voltage will drop. This sudden drop of voltage at the anode of V 2 is communicated to the grid of V u through C l (remember that the voltage across Ci cannot change instantaneously). This tends to cut off K 2 and cause its anode voltage to rise. This, in turn, is communicated to the grid of V 2, causing the grid to go in a positive direction. This cumulative action takes place very rapidly so that V x is cut off and V 2 is made conducting, thus discharging C and producing the flyback. The negative voltage cutting off Vi does not remain as C t discharges through R and R^ This time constant is made much smaller than that of the grid of V 2 so that a short flyback is obtained. As soon as K, starts to conduct, its anode voltage drops, the grid of V 2 is driven negative, V 2 is cut off and this, in turn, causes the anode (and so the grid of VJ to go in a positive direction. Thus, V 2 is conducting and V x cut off during flyback; V 2 is cut off and V t is conducting during the scan. The frequency can be varied by variation of the resistor R 2, or R 2 may be taken to a variable potential as in the blocking oscillator timebase. The time base can be triggered by a positive voltage fed to the grid of V 2, since this is non-conducting during scan. More commonly, a negative pulse is fed to the grid of V u which is converted to a positive pulse at the anode of V t and fed to the grid of V 2, thus making V 2 conduct and start the cumulative action. "u \r fc I t\ 1 \. T" 1 FIG TYPICAL MULTIVIBRATOR TIMEBASE CIRCUIT

62 TIMEBASES 57 A typical line timebase circuit is shown in figure R (33kQ) and C (500pF) form the charging resistor and capacitor. The anode voltage of V 2 is fed to the grid of V 1 through C 2 (6pF) with grid resistors R 3 (33kO) and fl 4 (68kfi). The anode load of V t is R x (15kO) and this is coupled to V 2 through C 3 (140pF). The grid resistor of V 2 is R 5 (1-8MO) which is fed from the potential divider formed by R 2 (100k i) and R 6 (looko), the former acting as the hold control. The potential divider is by-passed by C 4 (005^F). Negative synchronizing pulses are fed to the grid of V x through R 3. A variation of this type of circuit is shown in figure One coupling between the two valves is now by means of the common cathode resistor R t. FIG CATHODE-COUPLED MULTIVIBRATOR TIMEBASE CIRCUIT Suppose that V x is conducting and that V 2 is cut off.due to the charge on d. C will charge through R and the anode voltage of V 2 will rise. Eventually, V 2 will conduct due to the discharge of d through R 2 and R 3. This causes the voltage across R t to rise and so tends to cut off V t. This, in turn, causes the anode voltage of V x to rise and feed a positive voltage to the grid of V 2, making it conduct more. Thus, a cumulative action occurs and V t is cut off while V 2 is made conducting, so discharging C and producing the flyback. When C is discharged little current flows in V 2 (R is of large value) and so the voltage across R^ is small and, therefore, V x conducts. This results in a drop in anode voltage of V x which feeds a negative voltage to the grid of V2, so cutting off V 2 and starting the scan. As before, R2 may be made variable or taken to a variable voltage, and acts as the hold control. Synchronizing is now by negative pulses fed to the grid of V t (the grid of which is now free and has no voltage on it from the action of the timebase). These pulses are converted to positive pulses at the anode and are fed to the grid of V 2, through C u causing V 2 to conduct and thus initiating the flyback. Multivibrator timebases have increased in popularity in recent years, both for line and field timebases. It is a simple and reliable circuit and avoids the use of a transformer. Many other timebase circuits are available and used, for example, in cathode ray oscilloscopes, but not in television receivers at the present time. Within limits the higher the h.t. supply to the field timebase the better, because the timebase will give a greater output for a fixed linearity. Figure 3.13 will show for example that the greater the h.t. voltage the more linear will be the output for a constant output voltage; or for a given linearity the output can be greater. It is therefore usual to feed the field timebase from the boosted h.t. supply (from the line output stage: see Chapter 5). Alternatively, only part of the field timebase may be fed from the boosted h.t. In this case the charging capacitor is fed from the boosted h.t. For example, resistor R (of figure 3.18) would be taken to the boosted h.t. supply, and the valve Kj would be fed from the normal h.t. supply. As the field frequency is the same on 625 and 405 lines no changes are necessary when going from one system to the other. A resistor may be added in the h.t. lead on 625 lines because the boosted h.t. voltage may be greater.

63 58 TELEVISION SERVICING TRANSISTOR TIMEBASES Transistors may be used as timebases in place of valves but they are at present mainly used in portable receivers operating on a 12V supply. The timebase may be a blocking oscillator circuit or a multivibrator circuit. A basic blocking oscillator circuit is given in figure 3.19 where typical values are shown for a timebase operating at field frequency. This is the direct equivalent «3 S-6Krt5 C3 5Q(lF FIG BLOCKING OSCILLATOR TRANSISTOR TIMEBASE USING FEEDBACK BETWEEN COLLECTOR AND BASE CIRCUITS of the valve circuit, feedback being by transformer T x between the collector and base circuits. The frequency is mainly determined by the base C-R circuit, consisting of R lt R 2 and C 2. When the transistor conducts it discharges C 3, producing the flyback, and then C 3 charges through R 3 during the scan period. As in the valve circuit C 2 is charged by base current during the flyback and this cuts off the transistor during the scan period, until flyback is initiated by the negative synchronizing pulse. An alternative circuit is shown in figure 3.20, typical component values being given for a field timebase. In this case the feedback is between collector and emitter circuits by means of transformer T t. FIG BLOCKING OSCILLATOR TRANSISTOR TIMEBASE USING FEEDBACK BETWEEN COLLECTOR AND EMITTER CIRCUITS

64 TIMEBASES 59 Assume that C 2 has been charged (base positive) by the last flyback action and hence the transistor 7>, is not conducting. Thus, C 2 discharges through R x and R 2. At the same time C 4 charges through R 3 from the supply. Eventually Tr x will become conducting either because C2 is discharged or, more usually, by the application of a negative synchronizing pulse fed through C x. The rise of current in the collector circuit will induce a voltage in the collector winding of T x and also in the emitter winding. The latter voltage is such as to make the emitter positive and so increase the collector current, i.e. positive feedback. This also causes a base current to flow and charge C 2 (current coming out of the base) so that upper plate of C 2 is positive. As soon as the rate of rise decreases, the voltages induced in T x are reduced and the rate of rise is further decreased. This cumulative action continues until the collector current starts to decrease when, by the positive feedback action, the transistor is cut off and is maintained cut off by the charge on C 2. Thus a heavy pulse of collector current flows which discharges C 4, so producing the flyback. We are now back at the starting point where C 2 discharges through R x and R 2 (forming the frequency or hold control) and C«charges through R 3 so producing the scan voltage. Another blocking oscillator circuit is given in figure 3.21 where the values of the components shown are for a line timebase. In this circuit there is FIG BLOCKING OSCILLATOR TRANSISTOR TIMEBASE USING FEEDBACK BETWEEN COLLECTOR AND BASE CIRCUITS feedback between collector and base circuits by transformer T x. The circuit is shown as being controlled by the control voltage rather than by synchronizing pulses as the circuit is intended to be operated by a flywheel synchronizing circuit (see Volume 3). During the conducting state C x is charged through the emitter of Tr x, so that the emitter becomes negative. Provided this voltage is greater than the control voltage the transistor is cut off. However, C x discharges through R 2 and, eventually, the emitter voltage becomes the same as the base voltage and the transistor starts to conduct. As soon as collector current flows a voltage is induced in the base winding of T x driving the base negative and so increasing the collector current, and producing the flyback. During this period C, is recharged by the flow of emitter current (which is approximately equal to the collector current). As in other circuits the rise of collector current is limited, the induced voltages are reversed and Tr x is cut off. The diode D x and resistor R 3 are added to limit the voltage across the collector winding of T x so that excess voltage is not applied to the transistor.

65 60 TELEVISION SERVICING If the frequency is not correct then the control voltage changes so that the instant at which the transistor becomes conducting is changed. An output pulse (not sawtooth) is produced across the third winding of 7i and is used to drive the driver stage which, in turn, drives the line output stage. A transistor equivalent of the valve multivibrator circuit is shown in figure This circuit does not seem as popular as the blocking oscillator circuit. Suppose that transistor 7r 2 is cut off by the charge on Q and that FIG MULTIVIBRATOR TRANSISTOR TIMEBASE 7>! is fully conducting. Capacitor C 3 is charged through R t so producing the scan voltage. At the same time C\ is discharging through R 3 (and R t ). Eventually the base of Tr 2 reaches the potential of the emitter and collector current flows. The drop in collector voltage (i.e. the collector moves in a positive direction) is fed to the base of Tr r through C 2 and hence Tr x is cut off. The action of cutting off Tr x causes the collector voltage to go more negative. This change of voltage is fed to the base of Tr 2, through C lt so making Tr 2 conduct heavily. The large collector current of Tr 2 discharges C 3 so producing the flyback. Tr x remains cut off until C 2 discharges through R 2 (and R t) (the flyback time) and then the action is reversed, Tr 2 becomes fully conducting and 7>j is cut off, so starting the scan period. The scan period is mainly controlled by the base components of Tr 2, i.e. C t and R 3. Flyback is normally initiated by a negative synchronizing pulse fed to the collector of Tr 1 which results in the negative pulse being fed to the base of Tr 2, so causing it to conduct and start the flyback process. STABILIZATION Some form of stabilization is used in modern receivers which takes two forms: (i) Stabilization against changes in scanning amplitude due to temperature changes. (h) Stabilization against changes in scanning amplitude due to changes in a.c. supply voltage. Two devices are used: the thermistor; and the voltage dependent resistor. Thermistor All metals have a positive temperature coefficient of resistance, i.e. the resistance increases as the temperature rises. The change is not large but it may be appreciable. For example, copper changes about 0-4 per cent, per C.

66 TIMEBASES 61 Alloys used for wire-wound resistors have very small temperature coefficients. With few exceptions, semiconductors have negative temperature coefficients, the change of resistance being generally much greater than that of metals. A thermistor is constructed of semiconductor material (metallic oxides as a rule) and its resistance decreases rapidly with temperature. Figure 3.23 so TEMPERATURE WWW ( C) SYMBOL FIG CHARACTERISTIC OF A THERMISTOR COMPARED WITH COPPER shows a typical characteristic relating temperature of the thermistor to its resistance, with copper shown for comparison. It is important to note that the temperature of the thermistor may be raised by external means (by contact with another hot body) and/or by the passage of current through it. There will always be some heat produced by the current passing through the thermistor, the effect of which will depend on the magnitude of the power dissipated in the thermistor and on its physical size. The thermistor used in has little television receivers is relatively large and the power dissipated in it effect on its temperature. Thus, as regards the current in the circuits it acts as a linear resistor, the resistor changing in value if the temperature of the thermistor is affected by variations in the temperature of adjacent objects, e.g. scanning coils. [Thermistors are also made in small sizes and the effect of the current flowing through these causes large changes in resistance. Interesting and useful characteristics result but this type of thermistor is not used in television receivers]. The thermistor that is used in television receivers is available in various physical sizes with differing resistance values for a given temperature, varying from a few ohms to thousands of ohms at 25 C. If a thermistor is replaced the replacement must be of exactly the same type. Voltage Dependent Resistors (VDR) These are also known under the trade names of Thyrite, Metrosil and Atmite, but in television receivers they are usually known by the abbreviation VDR. These devices are constructed from silicon carbide crystals with a binding clay and they are fired at a high temperature. Their property is quite different from that of a thermistor and is nothing to do with temperature. A typical characteristic is shown in figure 3.24 from which it will be seen that the device does not obey Ohm's law, i.e. the current increases much more

67 62 TELEVISION SERVICING rapidly than the voltage. The characteristic can be expressed approximately by / = k.v, where k and n are constants, n being between 4 and 5. This means that if the voltage is doubled the current increases by 16 to 32 times. They are made in the form of rods (similar to resistors) and discs (similar to ceramic capacitors). Various types are available having different "resistances", i.e. requiring different voltages to pass the same current. It will be seen from the figure that the characteristic is the same in both directions and FIG CHARACTERISTIC OF A VOLTAGE DEPENDENT RESISTOR (VDR) OR METROSIL therefore it is not a rectifier. However, it can be used to produce d.c. from a pulse input (explained later); but it will not produce d.c. from a sinusoidal a.c. input. It is this use with a pulse input that tends to cause confusion. The characteristic is instantaneous and does not depend on any heating effect of the material. (The characteristic does change with temperature; the change is not large as in a thermistor, and is an effect that is not normally used). As the VDR has a non-linear characteristic, i.e. it does not obey Ohm's law, so one cannot quote or measure its resistance because this varies so much with applied voltage. One must state the current it passes when a certain voltage is applied to it. Again, a replacement must be exactly similar: it is useless measuring its resistance with an ohmmeter. TEMPERATURE STABILIZATION As the temperature of the scanning coils rises (due to the current passing through them and to the increase in the ambient temperature produced by heat inside the cabinet) the resistance increases. This causes a reduction in current flowing in the coils, which means a reduced field scan. Thus, if the receiver's field amplitude control is correctly set when the receiver is cold the vertical scan will be too small when the receiver has warmed up. Some means of preventing this is required which may be one of the following. (/) Including a thermistor of suitable value in series with the field scan coils and placing it close to the coils, i.e. in thermal contact so that when the coils are cold their resistance will be low, but the resistance of the thermistor will be high. As the scanning coils heat up, their resistance increases, and that of the thermistor decreases. If the thermistor is of a suitable value the increase in the resistance of the coils can just be compensated by the decrease in the resistance of the thermistor, i.e. the total resistance and the magnitude of the scan remain constant.

68 TIMEBASES 63 By a suitable choice of thermistor it may also be possible to compensate for the increase in the resistance of the field output transformer due to temperature. (h) Applying a greater scanning voltage to the field output stage as the temperature rises. Figure 3.25 shows a possible circuit using a multivibrator timebase, but the same idea can be applied to a blocking oscillator type of FIG USE OF A THERMISTOR TO STABILIZE A TIMEBASE AGAINST TEMPERATURE VARIATIONS timebase. During the scan, capacitor C is charged through the resistance R, which comprises thermistor T and resistor R t. The thermistor T is placed in thermal contact with the scanning coils. As the temperature of the scanning coils rises the resistance of T decreases, reducing the value of the charging resistance R and increasing both the amplitude of the timebase output and the output of the field output stage. If the components are chosen correctly the increase in output of the output stage will just compensate for the increase in the resistance of the scanning coils. Arrangement (0 seems to be the more popular. SUPPLY VOLTAGE STABILIZATION A decrease in the mains supply voltage causes a decrease in the h.t. supply of approximately the same percentage; the boost h.t. supply also decreases (which may be stabilized to some extent by the line output stage). This causes the output of the field timebase to decrease and hence a decrease in the magnitude of the scan. The picture size will decrease to an extent that might be annoying to the viewer. Some means of stabilization should be used which J R 2 v»f FIG USE OF A VDR TO STABILIZE A VOLTAGE (a) Circuit (b) Method of operation

69 64 TELEVISION SERVICING amounts to maintaining an approximately constant charging voltage for the timebase. If a linear resistor Ri and VDR (R 2) are arranged as shown in figure 3.26(a) the percentage variation of output voltage V^ is much smaller than the percentage variation of input voltage V in. Suppose that the input voltage increases and the current increases from ;'i to i\ [figure 3.26(b) ]. The change of voltage across the linear resistor R t will be from V R i to V' R1 ; across the VDR (R 2) it will be from V R2 to V' R2, which is much smaller than across the resistor R u The total input voltage variation is the sum of that across R r and that across R 2. From the figure it will be seen that the change in voltage (y^t) is much less than that across the input (V in) Obviously, the steeper the charac- across R 2 expressed as a percentage of the voltage. teristic of the VDR the better. particularly when This type of circuit can therefore be used to feed the charging circuit of the field timebase, an example being shown in figure The charging circuit is R and C which is fed from across the VDR, R 2. Thus, the voltage fed to R and C is much more constant and the variation of scan amplitude is less. FIG USE OF A VDR TO STABILIZE A TIMEBASE AGAINST SUPPLY VOLTAGE VARIATIONS

70 The CHAPTER 4 FIELD OUTPUT STAGE purpose of this stage is to provide a suitable current in the field deflecting coils placed round the neck of the cathode ray tube. Since it is the magnetic field that produces the deflection which is produced by the current in the coils it is the current which is important, and not the voltage across the coils. The waveform required is a sawtooth waveform of current similar to the voltage waveform produced by the timebases described in Chapter 3. The scan must be as linear as possible but non-linearity up to 5 per cent, is satisfactory. The shape of the flyback is not particularly important.., With no deflecting current in the coils the spot will be in the centre ot the screen of the cathode ray tube and therefore deflection is required in both directions. This means that the current in the coils must be pure alternating current since a steady d.c. component would shift the spot from the centre and cause the scan to be all on one side of the screen. The operation of the field output stage is different from that of an output stage operating a loudspeaker. The deflecting coils may be represented by the circuit shown in figured 1. L is the inductance of the coils and R is the ohmic resistance. FIG EQUIVALENT CIRCUIT OF FIELD DEFLECTING COILS Me: C represents the capacitance across the coils and R s represents the losses in the coils (other than that due to R) due to (say) hystensis loss in the core. At field frequency R s and C are not very important. Let us first consider the voltage which will be produced across the coils when a sawtooth current flows in the coils. First, the scan which should consist of a linearly rising current as shown in figure 4.2. This current, by CURRENT Scan Time FIG WAVEFORM ACROSS DEFLECTING COILS DURING SCAN Ohm's law, will produce a linearly rising voltage across the resistance R (remembering that R is not actually a separate physical component but the resistance of the coils) and V R = / x R. The current flowing trough L causes a voltage across L equal to: L x rate of rise of current. The rate of 65

71 66 TELEVISION SERVICING rise of current in this case is constant, hence the voltage across L is constant, independent of the actual value of the current. The rate of rise of current is Current / Time T where T is the time of scan. Hence, voltage across L(V L ) = L x I/T. Thus Voltage across R ** I.R R.T Voltage across L V L L x IIT L In practice R/L is 300 to 1,000 (R being in ohms and L in henrys), say 1,000 hence VJV L = 1,000.T Now, considering a 405-line system, the field scan time equals Field time blanking period (14 lines) = 20,000-1,400 = 18,600/xS. Hence, V R /V L = 1,000 x 18,600 x 10~6 = 18-6 Thus, if we assume that V R is unity voltage then V L = 1/18-6 = 0054V. It is seen that the voltage across R is much greater than the voltage across L and, during the scan, we can largely forget the effect of the inductance L. In the 625-line system the field frequency is, of course, the same as in the 405-line system namely 50Hz. The blanking period with the 625-line system is slightly different, being 20 lines which equals 20x64 = 1280/aS, compared with 1400/nS for the 405-line system. This difference can be neglected for the present purpose apart from the fact that the flyback of the field timebase should be completed in less than 1280/xS. During the flyback the change of current is the same but the change of current takes place in a much shorter time. The voltage across R will be the same, as this does not depend on the rate of change of current but only on its value. On the other hand, the voltage across L will be increased due to the higher rate of change of current. Since the current is now decreasing, the voltage produced across L will be of opposite sign and, if the flyback is assumed linear (for simplicity), the voltage will be constant during the flyback. The complete voltage waveform is now shown in figure 4.3, drawn approximately to scale. The flyback time in a 405-line system is, at the most, 14 lines of l,400jtis (in a 625-line system it is 1,280/xS), hence the rate of change of current during flyback is 18,600/1,400 = 13-3 times as great as during scan. Thus, the voltage produced during flyback is also 13-3 times as great as the voltage across L during scan, i.e. 13-3x0-054 = 0-72V. The voltage across the actual coils {i.e. across L and R) is the sum of the voltage across L and the voltage across R, figure 4.3. This is seen to be substantially a sawtooth waveform but with a negative spike during the flyback. In practice the spike is likely to be larger than that shown because: (i) The flyback time may be shorter than the total blanking time. (ii) The flyback will not generally be linear so, at some period of the flyback, the rate of change of current will be greater than that calculated above, hence the voltage will be larger. To simplify the figures the voltage across R was taken as IV. In practice this will be much greater and will depend on the type (impedance) of the deflecting coils used. It will be seen later that the waveform of the voltage across the line deflecting coils is much different from that across the field coils. We may consider the field coils as consisting of a resistance with a little inductance,

72 FIELD OUTPUT STAGE 67 VOU»GE 1CROSS R CO VOLTAGE ACROSS L CO FIG WAVEFORMS ACROSS DEFLECTING COILS (405-LINE SYSTEM) whereas the line coils act as if they consisted of an inductance with a little resistance. HIGH IMPEDANCE DEFLECTING COILS As the voltage induced in the field deflecting coils is not very large, high impedance deflecting coils may be used, these being designed to match directly to a normal valve. The coils cannot be placed directly in series with the valve as the steady anode current would cause the raster to be displaced to one side of the tube. To prevent this they must be fed through a capacitor C as shown in figure 4.4. The inductance of the coils is 1-3 henrys and the xvv^ FIG 4.4. CIRCUIT FOR OPERATION WITH HIGH IMPEDANCE COILS resistance about 1,000-3,000 ohms, the current being of the order of 40mA peak to peak (12" tube). To prevent distortion of the current waveform C must be large, say 16 to 32/j.F and even then some correction is often necessary to obtain a linear scan. The valve is fed through the resistor R t which has a value of 2,000 to 5,000 ohms. The mam difficulty with this circuit is that it will only operate with a high h.t. voltage, say 350V. This voltage is not available in modern receivers, owing to the use of a.c./d.c technique. Hence the circuit is not used in modern receivers and we will not go into further details.

73 68 TELEVISION SERVICING LOW IMPEDANCE DEFLECTING COILS The deflecting coils used in modern receivers have a low impedance with an inductance of (say) 5-50mH and a low resistance of 2 to 20 ohms and must be fed through a step-down transformer from the field output valve in a similar manner to a loudspeaker. A transformer is a rather complex piece of equipment if we try to represent it by an equivalent circuit ; but an approximate circuit is shown in figure 4.5. R 2 represents the total resistance IDEAL TRANSFORMER FIG Primary Secondary EQUIVALENT CIRCUIT OF FIELD OUPUT TRANSFORMER of the transformer (both primary and secondary) and L 2 the leakage reactance. This is the effective reactance which is caused by the fact that all the magnetic flux produced by the primary does not pass through the secondary and vice versa, but produces what is known as leakage flux. This flux causes a voltage to be induced in one winding but not in the other and can, therefore, be represented by an inductance L 2. R 2 and L 2 can be added to the R and L of the deflecting coils and do not make much difference to the secondary circuit. R t represents the core losses in the transformer and is not very important. L x represents the primary inductance of the transformer. This is the inductance when nothing is connected to the secondary and the current which flows in L x is that necessary to produce the magnetic flux in the transformer core. During the scan we have seen that the waveform across the deflecting coils is approximately a linearly rising voltage and hence a similar voltage occurs across the primary of the transformer. The current flowing in R t would be of similar waveform but not so the current in L t. The current in L t must be such that it will produce the steadily increasing e.m.f. to oppose the steadily increasing voltage applied to it. Since the voltage is proportional to the rate of rise of current, this rate of rise must steadily increase as the scan progresses. In practice, since the voltage across and the current in the deflecting coils are pure alternating quantities, the voltage is decreasing from a negative value to zero during the first half of the scan and then increases from zero in a positive direction as shown in figure 4.6. Hence, the current in Z,j will decrease during the first half of the scan (so that the rate of rise and voltage is negative) and then increase during the second half of the scan (so that the rate of rise and voltage is positive). The way in which the current changes is in the form of a curve. The slope of the curve (i.e. the rate of change) at any point must be proportional to the magnitude of the voltage across L^ hence the curve shown. The total current is the sum of that in L x plus that in the primary of the transformer and the latter will be sawtooth like the current in the deflecting coils. It is seen from figure 4.6 that the total current is not a linear sawtooth waveform but is distorted or curved. Hence the valve must provide this distorted current waveform if we are to obtain a linear scan. The distorted component of current in L r can be reduced by increasing the inductance of the transformer but it cannot be made negligible with a practical design

74 FIELD OUTPUT STAGE 69 Current in Coils & V Haae aero si Coili t Primary of Transformer Current in L Total Current FIG WAVEFORMS OF VOLTAGE ACROSS, AND CURRENTS IN, TRANSFORMER of transformer. Methods of obtaining this distorted waveform from the valve will be considered later. The operation of the circuit on scan is not so very dissimilar to that of a normal output stage, since the deflecting coils, as we have seen, act largely as if they were a resistance, like a loudspeaker. On flyback (the valve tending to cut off) a voltage is induced in the primary of the transformer in a similar way to the deflecting coils. The rate of change of current on flyback depends on the ratio of resistance to inductance in the circuit and is greatest when the ratio is high. (The higher the inductance the more it tries to prevent any change of current in the circuit). Hence, with a pentode (which has a high internal resistance) a rapid change of current occurs which results in a short flyback time but may result in excessive voltage. To reduce this voltage a resistor may be placed across the deflecting coils. This decreases the resistance of the circuit and, therefore, increases the flyback time and so must not be too low in value or the flyback time will be excessive. This type of circuit is now universally used as it will operate successfully on a low h.t. voltage such as 180V. The ratio of the transformer is about 10/1 to 20/1. Returning to the question of the non-linear current that is required to be produced by the valve it is useful to note that this is, unfortunately, opposite to the type of distortion which occurs in a normal timebase, i.e. when an exponential rise of voltage occurs. Some correction usually occurs due to the characteristic of the output valve, as shown in figure 4.7. The linear input. ANODE OR lo OUTPUT CURRENT FIG CORRECTION FOR NON-LINEARITY BY VALVE CHARACTERISTIC waveform is "distorted" to the correct shape by the curvature of the valve characteristic. It is not desirable to rely too much on this type of correction

75 70 TELEVISION SERVICING as it is likely to change from valve to valve and depends on the bias of the valve. Therefore, other types of correcting circuits are often used. Before discussing these circuits let us consider the effect of partial integration and partial differentiation of the sawtooth waveform. The effect of passing a sawtooth waveform through an integrating circuit is shown in figure 4.8: (a) is with a small time constant, (b) with a larger time constant FIG INTEGRATOR CIRCUIT AND WAVEFORMS and (c) with a large time constant. The general effect of an integrating circuit is to round off the corners of the waveform. The corresponding effect of a differentiating circuit is shown in figure 4.9. In this case (a) is with a large time constant, (b) with a smaller time constant and (c) with a small time FIG DIFFERENTIATOR CIRCUIT AND WAVEFORMS constant relative to the time of the sawtooth wave. The general effect of the differentiating circuit is to produce a large output when the input waveform is changing rapidly but only a small output when the rate of change of input voltage is small. The principle of many distortion correction circuits is shown in figure Suppose that the input waveform shows some distortion of the type commonly produced by a timebase as shown at (a). If this waveform is integrated, by a circuit with suitable time constant, the waveform of (b) is produced. If the two waveforms are added the result is as shown at (c). It will now be seen that the curvature of the waveform is opposite to that at (a) and of the type required for feeding the grid of the field output valve, in order to overcome the effect of the finite inductance of the transformer. The amount of

76 FIELD OUTPUT STAGE 71 I A /\ ADDITION OF \y\y\ - FIG PRINCIPLE OF CORRECTION CIRCUITS correction obtained depends on the relative magnitudes of the waveforms at (a) and (b). A circuit for producing this correction is shown in figure C x and C 2 form the charging capacitors of the timebase and the waveform across these will be approximately sawtooth. The voltage across C x is integrated by R U C 3. The integrated voltage across C 3 is added to the voltage "I -vwv- INPUT c, OUTPUT,-VV FIG DISTORTION CORRECTING CIRCUIT across C 2, to give the corrected output waveform. The circuit is not too convenient in practice owing to the need for a split charging capacitor, which must be split in the correct ratio to give the required correction and yet be of the correct total capacitance. A circuit which produces a similar result is shown in figure If R x and C t have suitable values the waveform of the current in the circuit FIG DISTORTION CORRECTING CIRCUIT will be approximately the same as the waveform of the input voltage, and so the voltage across R t will also have the same waveform. The voltage across Cj is the integrated waveform of the input voltage. By taking the voltage across Q and a fraction of that across R t a corrected waveform is obtained, the amount of the correction depending on the position of the slider on R x. The most common correction circuit is a negative feedback circuit and shown in figure V is the field output valve with output transformer T feeding the deflecting coils D. C x and C 2 form the usual charging capacitor of the timebase, the waveform across them being fed to the grid through the

77 72 TELEVISION SERVICING FIG FEEDBACK CIRCUIT FOR CORRECTION OF DISTORTION coupling capacitor C 3. Feedback is provided from the anode through the differentiating circuit C t,r 2 and the integrating circuit R 3,C 2. The voltage at the anode is similar to that shown in figure 4.14(a). This is of opposite FIG, JVJV WAVEFORMS IN CIRCUIT SHOWN IN FIGURE 4.13 polarity to that across the coils shown in figure 4.3. After differentiating by C A,R 2 the waveform is distorted as shown at (b) (compare with figure 4.9). The effect of the integrating circuit C 2,R 3 is to remove the sharp pulses and give a result similar to that at (c). The waveform when added to the normal voltage of the timebase, such as that across C t, gives the curvature required (i.e. as figure 4.10(c) ) to correct for the distortion produced by the transformer T. One may consider the circuit in another way. The ideal feedback would be current feedback (i.e. a voltage feedback proportional to the current in the deflecting coils) which would result in a linear sawtooth current in the coils. This is not a practical proposition using a simple circuit, as the voltage that could be obtained by passing the current through a resistor in order to obtain a voltage to feed back to the grid would be small, unless the resistor was so high that it upset the operation of the circuit and caused a power loss appreciable compared to the power in the deflecting coils. Now, the inductance of the transformer, together with the internal resistance of the valve, can be shown to form a differentiating circuit and is shown in figure This is then coupled to the deflecting coils. The current which flows in the deflecting coils can be shown to be approximately the integral of the voltage across them (i.e. as regards the current in them they behave as an integrating circuit). Thus, we have a differentiating circuit followed by an integrating circuit. By taking the voltage from the anode of the output valve, passing it through a capacitance-resistance differentiating circuit (C 4,R 2) and then through a resistance-capacitance integrating circuit (R 3,C 2) we shall have a voltage output which is the same waveform as the current in the deflecting coils. For this to be true the resistance-capacitance integrating and differentiating

78 FIELD OUTPUT STAGE 73 R-C EOUIV1LENT CIRCUIT FIG EQUIVALENT CIRCUIT OF FIGURE 4.13 circuits must be equivalent to the corresponding circuits in the anode of the output valve, i.e. they must have suitable values. If then we feed the voltage obtained from this circuit back to the grid of the output valve we are really getting current feedback. The effect of the circuit can be varied by varying R 2 and/or R 3. Commonly, R 2 is the linearity control and R 3 is fixed or preset. Typical component values are shown in figure 4.13, which is the type of circuit generally used. It might be mentioned that in the above we have assumed that a sawtooth current is required in the deflecting coils. When the screen is almost flat this is not so (see the next chapter on line output stage) but the distortion-correcting circutis still operate and are used to provide the required distorted current in the deflecting coils. A typical field output stage is shown in figure Valve V x forms the discharge valve of a multivibrator timebase, C\ (OljuF) being the discharge II-»-TO "^ CWD OF CRT FIG TYPICAL FIELD OUTPUT STAGE capacitor. This is charged through the height control J?! (lmii) and R 2 (470kH). The voltage across C t is fed through coupling capacitor C 2 (01/aF) to the grid of the output valve V 2, through the grid stopper R 3 (12k i). R t (2-2MQ) is the grid resistor. The deflecting coils L t and 2 are fed through transformer T. Feedback is provided by the differentiating circuit C 5 (005/aF), R s (250kO) and R? (82kO) and integrating circuit R 7 (50kQ), R 6 (120kQ) and C 3 (OOluF) and the capacitance across the resistor /? 4 formed by Q (0-1/iF) and C 2 (O-luF). Grid bias is by R 5 (lkfi) by-passed by C 4 (50^F). To reduce the damped oscillation which is commonly induced in the field deflecting coils by the line coils R n (lkfi) and R 12 (lkq) are placed across the field coils. The flyback pulse across the field deflecting coils is suitably shaped by R 10 (22kii) and C6 (0002fiF) and fed through C 7 (0-01/aF) to the grid of the cathode ray tube so that the spot is blacked out during the field

79 74 TELEVISION SERVICING flyback time. This prevents the possibility of the field flyback lines showing on the screen when the brightness control is advanced rather more than it should be, so that a grey instead of a black background is obtained. This is important when mean level a.g.c. is used because the black level of the picture increases when most of the picture is black. In some early 110 receivers a direct current was injected into the field deflecting coils. In these receivers the linearity correction was such that the picture tended to be off centre. Although the picture might be centred by the normal magnets around the neck of the tube it was undesirable to do so because of the possibility of picture distortion. The direct current in the coils was usually obtained by connecting the coils and secondary of the field output transformer across a small resistor (of about half an ohm) in the h.t. lead of the receiver. TRANSISTOR FIELD OUTPUT STAGES At the present time most field output stages use valves, except those in portable transistor receivers intended for operation on a 12V supply. The transistor field output stages are similar to the valve circuits and a basic circuit is shown in figure FIG FIELD OUTPUT STAGE R lt R 2, Cj and C 2 form the timebase C-R circuit and capacitors C t and C 2 are discharged during flyback by the discharge circuit, which is the field timebase. The voltage across Cj and C x is fed to the base of transistor Tr x which is connected as an emitter follower. An emitter follower stage is used so that the loading across C x and C 2 is small (an emitter follower having a high input impedance) and because it has a low output impedance it is suitable for driving the output transistor Tr 2. The deflecting coils cannot be placed directly in the collector circuit of the output transistor Tr 2 as the steady collector current would deflect the picture from the centre of the screen. The coils must therefore be fed through a transformer or by a shunt feed circuit, the latter being shown in figure Thus the d.c. component of the collector current flows in the inductor L and the a.c. component through C 3 and the deflecting coils D. As the deflecting coils are of low impedance (due to the low voltage supply, normally 12 volts) C 3 must be large in order to have a low reactance. A voltage dependent resistor (VDR) R 7 is placed across L to reduce the peak voltage during flyback, which might damage the transistor Tr 2. Resistors R t and R 5 form a feedback loop to the charging capacitor C 2 to improve the linearity.

80 : : The CHAPTER 5 LINE OUTPUT STAGE operation of the line output stage is quite different from the field output stage just discussed and before dealing with circuits let us consider the waveforms produced across the deflecting coils. As in the case of the field coils, the line coils may be represented (approximately) by an inductance L and a resistance R in series, together with a capacitance C and resistance R s in parallel, as shown in figure 5.1. In the case of the line FIG EQUIVALENT CIRCUIT OF LINE DEFLECTING COILS deflecting coils it of great importance and R s is more important than in the case of the field will be seen later that R is not very important but C is now deflecting coils. Referring back to the field coils we showed that the ratio of the voltage across R to the voltage across L was VJV L = 1,000 T where T is the scan time. In the case of the line coils T (in the 405-line system) is Line time (synchronizing pulse time + back porch time) = (9 + 7) = = 84/xS. Hence, VJV L = 1,000 x 84 x 10-«= 0084 or V L = 1/0-084 = 12K S. In other words, the voltage across R is now small compared with that across L, opposite to that which occurs with the field coils. This is due to the fact that the current change takes place much more rapidly and hence the voltage induced in the coils is much greater but, for a given current, the drop across the resistance portion is constant, independent of the rate of change of current. The voltage across L during flyback will be greater than on scan. The scan time is 84/nS and the flyback time 16/xS and so the voltage on flyback is: 84/16 x voltage on scan = 5-3 x voltage on scan, thus the voltage on flyback equals 5-3 x 12 x V R = 64K* since V R is the same on flyback as on scan. In practice it may be much greater than this as the flyback is not linear and may take place in less than 16/iS. Typical waveforms (drawn approximately to scale) are shown in figure 5.2, the voltage across R being taken as unity voltage. The total voltage is the sum of V R and V L and, as will be seen, is quite different from that produced across the field coils. The waveform consists of an approximately constant voltage during scan and a large negative pulse during flyback. If we assume that the line and field coils have the same resistance and inductance and require the same deflecting current (approximately true in practice) then the voltage across the resistance component will be the same in both cases. This has been made unity voltage

81 ' 76 TELEVISION SERVICING in figures 4.3 and 5.2. It will be seen that the voltage across the line coils is about 13 times greater during scan and about 50 times greater during the flyback. CURRENT VOLTAGE ACROSS R (vr) 12 Volts,64 Volts VOLTAGE ACROSS L K).. _ FIG WAVEFORM OF CURRENT IN, AND VOLTAGE ACROSS, LINE DEFLECTING COILS In the 625-line system the Scan time T line time (synchronizing pulse time + back porch time) = 64 - ( ) = = 53-5/aS. Therefore V R jv L = 1000 x 53-5 x 10-* = or V u = 1/ V R = 18-7 V R This means that the voltage across L is now of even greater importance than on 405 lines. The scan time is 53-5 and the maximum flyback time is 10-5/u.S. Therefore, voltage on flyback 53-5 = x voltage on scan = 5-1 x voltage on scan. This is approximately the same ratio as on 405 lines.

82 LINE OUTPUT STAGE 77 But now the voltage on flyback is 5-1 X 18-7 x V = 95 V R, which is even greater than on 405-lines. Due to these high voltages we cannot use high impedance coils and so the line deflecting coils are fed through a step-down transformer from the line output valve. As already said the transformer is a complex device having winding resistance and leakage reactance, magnetizing current, core losses and winding capacitances; if we bring all these into the circuit it becomes so We must, therefore, make complex as to be almost impossible to deal with. some rather large approximations. First, the winding resistance may be added to the resistance of the coils and the leakage inductance can be added to the inductance of the coils. The losses in the transformer core can be added to the losses (represented by jr s of figure 5.1) of the coils and the capacitance of the transformer added to that (C) of the coils. Thus, we arrive at an equivalent circuit which is as figure 5.1, except that the values of the various components of the circuit will have different values from those of the coils themselves. In this way we can simplify the problem to one which we can deal with but, it should be remembered, quite drastic approximations have been made. Before dealing with the deflecting coils we must consider the behaviour of a parallel resonant circuit when the d.c. supply to the circuit is suddenly removed. Suppose we consider the circuit shown in figure 5.3(a) in which FIG CURRENT IN, AND VOLTAGE ACROSS, RESONANT CIRCUIT (a) With supply connected; (b) Jusc after removal of supply; (c) After current has reversed a steady current I L flows from the source V. The capacitor C will be charged as shown to voltage V and the current will all flow through R and L, the voltage being dropped across R (there being no drop in L as the current is of steady value). The current flowing in L will produce a magnetic field which corresponds to a certain stored energy, actually equal to ili L 2 joules. If R is small there is little energy stored in C as the voltage V is small. Assume that the supply is suddenly removed. The current in L cannot suddenly stop since the energy stored in L cannot suddenly be lost. I L starts to reduce but, in so doing, a voltage is induced in L tending to keep the current flowing, this voltage being equal to: L x rate of change of current. The current cannot flow through the external circuit and must, therefore, charge up the capacitor C with the polarity shown in figure 5.3(b) which is opposite to that shown at (a). The current continues to decrease as shown in figure 5.4 until it reaches zero at A. There is now no energy stored in L (since there is no magnetic field because the current is zero) and all the energy has been transferred to C (apart from a small loss in R). The energy is now equal to \C. V * c joules. Although the current is now zero the circuit cannot remain in this condition owing to the charge on C. Capacitor C now starts to discharge through L and causes a current to flow through L opposite to the original direction, as seen in figure 5.3(c) and figure 5.4. The current rises to a maximum value at B (figure 5.4) when the voltage across the capacitor is zero and all the energy (apart from the loss in R) has been transferred back

83 78 TELEVISION SERVICING FIG CURRENT IN, AND VOLTAGE ACROSS, RESONANT CIRCUIT AFTER SWITCHING OFF SUPPLY to L. The condition is now similar to that at the start, except that the current is flowing in the opposite direction. I L now decreases again and C is charged in the opposite direction, and so on. Each time the current flows in R energy is lost and so the currents and voltages become less and less and we get what is known as a damped oscillation, as shown in figure 5.5(a). The decrease in the peak values of the current and voltage is exponential. The frequency of the damped oscillation produced is given approximately by / = 1 2tt\ and is approximately the same as the resonant frequency of the circuit. If the CURRENT VOLTAGE Q FIG. 5.5(a) DAMPED OSCILLATION IN RESONANT CIRCUIT; (b) CRITICAL DAMPING OF RESONANT CIRCUIT resistance is high enough the current does not reverse but dies away as shown in figure 5.5(b); if the resistance is just high enough to prevent the current reversing, the circuit is said to be critically damped. As well as the series resistance R there may be a shunt resistance R s but the effect of this may be allowed for by increasing the value of R in the circuit diagram. Similarly the effect of R may be incorporated in R s by suitable choice of R s. Suppose that we now consider the deflecting coils fed off the line output valve as shown in figure 5.6. The coils are shown directly in the anode circuit for simplicity, but it must be remembered that they are actually fed through a transformer of suitable turns-ratio. The series resistance has been omitted and has been combined with R s (which is made of suitable value to allow for

84 LINE OUTPUT STAGE 79 I ( Increasing) FIG BASIC OUTPUT STAGE this). During scan the current in L will increase at a constant rate (it must do so if the scan is to be linear) and, since the voltage induced in the coils is L x rate of change of current this voltage must be constant, since the rate of change of current is constant. Now, the rate of change of current is I/T where / is the final value of the current and Tis the line scan time. This voltage (V s ) is the induced voltage in L (often called the back e.m.f.) tending to prevent the current rising and its direction is as shown in figure 5.6. The voltage is tending to make the anode negative and hence reduce the voltage on the anode of the valve. The capacitor C will charge to this voltage V s but, since this voltage is constant, no current will flow in C. There will be a small constant current in R s equal to V S IR S but of little importance. The waveform of current and voltage during scan is shown from A to B in figure 5.7. Assume that at the end of CURRENT I CURRENT IN, AND VOLTAGE ACROSS, DEFLECTING COILS WITH CRITICAL DAMPING FIG scan, B, the valve is cut off by a suitable negative voltage applied to the grid. The current in L cannot cease instantly due to the energy stored in the coils. The current, therefore, decreases and a voltage is developed tending to maintain the current, i.e. opposite to that during the scan, and this voltage makes the anode of the valve more positive (i.e. more positive than the h.t. supply). If the resistance R s is sufficiently low to critically damp the circuit, the current will decrease to zero as shown from B to C. At C the valve starts to conduct again and the next scan is produced. If R s is of higher value (less damping) a damped oscillation will be produced as shown in figure 5.8. FIG CURRENT IN, AND VOLTAGE ACROSS DEFLECTING COILS WITH SMALL DAMPING If this oscillation is allowed to continue into the next scan it will cause variations in the speed of the scan (or it may even cause parts of the scan to fold over owing to the spot reversing its direction of travel) causing dark and bright lines across the screen. That shown in figure 5.7 may appear ideal

85 80 TELEVISION SERVICING but in this case the decrease of current is rather slow and it is not normally possible to get the flyback completed in the short time allowed (11 or 16 ns). This means that the spot would not have returned to the left-hand side of the screen by the time that the picture signal started and the left-hand edge of the picture would be folded over. To increase the flyback speed the damping is reduced and some overshoot is allowed in the opposite direction (say 10 per cent.) as shown in figure 5.9. The rate of change of current in the 7"~ FLYBACK SCAN FIG CURRENT IN, AND VOLTAGE ACROSS. DEFLECTING COILS WITH INCREASED DAMPING circuit is now increased and the flyback time decreased. For this action to take place so that the flyback joins on to the scan it is necessary to adjust C and R s to the correct values. Suitable components could be connected across the primary of the transformer but the voltage is large and they are, therefore, normally connected across the secondary as shown in figure 5.10(a), i.e. /?i and d. In practice they are more commonly connected in series as O DEFLECT INC COILS FIG USE OF DAMPING CIRCUIT ACROSS LINE DEFLECTING COILS (a) Parallel circuit; (b) Series circuit shown at (b). Typical component values (for circuit shown at (b) ) are Ri = 5kfl and C t = 0002 to 001/xF. In this arrangement all the energy which has been put into the coils during the scan must be dissipated in the resistor Ri (and the core losses of the transformer and the resistance of the coils and transformer); therefore, the efficiency is poor. Since most of the energy of the line output valve is thus dissipated in if, this must be of high wattage, say 5-10W. If the power output of the valve is sufficient, the circuit is simple and easy to adjust but is most wasteful of energy. For that reason this arrangement is not used to-day and a more efficient circuit will be described later. OPERATION OF VALVE The operation of the valve is now quite different from that of a normal sound output valve. Since the voltage across the deflecting coils (and primary of the transformer) is constant during scan, the voltage on the anode of the valve must be constant (equal to the h.t. voltage (F s ) less the voltage (V L ) across the primary of the transformer). The load line is now a vertical line (corresponding to a constant anode voltage) and as shown in figure 5.11 and

86 LINE OUTPUT STAGE 81 toad LINE -Vr FIG OPERATION OF LINE OUTPUT STAGE not a sloping line as for a resistance load. During the scan the operating point moves along the line from A to B and B may be arranged so that it is to the right or left of the knee of the characteristic. When the circuit was not stabilized (as in older receivers) the load line was to the left of the knee and limiting occurred due to the knee of the characteristic. When the circuit is stabilized the load line must be to the right of the knee so that the output can be varied by varying the bias of the valve (see later). As the anode voltage is low during scan and the anode current is cut off during flyback, the anode dissipation of the valve will be low. Suppose that the h.t. voltage is 300V and that the mean anode current is 100mA. The power dissipated in a normal output stage would be 300 x 100 x 10-3 = 30W. (It would be somewhat less than this with a signal, owing to the power that is delivered to the loudspeaker, but the circuit must be designed so that the dissipation is not excessive with no signal). Now, if the voltage across the transformer primary is 200V during scan, the actual anode voltage is = 100V and the dissipation is 100 x 100 x 10" 3 = 10W. (It would actually be less because we have not allowed for the fact that there is no energy dissipated in the valve during flyback). For this reason it is important not to run the line output stage without a suitable driving voltage on the grid because, in the absence of the sawtooth drive, the anode dissipation in the above case would rise to 30W and most Owing to the fact that the anode voltage is low, the likely damage the valve. line output valve tends to take a large screen current and the power output may be limited more by screen dissipation than by anode dissipation. EFFICIENCY SCAN CIRCUITS The normal circuit just described has a low efficiency, since the energy put into the scan coil during the scan is dissipated as heat in the damping resistor during the flyback. With a small screen cathode ray tube and a low e.h.t. this is not very important, but the following factors make it essential to use a more efficient arrangement: (i) (k) (Hi) Increased size of tube with increased scanning angle. Increased e.h.t. Reduced h.t. voltage due to the use of a.c./d.c. technique. With (0 and (w) much more power requires to be fed to the deflecting coils (to be accurate this is volt-amperes, i.e. product of voltage and current, and not watts) than with small tubes operating at low e.h.t. Hence, the line output valve must give out greatly increased power. Factor (Hi) means that, for a given power in the output valve, decreasing the h.t. voltage results in increasing the anode current and it becomes extremely difficult to design valves which will take a heavy current with a low anode voltage and, of course, stand the high voltage peak during flyback. Neglecting any losses in the deflecting coils, it is important to note that no power is required to produce deflection since the energy that is put into the

87 : 82 TELEVISION SERVICING coils during scan may, if a suitable circuit can be developed, be recovered during the flyback and used for the next scan. The device is rather similar to a pendulum, where the energy is changed from potential energy due to the height of the weight (or bob) to kinetic energy due to motion at the bottom of the swing and vice versa. The energy required to keep the pendulum swinging is very small just equal to the friction losses. All we wish to do is to build up a current in the coils, which requires energy, and then rapidly reduce the current to zero, take out energy and store it in some way ready for the next scan. This idea is of no importance on the field scanning circuit since the frequency is low, i.e. 50Hz compared with the line of 10,125Hz or 15,625Hz, and the scanning power required is proportional to the frequency, if the power required to be fed into the circuit on each scan is the same. Most of the energy in the field coils goes into the losses in the coils and circuit and cannot be recovered. Suppose that we consider deflecting coils having no resistance or losses then, during the scan the voltage across the coils is constant since it is equal to L x rate of change of current and the rate of change of current is constant if the scan is linear. A portion of the scan is shown in figure The direction of the current and voltage FIG PRINCIPLE OF EFFICIENCY SCAN CIRCUIT across the coils is as shown in figure 5.13(a), i.e. the voltage across the coils is tending to make the anode of the valve negative. This continues during the portion of scan AB. At B energy is stored in the magnetic field produced by the coils. When point B is reached the valve is cut off and the current i decreasing FIG CURRENT IN, AND VOLTAGE ACROSS, LINE DEFLECTING COILS DURING SCAN AND FLYBACK (a) Scan; (b) Start of Flyback; (c) Flyback when current has reversed now starts to decrease and a voltage is produced (induced) in L (by the change of magnetic field) tending to maintain the flow of current. The direction is shown in figure 5.13(b), the voltage across the coils now being reversed so making the anode of the valve highly positive (tending to cause it to conduct).

88 LINE OUTPUT STAGE 83 This continues until the current is reduced to zero at C (figure 5.12). At C there is no energy stored in the magnetic field of the coils because there is no magnetic field. All the energy is transferred to the capacitance C. Thus, we have recovered the energy from the coils or, more correctly, from the magnetic field of the coils, since C is mainly formed by the capacitance of the coils. The energy has, therefore, been transferred to the electric field of C. The capacitance C now starts to discharge and the current in the deflecting coils increases in the opposite direction as shown at (c) of figure Eventually, the current will reach its maximum value at D. If there are no losses, the magnitude of the current at D will be the same as that at B, but the current is now flowing in the opposite direction. The energy has now been transferred back to the magnetic field of the coils and this magnetic field is of opposite direction to that at B. There is no energy in the capacitance C as there is no voltage across it at instant D. We have now removed the energy from the magnetic field of the coils and returned it back so that the current is flowing in the opposite direction, which is just what is required since, of course, in practice we always start the scan with the current in one direction, decrease this to zero and then increase it to the same value in the other direction. The current would now start to decrease if we allowed it to do so, but the decrease would be sinusoidal as shown dotted (figure 5.12) and the voltage would rise in a negative direction as shown in figure 5.14(a). Instead i decreasing + J- decreasinq FIG OPERATION OF EFFICIENCY SCAN CIRCUIT of allowing this to happen suppose that we allow the current to increase at such a rate that the voltage is just equal to the original value (V s) during scan, as at point E (figure 5.12). At this point suppose we connect a battery of voltage equal to that during scan (K s ) across the coil as shown in figure 5.14(b). If we assume that the battery has no internal resistance then the voltage across the coils L and across the capacitor C is fixed. Since the voltage is constant no current can flow through C and the current flows through the battery as shown. Also, since the voltage is constant, the rate of decrease of current in L must be constant in the same way as it was during the portion of the scan a constant rate of rise or fall of current results in a fixed voltage, A to B (i.e. also a fixed voltage means that the current must rise or fall at a constant rate). The current will therefore decrease at a constant rate until it is zero at point F (figure 5.12), when the switch S is opened. The energy that was stored in the will be noted that coil at D has now been transferred to the battery since it the current is in such a direction as to charge the battery not discharge it. Thus, we have recovered all the energy originally put into the scan coils during the half scan AB. At F the line output valve is made to conduct and produce the second half of the scan, corresponding to AB. The valve is now only required to conduct for half of the time so the energy output is less than for the normal circuit. It can be shown that this is a quarter of that required by the normal circuit. (It is a quarter because the magnitude of current is one half and it only flows for half the time). As well as this we have, of course, recovered the energy by charging the battery.

89 84 TELEVISION SERVICING EFFICIENCY PLUS BOOST CIRCUIT The energy that we put into the battery can be used to help with the second half of the scan. The arrangement is shown in figure During FIG OPERATION OF EFFICIENCY SCAN CIRCUIT the second half of scan, S is open and by connecting the h.t. + to the point shown, the voltage across the battery is added to the h.t. voltage and so the energy that was put into the battery during the first half of scan (i.e. EF of figure 5.12) is used to help to feed the line output valve during the second half of scan corresponding to AB. If the circuit is designed correctly the battery just feeds out the same amount of energy on the second half of scan as is put into it during the first half. [It is interesting to note that if we could make the resistance of the line output valve zero when it was conducting so that it had 100 per cent, efficiency, no h.t. supply would be required, for the energy in the battery would be sufficient to build up the current in the deflecting coils again (assuming no losses in the coils). This is far from the case in practice, but the voltage across the battery does make a substantial contribution to the power fed to the line output valve]. This method of boosting the h.t. is always used as it increases the efficiency of the circuit. Even without this boost arrangement, however, there is a large improvement in the efficiency of the circuit compared with that described earlier, therefore it is known as an efficiency scan circuit. PRACTICAL CIRCUITS It is obvious that a switch cannot be used as shown in figure 5.15 and it is not desirable to use a battery. The switch S is replaced by a valve (in commercial television receivers this is a diode but it could be a triode) and the battery is replaced by a capacitor (sometimes with the addition of a resistor). A simple circuit without boost is shown in figure 5.16(a). During the second half of scan the voltage and currents are as in the figure. V^ is conducting and the current in L is increasing. The voltage is such that V 2 is conducting, therefore Ci is charged to voltage V s and a small constant current flows in R u During flyback from B to D (figure 5.12) the voltage and currents are as shown in figure 5.16(b). V± is cut off and the voltage across L is now reversed and, therefore, V 2 is non-conducting and C t will discharge to some extent through jr t. At point E (figure 5.12) the conditions are as shown in figure 5.16(c). V t is still cut off but as soon as the voltage across L exceeds that across C t (due to the charge remaining) the valve V 2 conducts and connects the relatively large capacitor d across the coils. This maintains the voltage approximately constant during the period when the current in L decreases to zero. This corresponds to the portion E to F in figure During this period the current in L flows into d and tends to charge it i.e. the energy in L is fed into C x ; but, since C x is large, the voltage does not change appreciably. At point F, K t conducts again and the process is up,

90 LINE OUTPUT STAGE 85 i decreasing or increasing m opposite direction HI.+. I i decreasing FIG EFFICIENCY DIODE CIRCUIT (a) During second half of scan; (b) Sure of flyback; (c) During first half of scan repeated. The energy put into C 1 is dissipated in R t but this is about one quarter of that of the normal low efficiency circuit. By connecting the h.t.+ as shown in figure 5.17 the voltage across C x added to the h.t. and applied to the output valve K t. is Provided the circuit is + P. FIG EFFICIENCY DIODE AND BOOST CIRCUIT so designed that as much energy is fed out of C, to feed V x as is fed into it during the first half of the scan, the resistor R t is not required, and the most efficient operation then results. This means that the average current in V 2 must equal the average current in V t. It is not normally possible to connect the deflecting coils directly in the anode circuit of V t owing to the large voltage across the coils. It should be noted that no steady current flows in L in figure 5.17 and, apart from the high voltage, the circuit would be possible, a transformer not being theoretically essential as in the normal low efficiency circuit. In practice, because there are losses in the circuit, there would be some steady current in L. Although the efficiency diode V 2 (as it is called) could be connected across the primary of the transformer the voltages are rather high and it is, therefore, connected

91 ' I 86 TELEVISION SERVICING v, r v. FIG EFFICIENCY DIODE CIRCUITS across the secondary. Two arrangements are shown in figure 5.18, it being immaterial in which order V 2 and C^Ri are connected. Various arrangements are possible when a boost circuit is used and they are shown in figure In all cases it must be remembered that current flows through the valve V 2 from anode to cathode and, using this information, it is easy to see which way round capacitor Cj will be charged. In figure 5.19(a), which corresponds to figure 5.18(a), the capacitor will be charged as shown. By connecting the positive side of the capacitor to the primary of the transformer and the h.t.+ to the negative side of the capacitor, the voltage If the across d is added to the h.t. and used to feed the line output valve V x. during scan FIG EFFICIENCY DIODE AND BOOST CIRCUITS mean current in V t is equal to the mean current in V 2 then resistor R 1 is not necessary, since the charge put into C x through V 2, during the first half of scan, will be used to feed V t during the second half of scan. The voltage on the cathode of V2 now becomes important. In this case it is boosted h.t. (i.e. h.t. voltage plus the boosting voltage across Q) plus the voltage across the deflecting coils during flyback. During flyback (when, of course, maximum voltage occurs in the circuit) the voltage across the deflecting coils is such that the lower end is positive (so as to make V 2 non-conducting). Hence, it will be seen that the voltage across the coils adds to the boosted voltage at point X and results in a large cathode voltage on V 2. Precautions must be taken to

92 LINE OUTPUT STAGE 87 ensure that this does not exceed the permissible cathode-heater voltage of 2 An alternative arrangement is shown in figure 5.19(b) (corresponding to figure 5.18(b) ). Again, the voltage across d adds to the h.t. to supply K t. If the mean currents of V t and V2 are equal then R t is not required. In this circuit the cathode is at boosted h.t. voltage which is a considerable improvement on figure 5.19(a). A modification of (b) in this figure is shown at (c) where the capacitor C\ is returned to chassis, instead of to the lower end of the deflecting coils. If one considers no voltage in the deflecting coils, then d would charge up through V 2 to the h.t. voltage. When a scan voltage is induced in the deflecting coils this is added to the h.t. supply as shown, and is charged to h.t. + plus the deflector coil voltage, or what has been called, d the boosted h.t. voltage. The voltage across d is «sed to feed V x during the second half of scan. The mean currents of K, and V 2 should be equal. As in figure 5.19(b) the cathode voltage of V 2 is boosted h.t. voltage. A rather different arrangement is shown in figure 5.19(d). Instead of adding the boosting voltage in series with the h.t. positive lead it is now added in series with the negative lead to K i.e. in the cathode circuit as shown. The total voltage applied to V r is h.t. plus the voltage across d- In this case the cathode of V 2 is at earth potential, but the cathode of V^ is at boosting voltage negative with respect to chassis. In this case the boosting voltage is also applied to the screen circuit of V. x In all arrangements the steady d.c. component of current in Vi flows in the secondary of the transformer and in the deflecting coils, the ratio of the two currents being determined by the relative resistances of these two circuits. Any direct current flowing in the deflecting coils will cause the picture to be moved sideways and it is usual to place a capacitor in series with the coils to prevent this effect. The capacitor is made sufficiently large to have no appreciable effect on the scan current in the coils. It should be remembered that the lower the losses in the deflecting coils the better: for ideal operation the losses in the deflecting system should be zero Unless the losses are small an efficiency scan circuit is not worth while, as the amount of scan that can be obtained by the efficiency diode becomes a small fraction of the total scan. The losses in the deflecting system have been reduced by the use of Ferroxcube as the core material for the coils and the line output transformer. The use of this material is universal to-day. Circuits of this type were commonly used in the early sets where efficiency scan circuits were incorporated, but an autotransformer is now universally used. The principle of this is identical and will be considered later. CONTROL OF DIODE CURRENT The simple arrangement described does not give a linear scan and some means of correcting for this is usually necessary. Consider the operation of an ideal circuit as shown in figure 5.20(a). From A to B the efficiency diode conducts and produces the first half of the scan, while the line output valve produces the second half B to C. During the flyback C to D both valves are cut off. During the portion A to B the current is flowing as shown at (b) and it is assumed that the voltage across C is constant. So long as there is no resistance in this circuit the current must decrease at a constant rate. This is due to the fact that the induced voltage in Z, is: L x rate of change of current. As there is no other voltage in the circuit the induced e.m.f. in the coil L must equal that across Ci and, therefore, if the voltage is constant, the rate of change of current must be constant. This means that the current must decrease to zero in a linear way as shown in figure 5.20(a), resulting in a

93 r j 88 TELEVISION SERVICING, T Induced Voltage Conducting Induced Voltage FIG EFFECT OF CIRCUIT RESISTANCE (a) Waveform; (b) Circuit with no resistance; (c) Circuit with resistance linear scan. Unfortunately, in practice there is always some resistance in the circuit. The coil L will have some resistance and, more important, the diode will act as a resistance even in its conducting direction. The valve will be assumed to have a constant resistance, although in practice its resistance does vary with the ourrent. The circuit now becomes as shown in figure 5.20(c) where R represents the total resistance of the circuit. The induced voltage must now equal the voltage across C (constant) plus that across R. As the current decreases the voltage across R decreases and so the induced voltage must decrease. This also means that the rate of change of current must decrease and, instead of obtaining a linear decrease, the decrease will be curved as shown dotted in figure 5.20(a). This would, of course, result in non-linearity particularly near the centre of the screen. Correction may be obtained by allowing the line output valve to start conducting before B so that the overall current change is linear, as shown in figure The CURRENT OF V, A CURRENT OF V 2 FIG ONE METHOD OF REDUCING NON-LINEARITY difficulty is that the earlier the line output valve conducts the less will be the efficiency of the circuit. Other means may, therefore, be used in modern circuits where efficiency is most important. Another method of overcoming this non-linearity due to the resistance of the circuit is to control the current through the diode V 2. In other words, instead of allowing the current in the diode to decrease in a curve as shown in figure 5.21 we inject an e.m.f. in series with it so as to control the current and force it to decrease linearly. The basic idea is shown in figure 5.22(a) which shows an efficiency circuit without boost. This is as figure 5.18(b) except for the addition of the transformer L^L^, capacitor C 2 and resistor R 2. L lt L 2 form a transformer with L 2 tuned by C 2, but heavily damped by R 2. During the second half of the scan, when V t is conducting, L 2 is wound in such a direction that the secondary voltage tends to cut off V 2 (polarity as

94 LINE OUTPUT STAGE 89 Hflybock FIG (a) CIRCUIT FOR ELIMINATING EFFECT OF RESISTANCE OF CIRCUIT; (b) WAVEFORM shown) since we do not desire any current in V 2 during this period. When V 1 is cut off, a damped oscillation occurs in the L 2,C 2,R 2 circuit but, by suitable choice of R 2, this circuit can be damped so that the voltage across the circuit is as shown in figure 5.22(b). The voltage is seen to reverse (opposite to that shown in figure 5.22(a) ), hence the voltage is tending to send current through the diode. The voltage decreases towards zero at the end of the first half of scan. In other words the voltage is very similar to the resistance drop in the circuit (which is a maximum at maximum current at the start of scan and decreases to zero when the current in V 2 is zero) except that it is in the opposite direction [see figure 5.20(c)], hence cancels out the effect of the resistance. The diode current is, therefore, controlled so that it decreases at an approximately constant rate resulting in a linear scan. If h.t. boost is required the h.t. supply [figure 5.22(a)] is taken to point Y, and point A" feeds the line output valve V 1 through L^ and T. When h.t. boost is used (as is general) the transformer formed by L x and L 2 is usually made as an autotransformer and the circuit is then as shown in figure L 2, C 2 and R 2 form the damped tuned circuit and L t forms the primary winding of the transformer. L 2 is adjusted by movement of a Ferroxcube core to adjust linearity. An alternative arrangement is shown infigure 5.24 when the action FIG CORRECTION CIRCUIT TO OVERCOME THE EFFECT OF CIRCUIT RESISTANCE ALTERNATIVE CORRECTION FIG CIRCUIT TO OVERCOME THE EFFECT OF CIRCUIT RESISTANCE is similar to the above, the resonant circuit being formed by L u C x and C 2, damped by R2. L± is adjustable and forms the linearity control. Another method of linearity control is the use of a small coil on a Ferroxcube core in series with the deflecting coils. The core is saturated by a small permanent magnet (or by a winding carrying d.c). When the core is saturated the flux in the winding does not change and there is no induced voltage in the winding. When the core is not saturated the

95 90 TELEVISION SERVICING flux changes rapidly in the core and induces an e.m.f. in the winding, the e.m.f. being in such a direction as to oppose the change of current flow. Thus, if the permanent magnet is in the correct direction and of correct magnitude, the core will remain saturated until the current in the coil produces an m.m.f. almost equal and opposite to that of the permanent magnet. When this occurs an e.m.f. opposing the current change will be induced in the coil. If this is made to correspond to the part of the scan where the rate of current change is greater than that required, this back e.m.f. will tend to reduce the rate of change and so improve the linearity. Another method of linearity control used is the shorted loop placed under the line deflecting coils and is known as linearity correcting coils. Two loops are used, one under each line coil. The loops, which are usually copper foil, are moved in and out relative to the deflecting coils to adjust the amount of correction. During the scan an e.m.f. will be induced in the shorted coil and it will be constant if the scan is linear. As a result of this e.m.f. a current will flow in the coil and will rise exponentially as the shorted coil has resistance and inductance. The time constant is small compared with a cycle and hence the current rises fairly rapidly at the start of the cycle to an approximately constant value. The m.m.f. produced by this current opposes the m.m.f. produced by the scanning coils. This is an application of Lenz's law. The rapidly increasing m.m.f. at the start of the scan tends to reduce the rate of rise of flux but later in the scan the opposing m.m.f. is approximately constant and has little effect. Thus the initial rate of rise of m.m.f. is decreased which is what is required to make the scan more linear. The effect is shown in figure This is a NON- LINEAR SCANNING WAVEFORM I IN SHORTED COILS FIG EFFECT OF SHORTED LOOP ON LINE LINEARITY simple explanation of a rather complex matter and it is not possible to go further into it here. USE OF AUTOTRANSFORMER Up to now we have considered the use of a double-wound line output transformer, a transformer being essential to reduce the voltage on the deflecting coils. The main difficulty of a double-wound transformer is the leakage flux between the primary and secondary, which causes damped oscillations to occur and upset the ideal operation of the circuit. The leakage flux is reduced by using an autotransformer, also some saving of copper results.

96 LINE OUTPUT STAGE 91 Polarity flyback J ^'±-m* BOOSTED H,T. VOLTAOE FIG AUTOTRANSFORMER LINE OUTPUT CIRCUIT (a) Double-wound transformer; (b) Autotransformer; (c) Autotransformer with efficiency diode fed with larger voltage than deflecting coils; (d) As (c) but redrawn; (e) with Ci connected to chassis instead of h.t. + Suppose that we consider the basic boost circuit shown in figure 5.26(a) (same as figure 5.19(a) ). The corresponding autotransformer circuit is shown in figure 5.26(b). In order to keep down the voltage on the deflecting coils L, but obtain a high boost voltage, the arrangement shown at (c) is usually used. The deflecting coils are fed from points a and b of suitable voltage, while the efficiency diode and boost circuit are fed from points a and c which are farther apart and, therefore, produce a higher voltage. Point c may be at the anode of the line output valve so that the full voltage is available for the boost circuit. Figure (c) may be drawn differently as in figure (d) although the actual electrical circuit is the same. Instead of connecting the capacitor C\ to h.t.+ it may be connected to chassis as at (e) so that the total boosted h.t. voltage is now applied across C t rather than the boost voltage itself. This is similar to the modification in figure 5.19 from circuit (b) to that at (c). The difficulty with this circuit is that during flyback the cathode of the efficiency diode is at a high voltage. It is equal to boosted h.t. plus the

97 92 TELEVISION SERVICING flyback voltage across a-c which has been shown to be in such a direction as to add to the boosted h.t. voltage. If c corresponds to the anode of V x then the cathode voltage is the same as the anode voltage of the line output valve, up to (say) 7kV. Even if c is only a portion of the way down the winding between b and the anode of the line output valve V lt the cathode voltage of V 2 is still large. It should be noted that this is also the case for the double-wound transformer when used in the circuit of figure 5.26(a). But, with the double-wound transformer, other circuits can be used [e.g. those of figure 5.19(b) and (c)] but these are not possible with the autotransformer. If a normal valve is used for the efficiency diode V 2, this high voltage will break down the insulation between the heater and the cathode. This difficulty also arises in the case of the e.h.t. rectifier (see later in this chapter) which is overcome by feeding the heater from a winding on the line output transformer. This is possible because the heater power is small, but cannot be used for the efficiency diode as the heater power required is too great. The heater may be fed off a special filament transformer (on a.c. mains only) but the transformer is difficult to design and is costly. One method of overcoming the difficulty is to use a bifilar winding. This is shown in figure The heater is fed from the normal heater chain To heater chain FIG USE OF BIFILAR WINDING TO FEED HEATER OF EFFICIENCY DIODE (a) Circuit; (b) Equivalent arrangement through two windings x and y which are wound together (hence the name "bifilar") and in the same direction. The turns on x and y are the same as those between points p and q. Point p is at boosted h.t. potential. On flyback a voltage is induced between p and q in the direction shown so that the cathode of V 2 is driven positive. A similar voltage will be induced in x and y driving the heater positive by the same amount, hence the flyback voltage does not appear between the heater and cathode. The arrangement is like that shown at figure 5.27(b). The two batteries x and y cause the heater to be at a positive voltage with respect to earth but, as regards the voltage across the heater, they cancel out. Capacitor C [figure 5.27(a)] is added to prevent ringing {i.e. a damped oscillation) of these windings and causing a surge voltage to appear between the heater and the cathode. This arrangement allows the use of a normal efficiency diode to be used but is expensive and not used in modern receivers. Instead of reducing the voltage between the heater and the cathode the present method is to use a valve which will withstand this high voltage. Valves are available which will withstand a voltage of 6-6kV between heater and cathode (under these operating conditions only). In order that the valve will withstand this high voltage a ceramic sleeve is fitted between the heater

98 LINE OUTPUT STAGE 93 and the cathode. Unfortunately, this has the effect of increasing the time (approximately double) of heating the cathode. Since the anode of the valve is at a relatively low potential and the cathode at a high potential the latter is brought out to the top connection, while the anode is brought out at the base with the heater leads. This overcomes the voltage difficulties of the circuit and the arrangement is now almost universally used. The voltages produced in modern line output stages are high owing to the difficulties of producing the wide angle deflection now required. On flyback the voltage on the anode of the line output valve is about 7kV and on the cathode of the efficiency diode up to 6-6kV. A voltage of l-5kv may be developed across the line deflecting coils which must be suitably insulated with polythene insulation. The ratio of the line output transformer is now rather low (say) 2/1 to 5/1. The amount of overswing obtained in modern deflecting coils (i.e. the ratio of the current at D to that at B [figure 5.12]) is 0-9 to 0-95 due to the low losses now possible in the coils. The amount of boost voltage obtained from the efficiency diode is large (say) up to 600V, resulting in a boosted voltage (applied to the line output valve) of 600 to 800 volts. PRODUCTION OF E.H.T. VOLTAGE FOR CATHODE RAY TUBE This is considered in this section as the e.h.t. is now always obtained from the line output transformer. In very early receivers the e.h.t. (then about 5kV) was obtained from a mains transformer, half- wave rectifier and resistance-capacitance smoothing circuit. Apart from the voltage the circuit was conventional, but as it is not used today we will not go into further details. The disadvantages are: (0 Insulation difficult on transformer. (if) Transformer expensive, bulky and heavy. (Hi) Smoothing capacitor is also expensive, bulky and heavy (01/ttF, 5kV working). (iv) It is difficult to protect the circuit should a fault occur, such as a short circuit on the e.h.t. lead. (v) Large power available so that circuit may be dangerous if touched. These serious disadvantages now prevent its use in television receivers. An alternative arrangement is the use of an r.f. oscillator which, as it is not used in commercial receivers, will be considered only briefly. The arrangement consists of an r.f. oscillator operating at about 50kHz. The tuned circuit is made in the form of a step-up transformer so that a large r.f. voltage is obtained. This is rectified by a half-wave valve rectifier and then smoothed. The arrangement is good but expensive compared with the method to be described. We have already seen that during flyback a large voltage is produced across the primary of the line output transformer, making the anode of the line output valve highly positive. This voltage pulse may be up to 7kV. By rectifying this pulse and smoothing it, an e.h.t. voltage would be obtained. This is not sufficient for the operation of a modern tube and, although voltage doubling circuits have been used they are not used in modern monochrome receivers. Instead, an extra winding is placed on the transformer to step up the flyback pulse. This is shown in figure 5.28 from which, for simplicity, the efficiency diode has been omitted. The normal primary winding a-b (or the section of the autotransformer acting as the primary) is extended by the addition of winding b-c. Since the winding is in the same direction a larger voltage pulse will be produced at c than at b, the magnitude depending on the ratio of the turns between a and b and between b and c. o

99 : 94 TELEVISION SERVICING FIG FLYBACK E.H.T. CIRCUIT This pulse must now be rectified, which is achieved by a high voltage rectifier, V 2, specially designed for this purpose. Since the heater of this valve is at e.h.t. potential it cannot be fed from the normal heater chain. Instead, it is fed from a winding on the line output transformer. The power required for the heater must, of course, come from the line output stage, but the heater power required is small (say) 6-3V at 90mA = 0-567W. The output from the rectifier must be smoothed by a capacitor, Cj (500 to l,o00pf). The operation of the rectifier and smoothing capacitor is shown in figure Output Voltoqe b _/ d FIG OPERATION OF FLYBACK E.H.T. CIRCUIT The rectifier only conducts between a-b and c-d, and, during this period charges up the capacitor to almost the peak value of the pulse. Between pulses (i.e. between b and c, etc.) the capacitor supplies the cathode ray tube. If the capacitor is large enough the drop in voltage between b and c is small. The capacitor Q may be a high voltage capacitor but, more commonly, it is the capacitance between the inner and outer graphite coatings of the cathode ray tube. The use of a single capacitor (or the capacitance of the cathode ray tube) in this way is sufficient smoothing, largely due to the high frequency (10,125Hz or 15,625Hz) fed to the rectifier from the line output transformer. This method of obtaining e.h.t. has the following advantages (0 Cheapness. (n) Not lethal as the power output is limited and no large smoothing capacitors are used. (hi) Unlikely to cause much damage due to a short circuit on the e.h.t. as the power output is very limited. (iv) May be used with a receiver operating on a.c. or d.c. supply, (v) Less weight and space than 50Hz supply, (vi) Smaller smoothing capacitor due to higher frequency. (vii) E.H.T. soon drops when set switched off, due to the use of small smoothing capacitor. (yiii) E.H.T. not applied until the cathode ray tube warms up. The e.h.t. is not available until the efficiency diode is conducting and this is the last valve to reach operating temperature (when using a valve with a high voltage rating between heater and cathode).

100 : LINE OUTPUT STAGE 95 The disadvantages are not serious and are (/) Regulation not too good but depends on design. (h) Line amplitude control may vary e.h.t. (iii) E.H.T. not available for test if line output stage or timebase is faulty. The foregoing method may be used to obtain e.h.t. voltages up to 15kV or more, and is universally used in modern television receivers. A Metrosil rod (VDR) is sometimes placed across the e.h.t. line to improve the regulation of the e.h.t. supply. The use of a Metrosil to improve the voltage regulation has been described in connection with the stabilization of field timebases. In this case the series resistor is not a physical component but is the internal resistance of the e.h.t. supply. As well as the above e.h.t. to feed the final anode of the cathode ray tube a voltage is also required to feed the first anode when a tetrode tube is used. The voltage required is V above the cathode voltage of the cathode ray tube. This is usually obtained from the boosted h.t. line, often through a simple R-C filter circuit. The current required is extremely small. In modern electrostatically focused tubes a voltage of 20O-600V is required for the first anode and is generally obtained from the boosted h.t. line. The third anode may also be fed from the boosted h.t. voltage through a suitable potential divider, which may be made variable to form a focus control. The boosted h.t. is also used in some receivers to feed the field timebase. In some older receivers a Metrosil disc was used to feed the first anode of the tube, the Metrosil acting as a rectifier (due to the fact that it is fed with pulses and not a symmetrical sine wave), but this arrangement is rarely used now. More details about the use of a Metrosil, or a voltage dependent resistor (VDR), as a rectifier are given in the section on stabilization. CONTROL OF WIDTH Some method of controlling the width of the picture is necessary. If the output of the timebase is varied this will also vary the e.h.t., so that this cannot normally be used. Some method of controlling the current in the actual deflecting coils must be used. There are three ways of doing so and they are: (i) A variable inductor in series with the deflecting coils as shown in figure 5.30(a) (overleaf). (h) A variable inductor across a section of the transformer, so shunting some of the current from the deflecting coils as at (b). Both these upset the operation of the line output stage as they alter the effective inductance of the line output transformer. (iff) Series and shunt coils as shown at (c). The inductor is varied by altering the position of the Ferroxcube core. As the core moves out of L t it moves into L 2. Thus, more current flows in, and less in L 2 and the deflecting coils, but the effective inductance of the transformer remains approximately constant. A damping circuit consisting of a resistor, or a resistor and capacitor, may be used across the width adjusting inductor, to reduce any damped oscillation set up in it. An alternative arrangement is shown at (d) which operates in the same way as that at (c). In this case the two coils are electrically isolated. When a stabilized output stage is used other methods are possible (see later). TYPICAL LINE OUTPUT STAGE ««., A typical line output stage for a 90 tube is shown in figure The line output valve K, is fed from the timebase through C t (0-OOlftF). The screen grid is fed through R2 (2-2kfi) and by-passed by C 2 (300pF). The autotransformer 7\ feeds the deflecting coils L through the width control L lt

101 96 TELEVISION SERVICING Width Control 7T WIDTH CONTROL DEFLECTING COILS LINE OUTPUT TRANSFORMER FIG WIDTH CONTROL CIRCUITS (a) Series inductor; (b) Shunt inductor; (c) Series and shunt inductors; (d) Series and shunt inductors electrically isolated FIG TYPICAL LINE OUTPUT STAGE FOR 90 TUBE damped by R 3 (1-5kO) and the linearity control L 2, damped by R t (1-2kO). L 2 is provided with a permanent magnet to saturate the core. The transformer also feeds the efficiency diode V 2, the boosting voltage being obtained across the capacitor C 3 (0-25/x.F). The boosted h.t. appearing at A" is smoothed by R5 (220k 2) and C 4 (0-01ju,F) and used to feed the first anode of the cathode ray tube. E.H.T. is obtained from V 3 which is smoothed by C 5, the capacitance of the cathode ray tube. It should be noted that no bias is provided on the line output valve apart from that which is obtained by self-bias when a signal is applied (i.e. due to the flow of grid current as in an oscillator). Although we have assumed that a sawtooth waveform is fed to the grid of the line output valve this is not essential. In many cases a negative pulse is fed to the grid so as to cut the valve off during flyback. The line output stage produces a sawtooth current due to the rise of the current in the

102 LINE OUTPUT STAGE 97 inductanc.e formed by the primary of the line output transformer. This action is similar to that of the sawtooth current generators which are considered in the next chapter. 110 SCANNING CIRCUITS Although no fundamental changes are made regarding the line output stage, various alterations are necessary when a 110 scanning angle is used. These are considered below. S Correction We have assumed that the deflecting current in the scanning coils should be a linear sawtooth waveform but in fact this is not so. This requirement is necessary for all cathode ray tubes but it becomes more important when tubes are used with large deflecting angles and "flatter" faces. CENTRE OF SCANNING FIG S CORRECTION (a) Effect of equal deflecting angles on a small angled tube; (b) Eflect of equal deflecting angles on a large angled tube; (c) Change of angle necessary to maintain linear deflection on a large angled tube If the tube face was curved as in figure 5.32(a) so that its radius was such that the centre of this radius corresponded to the centre of the scanning (i.e. approximately the centre of the scanning coils) then equal angles 6U 62, etc. would result in equal displacements a x, a 2, etc. on the screen and a linear scanning current would result in a linear picture. In older tubes using a small deflecting angle this was approximately true. When the angle is increased and the tube face is made flatter then the condition is as at (b). If we now make Q\ = 2 = #3 then from the diagram it is obvious that a t will not equal a z and that a 2 will not be equal to a 3. In other words, for a given deflecting angle 6 (which is approximately proportional to the deflecting current) the

103 98 TELEVISION SERVICING movement of the spot is greater at the edges of the screen than at the centre. Obviously, for a linear scan the distances a lt a 2, a 3, etc. must all be equal as at (c). This means that d t must be greater than 8 2 and that 6 2 must be greater than 6 3, etc. Assuming that the angle of scan is proportional to the deflecting current then the deflecting current must increase at a smaller rate towards the edges of the screen. Thus, it must have a shape as in figure 5.33(a), i.e. it must be shaped like the letter S rather than linear. It is known as S correction. DEFLECTING CURRENT DEFLECTING COILS FIG S CORRECTION (a) Waveform of scanning; (b) Derivation of S correcting waveform The method of achieving this result is to place a capacitor of suitable value in series with the deflecting coils which has the effect of reducing the rate of current rise at the beginning and at the end of the scan. The required value will be different on 625 lines than on 405 lines. The value of the capacitor required is approximately 01 to 0-3/^F. S correction operates by resonance between the deflecting coil inductance and the correcting capacitor. The circuit is pulsed into a damped oscillation by the flyback and it will be seen that the shape of the curve required is part of a sine curve. By arranging that the resonant frequency is 1/3 to 1/2 of the line frequency a suitable portion of the sine curve is used as shown in figure 5.33(b), i.e. only the portion A-B is completed in the time of the scan. By reducing the value of the capacitor and increasing the resonant frequency the portion of the sine curve completed during the time of the scan is increased and the amount of S correction is therefore increased. In a typical case with 5mH deflecting coils when a capacitor of 0-12//.F is used on 625 lines, the resonant frequency works out to be 6,500Hz compared with a line frequency of Hz. Third Harmonic Tuning This principle was used on some 90 scanning circuits but is more important in 110 scanning circuits where greater scanning power is required. Without third harmonic tuning it was usual to wind the overwinding (i.e. the e.h.t. winding) on the same limb as the main winding. The main winding was made broad so that the leakage inductance was kept to a minimum. The e.h.t. winding, however, was narrow, it being necessary to keep down the capacitance and make the insulation of the winding easier.

104 i I t i 1 I 1 I 1 I 1 I LINE OUTPUT STAGE 99 Owing to the shape of the e.h.t. winding, considerable leakage inductance resulted between the two windings (due to flux which cuts one winding but not the whole of the other as illustrated in figure 5.34). This leakage inductance E.H.T. -~ WINDING MAIN- FLUX 4 T CORE ' v - MAIN ^ WINDING u. 1 I, LEAKAGE FLUX FIG LEAKAGE FLUX IN LINE OUTPUT TRANSFORMER may be represented in an equivalent circuit by an inductance in series with the main winding, similar to an external resistor to represent the internal resistance of a cell. An approximate equivalent circuit of the line output transformer is shown in figure 5.35, h x representing the main winding together with deflecting coils. FIG APPROXIMATE EQUIVALENT CIRCUIT OF LINE OUTPUT TRANSFORMER SHOWING OVERWINDING AND LEAKAGE INDUCTANCE This winding has a capacitance C lt which may be stray capacitance or partly composed of a physical capacitance. This capacitance controls the rate of flyback. The overwinding is L 2 with the leakage inductance represented by inductance L 3. Associated with L 3 is its stray capacitance C 3. The capacitance C2 represents the capacitance of the e.h.t. rectifier, etc. connected to the live end of the e.h.t. winding. The main winding is, of course, connected to the line output valve which can be represented by the battery B with switch S. During the scan switch 5 is closed and current builds up in,. At the start of flyback S is opened and a high voltage is induced in h x and L 2 : as has been seen earlier this is a half sine wave. The direction of these induced e.m.f.s is shown by the arrows. The sudden voltage applied to L 3 (note that no voltage is induced in L 3 as it is a leakage inductance, and flux in L t and L 2 does not cut L 3) causes a damped oscillation at a frequency determined by L 3 C 3. If the magnitude of this oscillation is large it extends into the scan and causes striations on the picture because it causes variations in the scan speed. This is due to the fact that part of the voltage across L 3 appears across L t and the deflecting coils. Apart from this trouble the energy in the circuit is a loss as

105 100 TELEVISION SERVICING it is not recovered by the efficiency diode circuit. In modern line output to approximately transformers use is made of this oscillation by tuning L 3 C 3 three times the resonant frequency of Z,iC x. The voltage across L 3 C 3 is applied partly across C 2 and partly across L x Cx but it is applied in opposite senses to C 2 and L X C X. Assume that the voltage across L 3 is in the direction shown. At any instant it will tend to increase the voltage across C 2 but decrease that across Z-iQ. The effect is shown in figure It will be seen that by tuning the leakage reactance to RESULTANT VOLTAGE APPLIED TO E.H.T. RECTIFIER VOLTAGE ACROSS Tima VOLTAGE ACROSS Lj.Cj VOLTAGE ACROSS L,,C (wirtiout «ff«ct of Lj.Cj) RESULTANT VOLTAGE ACROSS L,,C, EFFECT OF VOLTAGE ACROSS L 3,C 3 FIG THIRD HARMONIC TUNING about three times the resonant frequency of L x Ct the voltage across L 3 C 3 is applied to the e.h.t. rectifier (i.e. across C 2) in such a direction as to increase the e.h.t. voltage. On the other hand the voltage across L 3 C 3 reduces the voltage across LiC u i.e. on the anode of the line output valve which is obviously desirable. The voltage across L 3 C 3 is not quite sinusoidal owing to the capacitance of the circuit being charged and discharged by the flyback pulse and the resonant frequency of L 3 C 3 is normally made about 2-7 times that of InQ. Also, since the voltage across L 3 C 3 is zero at the end of the flyback there are less chances of striations on the scan. The voltage across L 3 C 3 may appear to be the opposite to that expected. The voltage is rising across L t and L 2 at the start of the flyback but it must be remembered that there is no voltage induced in L 3. The circuit L 3 C 3 is shock-excited by the voltage applied across it and if the voltage across L 2 is rising then that across L 3 will tend to decrease and hence the voltage across L 3 C 3 is in the phase shown in figure A suitable leakage inductance is obtained by having the main winding and the e.h.t. winding on different limbs of the core of the transformer. The leakage inductance can be varied by using a coupling winding between, and L 2 and this will be referred to later in connection with 625-line operation. Stabilization In older designs of output stages, variations in supply voltage caused variations in scan power and e.h.t. Because the e.h.t. and scan power decrease together, the amplitude of scan tends to remain constant; but this does not apply to the field scan where any variations directly affect the scan

106 LINE OUTPUT STAGE 101 amplitude and, similarly, variations of e.h.t. cause variations of field scan amplitude. To correct this it is the practice in modern circuits to use a stabilized line output stage, which is really a type of automatic gain control. A basic circuit is shown in figure The line output valve K, is fed from the FROM LINE-*- c 2 -n- FROM DEFLECTING / COILS VDR + J_ FIG STABILIZATION OF LINE OUTPUT STAGE * line timebase through C 2, and R t is the normal grid resistor. A positive flyback pulse is fed through d from the line output transformer, and is applied across the voltage dependent resistor VDR. Let us now consider how a VDR may act as a rectifier when a pulse is applied to it. Take the basic circuit of figure 5.38(a) where capacitor d is fed through the VDR from a transformer with a waveform consisting of short positive pulses. Since there can be no d.c. component across the transformer the two shaded areas being equal. winding the waveform will be as at (b), Thus, the voltage applied in a positive direction is much greater than that in the negative direction. If the VDR were replaced by a resistor no direct voltage would exist across d- The resistor and d would act purely as a smoothing circuit and if we smooth the waveform of (b) the result is zero (since it has no d.c. component). This may be looked at in another way. During the period of the positive pulse, current would flow through the resistor to charge d- The current would be large (being proportional to the voltage) but its duration is short. During the remainder of the cycle the voltage is reversed and of much smaller amplitude, but it flows for a longer time. It might be expected that the effect of these two currents would be the same as regards the charge fed to d. This, in fact, is so since the product of magnitude and time (assuming that the pulse is rectangular for convenience) must be the same for positive and negative portions because these products are the shaded areas shown at (b). Thus the charge flowing into the capacitor during the positive pulse will equal that flowing out during the negative portion of the Therefore, there will be no resultant charge on the capacitor. cycle.!_l VDR tl 'i^esasssssssi FIG. S.38. OPERATION OF VDR AS RECTIFIER TO PULSES (a) Circuit; (b) Input waveform

107 102 TELEVISION SERVICING JCURRENT DUE TO ^ POSITIVE PULSE CURRENT DUE TO NEGATIVE PORTION OF CYCLE O VDR CHARACTERISTIC EQUAL AREAS. VDR CHARACTERISTIC FIG OPERATION OF VDR AS RECTIFIER TO PULSES (c) Operation (initially); (d) Operation final When the resistor is replaced by a VDR the operation is as in figure (c) During the positive peak a large current flows owing to the shape of the VDR characterisitic; but during the negative portion of the cycle an extremely small current flows. Due to the shape of the characteristic the current that flows is no longer proportional to the applied voltage, hence the charge fed into the capacitor during the positive pulse is much greater than that flowing out during the negative portion, and the capacitor charges up in the direction shown at (a). Obviously, this action cannot continue indefinitely and what happens is that the waveform is displaced to the left of the VDR characteristic as at (d) so that equilibrium is reached, and equal current flow into and out of the capacitor.! This displacement to the left is due to the charge built up across C x. Thus the capacitor charges up to some voltage which is less than the peak of the pulse. The magnitude of the output will depend on the pulse magnitude and its duration, and on the exact shape of the VDR characteristic. In figure 5.37 the capacitor and VDR are interchanged so that the circuit looks like that in figure The output is now the pulse plus the d.c. voltage across C, which is in a negative direction. The pulse is removed by the smoothing circuit (R x and C) and the resultant output is the negative voltage across C". In figure 5.37 the smoothing is by grid resistor R t and FIG BASIC LINE OUTPUT STAGE STABILIZING CIRCUIT

108 LINE OUTPUT STAGE 103 the charging capacitor C of the timebase and C 2 in series. The negative voltage across the VDR is applied as bias to V 1 but by itself it would be toelarge Accordingly, a positive voltage from a preset potentiometer P is fed through R2 so as to partly oppose the bias from the VDR. Thus, if the amplitude of the scan tends to increase then the flyback voltage is increased, and the bias increases thereby reducing the power fed out of the line output valve Potentiometer P is set to give the correct boost h.t. voltage or may be used as a width control. The voltage fed back is, of course, the pulse voltage across the line output transformer, but if the circuit values are constant this is proportional to the scanning current in the deflecting coils. In order that the valve output may be varied by this bias the valve must not be run below the knee of the characteristic, and this causes some loss of efficiency. Some variations of this circuit are possible but all use the same principles. DESATURATED LINE OUTPUT TRANSFORMER In a standard line output transformer the direct current taken by the line output valve flows through the winding in one direction and produces a d.c. magnetization of the core, as well as the normal a.c. component of magnetization. If the d.c. magnetization can be removed it is possible to use higher flux densities in the core and the size of the core can be reduced. A whistle is produced by the core of the line output transformer due to magnetostriction, i.e. a change in the dimension of the core with magnetiza- When the core has a d.c. magnetization the size of the core varies as the tion a c component varies, i.e. increases and decreases in size at the line output frequency {i.e. 10,125Hz or 15,625Hz). If the core is not saturated it will vibrate at twice the line frequency and therefore will be inaudible. This is because the core moves in the same direction for both negative and positive half cycles since the change of dimension does not depend on the direction of the magnetization (i.e. both directions of magnetization cause an increase in dimension). This effect is illustrated in figure 5.40 (overleaf). One method of overcoming this d.c. saturation or magnetization is shown in figure V x is the line output valve and V 2 is the efficiency diode. The capacitor C t forms the efficiency diode capacitor. If the h.t. is fed in at point X current will flow in the two sections of the line output transformer a-b andc-rf in opposite directions; and if the number of turns on the two sections are equal, the d.c. magnetization will be cancelled out. The point X cannot be fed directly as it is at an a.c. potential with respect to chassis and hence is fed from the h.t. through the isolating inductor,. The top of the line output transformer is maintained at a steady potential by the capacitor C 2, i e C 2 forms the a.c. short circuit path to h.t.+ or chassis. Alternatively, the feed for valve V 2 may be taken through a secondary winding as in figure The current for V 2 now flows in winding W which sets up an m.m.f. in the direction shown which opposes the m.m.f. in section c-d of the main winding. To prevent the induced voltage in W being applied to V 2 the inductor L x is added together with the smoothing capacitor C u so that L and d act as a smoothing circuit and all the a.c. voltage in winding W appears across inductor L x. 625-LINE OPERATION When a change is made to 625-line operation the frequency of the line timebase must be changed from 10,125 to 15,625Hz. The change of timebase frequency is easily done as it is only necessary to alter the value of the charging capacitor to the required value. So that the timebase may be accurately adjusted to the correct frequency for both systems it is usual to fit separate line hold controls and to switch them when changing from one system to the other.

109 104 TELEVISION SERVICING A.C. MAGNETISATION MAGNETISING M.M. F. j i DC. MAGNETISATION RESULTANT MAGNETISATION CHANCE OF SIZE CHANGE IN S12E (at line frequency) MAGNETISING M.M. F. A.C. MAGNETISATION CHANGE SI7F OF CHANCE IN SIZE (ar 2 X line frequency FIG EFFECT OF DESATURATION OF CORE OF LINE OUTPUT TRANSFORMER (a) With d.c. saturation of core (b) Without d.c. saturation of core L r b uj% ii tn *u fx T C2 BOOSTED., f. H.I. * VOLTAGE I 3 1 -DEFLECTING COILS ^ VALVE i«-line OUTPUT VALVE FIG. S.41. ONE METHOD OF DESATURATION OF CORE OF LINE OUTPUT TRANSFORMER

110 LINE OUTPUT STAGE 105 -HT.+ BOOSTED ^. H.T. VOLTAGE c 2 : EFFICIENCY DIODE VALVE LINE OUTPUT VALVE FIG ALTERNATIVE METHOD OF DESATURATION OF CORE OF LINE OUTPUT TRANSFORMER The main difficulty lies in the line output stage and this will be considered in terms of a dual-standard receiver. When changing from one system to the other there are two possibilities: (i) Use of constant flyback time; (ij) Use of constant flyback ratio. (0 Constant flyback time The scan time of the two systems is different which means that the ratio of scan to flyback time will also be different. The scan time on 405 lines = 80-5ju.S 625 = 520/iS Line pulse + back porch on 405 lines = 16-5/u.S 625 = 10-5/xS Unless part of the picture is to be lost, the flyback must take place in the period of the line pulse + back porch. To be constant this must be the shorter of the two periods, i.e. the 625-line system equal to 10-5/aS. In the 405-line system the ratio of scan time to flyback time is therefore 80-5/10-5 = 7-7. In the 625-line system the ratio of scan time to flyback time is 52-0/10-5 = 4-9. There is, therefore, quite a large change in the ratio. When the flyback time is maintained constant the voltages on flyback remain constant; but if the flyback is made as short as is required by the 625-line system the flyback voltages tend to be high. Similarly, if the flyback time is constant any third harmonic tuning will be correct on both systems. Figure 5.43(a) and (b) shows the waveform of the current in the deflecting coils in the two systems assuming that the flyback is linear and that the flyback time is that required on the 625-line system, i.e. 10-5/xS. Since the magnitude of current for the same amplitude of scan is independent of the frequency, the amplitude of the waveforms is the same. (This is not quite correct if the picture widths are to be the same, but near enough for the present purpose). Neglecting the resistance of the deflecting coils the waveform of voltage across the coils is shown at (c) and (d). The flyback time is the same so the rate of change of current during the flyback is the same, and the flyback voltage is constant. On the other hand the rate of rise of current is greater in the 625- line system and the voltage is greater. The ratio of scan voltage on 625 lines to that on 405 lines is 80/52 = This is the voltage applied to the e.h.t. rectifier heater, and it is obvious that voltage at (d) is larger than that at (c).

111 106 TELEVISION SERVICING Time ^y *=" FIG. S.43. CURRENT IN AND VOLTAGE ACROSS LINE DEFLECTING COILS (a) Current in coils on 405-line system; (b) Current in coils on 625-line system; (c) Voltage across coils on 405-line system; (d) Voltage across coils on 625-line system As regards heating it is the r.m.s. value which is important and it can be shown that the voltage on 625 lines is approximately 1-34 times that on 405 lines. Some arrangement is needed to keep the voltage on the e.h.t. rectifier heater constant although the voltage induced in the e.h.t. rectifier heater winding will vary. Because the heater winding is at a high potential (10 to 15kV) it is impractical to do any switching, and the arrangement used consists of an inductor or choke in series with the heater. The heater winding must now be designed to produce a voltage higher than normal to allow for the drop in the inductor when on 405 lines. When on 625 lines a higher voltage is induced in the heater winding; but due to the higher frequency the impedance of the inductor will increase and if the circuit is designed correctly the increased drop across the inductor will just compensate for the higher voltage across the heater winding. The inductor is normally placed in the e.h.t. rectifier valveholder. The voltage across the line output transformer is also shown in figure 5.43, as it is the same waveform as that induced in the heater winding. The voltage applied to the anode of the line output valve during scan is the boosted h.t. voltage, minus that across the line output transformer winding feeding the line output valve anode. For 625 lines the circuit must be so designed as to leave sufficient voltage across the valve and in the stabilized line output stage this voltage must be just above the knee of the valve characteristic. Due to this comparatively low anode voltage the anode dissipation is low (as explained earlier). When switching to 405 lines, unless some change is made the voltage across the line output transformer is less, and a larger voltage is applied to the anode of the line output valve during the scan. This means an increased anode dissipation which may be excessive. One may look at it in another way. On 405 lines less power has to be fed to the line deflecting coils, and unless conditions are changed more power is dissipated in the valve. This greater dissipation may be reduced by placing a resistor in series with the boost diode (with suitable decoupling) when on 405 lines, thereby decreasing the voltage

112 LINE OUTPUT STAGE 107 fed to the circuit. Owing to the reduction in the boost voltage it may be necessary to use switching in the supply to the first anode of the cathode ray tube as this should be maintained constant. As the S correction changes with frequency the value of the S correcting capacitor must be changed between systems. When changing over from one system to the other unless corrected the picture will be displaced sideways on one system relative to the other. This is explained with the help of figure At (a) is shown the 405-line conditions, " FIG EFFECT OF BLANKING PERIOD ON POSITION OF PICTURE In both systems the flyback will be and at (b) is shown the 625-line conditions. started by the leading edge of the line synchronizing pulse and the flyback time is the same on both systems. Assuming the flyback time to be the same as the time of the 625-line synchronizing pulse, plus back porch, the scan starts at instant B. On 625 lines the picture signal starts at this instant and hence this corresponds to the left-hand side of the picture (not raster). On 405 lines the picture signal does not start until instant C which is 6jxS later than instant B. Thus, C now represents the left-hand side of the picture (not raster). right-hand side is settled by the scanning and although it takes place at different rates the amplitude of scan is assumed to be the same, hence the spot will reach the same position on the right-hand side before flyback occurs. There is of course a front porch, but this is approximately the same on both systems. As a result of the foregoing the picture on 625 lines appears to be moved to In other words, if the the left, although the rasters are in the same positions. 625-line picture is adjusted to be central, when a change is made to 405 lines there will be a black margin on the left-hand side of the screen. This may be obviated by delaying the synchronizing pulse on 405 lines. It should be noted that it is the pulse that is being delayed and not the picture signal. Thus, when the pulse is delayed its position relative to the blanking period is now as at (b) in figure 5.45 rather than as at (a). In a sense we are increasing the front porch and reducing the length of the back porch. If figure 5.44 is examined it will be seen that if the period B-C is reduced by the correct amount the picture will be central. The delay is introduced to the pulses only and may be obtained by a simple delay network as m figure 5.45(c). Although the picture has now been centred, its width is not the same, although the size of the raster is constant. The original black margin on the left of the picture has been divided equally between the right- and left-hand sides. Fortunately, as already explained, due to the shape of the rectangular The

113 108 TELEVISION SERVICING BLANKING PERIOD SYNCHRONISING PULSE IN NORMAL POSITION SYNCHRONISING AFTER DELAY PULSE O _I 1 'WOW 1 i I_ FIG CORRECTION OF PICTURE POSITION (a) Normal position of line synchronizing pulse; (b) Delayed line pulse; (c) Delay circuit tube the picture must be overscanned in the horizontal direction and so the black margins do not show. This really means that almost the whole of the 405-line picture can be seen, but on 625 lines a small portion of the picture is cut off on both sides. Instead of making the flyback time as short as is required on the 625-line system it is sometimes made longer (as this is easier and reduces the voltages induced in the line output circuit). On 625 lines, the flyback will not be completed before the picture signal starts and the picture will be folded over on the left-hand side. If the time of fold-over is not too long it will occur in the overscanned portion and will not be visible. To be certain that it is not visible the spot may be suppressed during the line flyback in a manner similar to that used on field flyback. This is usually done by feeding negative pulses from the line output transformer to the grid of the cathode ray tube through a capacitor. Alternative negative pulses may be fed to the first anode of the tube Ȧ typical stabilized constant flyback time type of line output stage is shown in figure Valve V x is the line output valve which is fed from the line timebase, the screen grid being supplied in the normal manner. The line output transformer is operated in a desaturated condition by feeding the h.t. to point X, the line output valve current flowing as shown. The current in section a-b produces ampere-turns to oppose the ampere-turns produced by the current in section c-d. Capacitor C x is the efficiency diode capacitor which is charged by the efficiency diode valve V 2. The flyback time is controlled by the capacitor C 2. The e.h.t. rectifier V 3 is fed from the overwinding and has the inductor L t in series with the heater in order to maintain a constant heater voltage, as already explained. Point A" which is at an alternating voltage is fed from the h.t. supply through inductor L 2 and the top end of the line output transformer is maintained at a constant potential by capacitor C 3. The deflecting coils are fed from section a-b of the transformer while capacitor C 4 is the S correcting capacitor on 625 lines. Capacitor C 5 is added in parallel on 405 lines. To limit the dissipation of valve V x on 405 lines a resistor R x is inserted in the h.t. lead. The boosted h.t. voltage from d is used to feed the first anode of the c.r.t. through R 2 on 625 lines. On 405 lines R 2 is shorted out as the boosted h.t. voltage is less. A stabilizing voltage is obtained from across the VDR due to pulses fed through C 6. This voltage is offset by a positive voltage from the potential divider R 3 feeding through resistor R t. The potential divider settles the operating point of the line output valve. (ii) Constant Flyback Ratio In this case the ratio of scan to flyback time is

114 U LINE OUTPUT STAGE 109 Oil y o-wi/v-l FIG TYPICAL LINE OUTPUT STAGE. STABILIZED CONSTANT FLYBACK TIME CIRCUIT made the same for both systems. This has the advantage that the waveform of voltage across the deflecting coils is the same, i.e. the ratio of scan voltage to flyback voltage is the same. The actual voltage across the deflecting coils will be less on 405 lines due to the slower rate of rise of current. As far as the output valve is concerned this voltage can be made the same by having the scan coils connected across a different number of turns of the line output transformer on the two systems. In this way the voltages on the line output valve, the boost voltage and e.h.t. rectifier heater voltage remain constant. Because the time of flyback is now different the third harmonic tuning is upset. To correct this it is necessary to change the leakage reactance of the e.h.t. winding on the two systems. This may be done by placing a winding on the same limb as the e.h.t. winding and connecting it across a section of the primary when operating on 625 lines, so reducing the leakage inductance by increasing the coupling between the two windings. Since the scanning frequency is different the S correcting capacitor must be changed between the two systems; usually an additional capacitor is connected in parallel on 405 lines. In some cases the line output stage operates with neither a constant flyback time nor with a constant flyback ratio but somewhere in between the two. A typical non-stabilized constant flyback ratio type of line output stage is shown in figure V ± is the line output valve which is fed from the line timebase. The e.h.t. rectifier V 3 is fed from the overwinding in the usual manner. The core of the line output transformer is desaturated by feeding the h.t. through the separate winding W x. To prevent the voltage in this winding being applied to the efficiency diode, inductor i t and capacitor C t are added. These act as a smoothing circuit so that the anode of the efficiency diode is supplied with a constant direct voltage. Capacitor C 2 is the efficiency diode

115 110 TELEVISION SERVICING FIG TYPICAL LINE OUTPUT STAGE. CONSTANT FLYBACK RATIO CIRCUIT capacitor which is charged from the efficiency diode V 2. The deflecting coils are connected to different sections of the line output transformer on the two systems by the switch S x. Switch S 2 selects the different S correcting capacitors for the two systems. To tune the leakage reactance (third harmonic tuning) for the two conditions winding fv 2 and W 3 are added and connected together by switch S 3 on 625 lines. The width is varied by varying the h.t. supply by means of the resistor R u Capacitor C 3 controls the flyback time. The general demands on the line output valve are greater when operated on 625 lines because the scanning energy required is about li times more than on 405 lines. To deal with this increased power improvements have been made in line output valves. The screen current is a loss since it does not contribute to the output of the valve and should be reduced to a minimum. Various methods have been used to reduce this current which, apart from direct collection of electrons in the electron beam, is caused by electrons reflected from the anode and secondary electrons emitted from the anode. The collection of electrons from the electron stream can be reduced by shadowing the screen grid by grid wires, i.e. arranging that the screen grid wires are behind the control grid wires. This is known as aligned grid construction. Secondary emission from the anode may be reduced by coating it with a material of low secondary emission ratio. In one type of valve the screen current has been reduced still further by the "cavitrap" construction (Mullard), the general arrangement being as shown in figure Attached SCREEN CRID Ha» S~ ^ III -' CAVITRAP ANODE FIG y CATHODE 121 N I _r\ X - ^ODE CONTROL GRID "CAVITRAP" LINE OUTPUT VALVE

116 LINE OUTPUT STAGE 111 to the anode are dividing walls which tend to collect most of the secondary or reflected electrons from the anode. Compared with normal construction, this type of construction roughly doubles the ratio of anode-to-screen currents. TRANSISTOR LINE OUTPUT STAGES Transistor line output stages are not common and at present are limited to portable transistor receivers. Efficiency diode circuits are used; but a shunt efficiency diode circuit is used rather than the series efficiency diode circuit already described in connection with valves. If we consider the valve circuit as in figure 5.49 (which is the same as the basic circuit given in figure 5.17) is applied it is seen that the voltage V s across the efficiency diode capacitor C t «i? t c ' h P, FIG EFFICIENCY DIODE AND BOOST CIRCUIT (SERIES) in series with the h.t. supply to the line output valve V x. Thus, the effective supply voltage for V x is the h.t. supply plus the voltage across C,. Hence the voltage applied to V t (the boosted h.t. voltage) may be volts compared with an h.t. of volts. This is advantageous with valves, since the higher the applied voltage the less the current for a given output. It is easier to design a valve for a high voltage and low current rather than vice versa. Further, a valve tends to be more efficient when it is supplied with a high voltage since it is a high impedance device. Such a circuit is not suitable for present-day transistors which have a relatively low maximum collector-emitter voltage rating. A transistor is a low voltage device but can easily be made to pass a relatively large current. For this reason the shunt efficiency circuit is used. The basic circuit is shown in figure 5.50 where L represents the inductance of the deflecting coils and C represents the self-capacitance of the coils. The FIG BASIC SHUNT EFFICIENCY DIODE CIRCUIT

117 112 TELEVISION SERVICING efficiency diode is Re t this now being a semiconductor diode. An additional capacitor, C u is included to control the flyback time and hence control the flyback voltage. Transistor Tr x corresponds to the line output valve but is operated more as a switch, being either oft" or fully conducting by a suitable voltage and current fed to its base. During the second portion of the scan (that supplied by TrJ transistor 7>, is made fully conducting (i.e. bottomed) and hence the voltage across it will be low, say 1 volt. The current flows in the direction shown and increases in value as the scan proceeds. Hence the voltage across L and C is in the direction shown, i.e. it is a voltage tending to oppose the rise of current. At the start of flyback transistor Tr t is cut off and this corresponds to point B of figure 5.51 (which is the same as figure 5.12), i.e. just before flyback the current is a maximum, and there is constant voltage V s across the coil L. As soon as transistor Tr t is cut off the current FIG OPERATION OF EFFICIENCY SCAN CIRCUIT falls corresponding to the first part of a damped oscillation. Since the current is now decreasing the voltage across L reverses and rises as shown in figure This action continues until the current reaches zero at C. Since, at this point, the current is changing at a maximum rate the voltage is a maximum, the direction now being such as to add to the h.t. voltage as seen in figure and also into The current flowing in L now flows partly in C so charging it, Q increasing the voltage across this capacitor. The voltage applied to C x is now V s p+ V L. This same voltage is applied across the transistor Tr x and the FIG OPERATION OF EFFICIENCY DIODE CIRCUIT DURING FIRST HALF OF SCAN PERIOD efficiency diode Re u which must both withstand this voltage without damage. At point C all the energy originally stored in the magnetic field of L has been

118 LINE OUTPUT STAGE 113 transferred to C x (and C) and stored as an electric field. The current in L now reverses and the voltage decreases to point D (figure 5.51) when the voltage across L is zero and the current is a maximum in the reverse direction. The voltage across C t is now equal to V sup. If nothing else happened the current would decrease as part of a sine waveform as shown dotted. However, as the current starts to decrease again, the voltage reverses until point E is reached, at which point the voltage across L (v L ) is equal to the supply voltage V sup. The conditions are as in figure The voltage on the collector of Tr t and across C t is now zero and hence Rei conducts and current flows as shown. Neglecting the drop across Re lt the voltage v t across L is constant and hence the rate of fall of current must be constant since L di dt FIG OPERATION OF EFFICIENCY DIODE CIRCUIT DURING SECOND HALF OF SCAN PERIOD Thus the current falls in a linear manner to zero at point F (figure 5.51) so producing the first portion of the scan. In practice the decrease in current is not quite linear, due to the resistance of the circuit. Since this voltage V L is approximately the same as V s during the portion A-B (they would be equal if the drop across 7V, was neglected when it is fully conducting) the rate of current change from E to F is the same as that from A to B. At F transistor Tr x is made fully conducting and the second part of the scan corresponding to A-B is produced. During the decrease of current from E to F the energy of the coils is returned to the supply. Ideally, if there were no losses all the energy fed into the coils from the supply, during the portion A-B, would be returned to the supply during portion E-F. It should be noted that no efficiency diode capacitor is required since the energy has not to be stored for the next half of the scan, as in the valve circuit. If C, is increased then the flyback time is increased, because the frequency of the damped oscillation is reduced. This also results in a decrease in the peak voltage across C t and hence across Tr t which occurs at point C of figure This is due to the fact that the energy stored in a capacitor is equal to icv 2 ; thus, for a given energy, V is reduced if C is increased. Since it is necessary to keep the maximum voltage across the transistor at a minimum, the flyback time is increased as much as possible

119 114 TELEVISION SERVICING (commonly up to 23 per cent, of the total time). This is greater than the blanking time, and advantage is taken of the fact that in order to fit the picture into the aspect ratio of the c.r.t., some horizontal overscan is present. Hence, if the flyback is too long the effect will not be seen, provided, of course, it is not excessive. The voltage across the transistor during flyback is commonly about 70 volts with a 12-V h.t. supply. During the portion A-B the transistor is fully conducting and, at the start of flyback, it is required to cut off the transistor. Unlike a valve there is a delay due to an excess carrier density built up in the base of the transistor during the conducting period (hole storage effect). Thus, although the base current may be reversed, the collector current will continue to flow until the excess carriers have been swept away. This means that there is a delay in the start of the flyback and so direct synchronizing of the line timebase is not possible; a flywheel synchronizing circuit must be used (see Volume 3). In the basic circuits an n-p-n transistor has been shown, together with a positive supply. This was used because it was easier to compare with the corresponding valve circuit. Transistors of the p-n-p type may be used (and are more commonly used) in which case the polarity of the supply is reversed and the efficiency diode is reversed, as in figure 5.54(a). The operation is T % r O C i 1 L Tri R., (b) FIG (a) COMMON EMITTER LINE OUTPUT CIRCUIT (b) COMMON COLLECTOR LINE OUTPUT CIRCUIT

120 LINE OUTPUT STAGE 115 exactly similar to that described except that all currents and voltages are reversed. The driver transformer 7\ has been included in this figure. An alternative is to use a positive supply and reverse the p-n-p transistor as shown at (b). This does not alter the operation of the circuit and such changes are easily made with a transistor but not, of course, with a valve. *jii* w,.gi : oil li- LZi (7r eht --, w» "»«m FIG TRANSISTOR LINE OUTPUT STAGE A more complete circuit is given in figure Transistor Tru C t and Rei are as the previous circuits but, instead of feeding the deflecting coils D directly they are fed through the autotransformer T 2. Transistor TVi is driven through transformer T t from the driver transistor Tr 2. The width ot scan is controlled by the variable inductor L t and the linearity by means of L 2 which has a core saturated by a permanent magnet, similar to that used in valve circuits. An e.h.t. supply is required for the final anode supply of the c r t and this is obtained using the flyback voltage, essentially as m a valve circuit. In figure 5.55 this is obtained by the use of an e.h.t. winding W^ on the line output transformer T 2. This is rectified by a valve rectifier V l whose heater is supplied from a heater winding W t. An overwinding (i.e. extension to the main winding) cannot be used in this case since the flyback pulse on the collector of Tr x is negative; an overwinding would therefore give a negative pulse which is useless when a positive e.h.t. supply is required. If a positive supply voltage is used (with an n-p-n transistor or reversed p-n-p transistor) then an overwinding can be used. As in valve circuits third harmonic tuning is used to reduce the flyback voltage on the transistor and increase the e.h.t. V Sets using transistor output stages are at present designed as portable receivers operating off a 12V supply or off a mains supply. In the latter case the mains is converted to a 12V d.c. supply. Thus there are no voltages available for focus electrode and first anode of the c.r.t. These are positive voltages and are obtained from an additional winding W5 on the line output transformer, using rectifier Re 2 and smoothing capacitor C 3. A first anode voltage of about 300 volts is required and a variable focus voltage is obtained from potential divider P 2.. The 12V supply is not sufficient for the transistor video amplifier which is required to produce a relatively large drive to the cathode of the c.rx Thus, a supply of higher voltage (usually about 90 volts) is required for the video output stage. This is obtained from winding W 3 on the line output transformer, together with rectifier Re 3 and smoothing capacitor C 2. A negative voltage supply has been shown so that a p-n-p transistor can be used. Since the video amplifier has a negative supply the cathode of the cathode ray tube

121 116 TELEVISION SERVICING will be negative with respect to chassis and hence the grid must be more negative. Thus the brightness control P t is fed from this same negative supply. The driver stage Tr 2 may supply a sawtooth current to Tr-, but more commonly a pulse to drive the line output transistor Tr x into full conduction A pulse could be provided to Tr x (to switch it on) by the action of switching transistor 7> 2 on, i.e. both Tr t and Tr 2 conduct together. Alternatively, 7>, can be switched on by the action of Tr 2 being switched off, the driving pulse coming from energy stored in 7\. The latter arrangement is normally used and is called the non-simultaneous mode, since both transistors do not conduct at the same time. The driver stage is, of course, driven from the line timebase. At present a transistor output stage costs more than that of a valve circuit and is restricted to those receivers where valves cannot be used. No doubt the position will change in a few years' time.

122 CHAPTER 6 SAWTOOTH CURRENT GENERATORS Instead of using a voltage timebase to produce a sawtooth voltage waveform (or a pulse in some cases) to drive an output stage which feeds the deflecting coils, a circuit may be used which generates directly the required sawtooth current waveform. The basic idea is shown in figure 6.1. Suppose FIG BASIC SAWTOOTH CURRENT GENERATOR CIRCUIT that the valve is cut off during flyback and the grid voltage is constant during scan. At the start of the scan the current in L (the deflecting coils fed through a transformer) will be zero but it will rise exponentially, in a way exactly similar to a capacitor charging through a resistor, at a rate determined by the time constant L-R of the circuit where L is the inductance of the circuit and R is the total resistance of the circuit. This is because an inductance sets up an e.m.f. tending to prevent the rise of current in it. At the end of scan the valve is cut off and flyback occurs, as described in Chapter 5. Thus, if the valve grid is fed with a large negative pulse corresponding to the flyback, the stage operates to produce a sawtooth current in the deflecting coils. This method of operation may be used with the circuits described in Chapter 5, by obtaining a pulse output from the timebase circuit instead of a sawtooth waveform. On the other hand, the grid drive may be obtained from the anode circuit of the output valve by a feedback circuit. The stage then acts as a sawtooth current generator or oscillator. A basic circuit is shown in figure 6.2. To obtain the grid drive a feedback winding L 2 is coupled to the anode coil L^. On switching on, the current in Transformer feeding deflecting coils FIG BASIC CURRENT TIMEBASE CIRCUIT L t rises. L 2 is wound in such a direction that the voltage induced in it drives the grid positive and so maintains or helps the anode current. Eventually, the rise of anode current will cease, due either to the application of a negative synchronizing pulse to the grid or to bottoming of the valve (i.e. reaching a point when no further increase in anode current can take place due to the low anode voltage). When the rise of anode current ceases the voltage induced in L 2 ceases (since it is proportional to the rate of rise of current in LJ and this causes the anode current to decrease because there is now no voltage fed 117

123 118 TELEVISION SERVICING to the grid. This causes the voltage in L 2 to reverse and drive the grid negative, the cumulative action cutting the valve off. At the end of flyback the voltage across L 2 is no longer present and the action starts all over again. There are a large number of variations of the basic idea, some using one valve and some two valves. The idea may be applied to line or field output stages. At one time this type of circuit was very common but it appears to have lost popularity, particularly as regards the field output stage. In view of the many arrangements that are possible only a few can be dealt with. In all the arrangements there must be some feedback circuit between the anode circuit (or screen grid circuit in some cases) and the grid circuit so that the stage will oscillate. SINGLE-VALVE CIRCUITS Figure 6.3 shows a typical circuit which is conventional apart from the feedback path produced by d.-r, and R 2. During scan the voltage on the FIG TYPICAL SINGLE-VALVE CURRENT TIMEBASE CIRCUIT secondary is in the direction shown, so tending to make the grid of V x positive causing it to conduct; the anode current will therefore rise. At the end of the scan a negative synchronizing pulse is applied to the grid, reducing the anode current. This causes a reversal of the voltage across the secondary (since it is proportional to the rate of change of current), the grid is driven highly negative, the valve being cut off so producing the flyback. When the voltage across the secondary is reversed again during the flyback, the valve K, will conduct and the operation will continue. The exact instant when V t conducts will depend on the capacitor Q and R U R 2 and the circuit is designed so that the efficiency diode V 2 provides the first portion of the scan. Another basic circuit is shown in figure 6.4 where the grid and screen grid are used as an oscillator. The operation is similar to that of figure 6.2 except 5 LINE OUTPU1 CIRCUIT FIG SINGLE-VALVE CURRENT TIMEBASE CIRCUIT that the screen grid takes the place of the anode. The frequency of operation of the circuit is altered by variation of the inductance of L x. This is coupled

124 SAWTOOTH CURRENT GENERATORS 119 to the transformer by winding 3 and, therefore, alters the effective inductance of winding 1, hence the rate of rise of current. During scan the valve is conducting and the anode (and screen) currents increase in an approximately linear manner. The rise of anode current operates a conventional scanning circuit connected to the anode. During flyback the valve is cut off and no anode or screen currents flow. It is important to note that the circuit does not operate as a blocking oscillator, in which the valve is cut off during the scan and conducts only during the flyback. TWO-VALVE CIRCUITS A typical two-valve basic circuit is shown in figure 6.5 and may be considered as a type of multivibrator. During the scan the valve V 2 is S CIRCUIT L* t.h.t FIG TWO-VALVE CURRENT TIMEBASE CIRCUIT conducting and the anode current will rise at a rate determined by the inductance in the anode circuit, while valve Vx is cut off. A negative synchronizing pulse fed to the anode of V x is communicated to the grid of V 2 through Q and tends to cut off valve V2 This causes a reversal of polarity. across the transformer and a positive voltage is fed through C 2 to the grid of V. t Hence V x conducts. The rise of anode current in Vx causes a drop in anode potential which is fed to the grid of V 2 and, due to cumulative action, V 2 is cut off, resulting in the flyback. At the end of the flyback the voltage fed back from the transformer is reversed so cutting off V v This, in turn, feeds a positive voltage to V 2 causing it to conduct. The frequency of operation of the circuit, in the absence of synchronizing pulses, is settled by R2 (and C 2) since Vx will start to conduct when the charge on C 2 (produced during the flyback pulse) has leaked away through R 2. Thus, R 2 is the hold or frequency control. Instead of feeding a constant voltage to the grid of V 2 during the scan, a sawtooth voltage may be obtained by placing a suitable capacitor across AB. This will charge through the anode resistor of K, during scan and produce a sawtooth voltage.

125 APPENDIX AVOIDING SWITCHING-OFF BURNS On older receivers it was usual for a spot to remain in the centre of the screen for some time after the set had been switched off. In some cases this was present for several minutes, and usually worried the non-technical viewer. A burn mark on the screen may have resulted. The spot was caused by the e.h.t. retained by the e.h.t. smoothing capacitor. The cathode of the c.r.t. takes some time to cool and there is sufficient emission to produce a small beam current for some time after switching off. The deflecting fields have, of course, collapsed and as there is no deflection a single spot results in the centre of the screen. This remains until the e.h.t. capacitor is discharged, which may take a considerable time because the beam current will be small. There are two basic methods of eliminating this trouble. (1) By maintaining a high negative voltage on the grid of the cathode ray tube until the cathode has cooled sufficiently for no emission to take place. Owing to the difficulty of maintaining this bias over a long period this method is not used. A disadvantage is that it leaves the e.h.t. capacitor charged, which may be a source of danger to service engineers. (2) By rapidly discharging the e.h.t. capacitor while the beam is still being scanned. The easiest way is to use the c.r.t. itself by arranging that the beam current increases immediately the set is switched off. There are a number of possible methods. (0 Return of the brightness potential divider to the mains side of the on-off switch. This is shown in figure 1. When the set is switched off the.to VIDEO AMPLIFIER MAINS TO RECTIFIER f ON-OFF SWITCH T BRILLIANCE CONTROL H looka g^ T _p»e -No j, I R :c OOljiF WJ& T I ' CRT. FIG. 1. RETURN OF POTENTIAL DIVIDER TO MAINS SIDE OF ON-OFF SWITCH whole of the potential divider takes up h.t.+ potential since the lower end of the potential divider is disconnected (i.e. it is no longer a potential divider). The result is that the grid goes positive and the c.r.t. rapidly discharges the e.h.t. capacitor. There are two disadvantages here. The first is that it is only effective if the set is switched off by the receiver switch. It does not operate if the receiver is switched on and off by (say) a wall socket. The second disadvantage is that 120

126 APPENDIX 121 there is a connection through the potential divider from the mains to the receiver even when the set is switched off. This may be a connection to the live side of the mains should the receiver be incorrectly installed or the plug reversed, which might prove dangerous. This could be overcome by the use of a triple pole switch; but it would be costly and does not overcome the first disadvantage. This circuit is not commonly used. (h) Use of a VDR for the potential divider. This is shown in figure 2. It will be seen that the grid is now fed off a circuit similar to that used to maintain a constant h.t. supply for the field timebase described earlier. Thus, if the h.t. voltage varies then the voltage across the VDR (and the voltage applied to the grid) will remain approximately constant. When the set is switched off the h.t. voltage drops and so therefore does the cathode voltage of the c.r.t. Due to the use of the VDR the voltage on the grid does not fall by the same amount and the grid tends to become positive with respect to the cathode and again the e.h.t. capacitor is rapidly discharged. This circuit is commonly used. TO VIDEO AMPLIFIER FIG. 2. USE OF VDR FOR POTENTIAL DIVIDER (iii) Use of a VDR as a rectifier. A basic circuit is shown in figure 3. produced across the VDR by the application of Here a negative voltage is line pulses in a manner which has been described in connection with the stabilization of the line output stage. Thus the potential divider is normally taken to a point of negative potential. When the set is switched off the pulses from the line timebase decrease rapidly and the negative voltage drops. This results in a higher voltage being applied to the grid of the c.r.t. and the e.h.t. capacitor is discharged. This circuit is rarely used. JO VIDEO AMPLIFIER ±-HI FIG. 3. USE OF VDR AS RECTIFIER

127 Afterglow Aligned grid construction of line 15 output valve 110 Aluminized screen 8, 15 Anode (first) supply of cathode ray tube 95 Aspect ratio 12 Atmite Automatic limiter (synchronizing 61 separator) 44 Autotransformer (line output) 90 Beam, deflection 9 Bifilar winding 92 Blocking oscillator timebase 51, 58 Boosted h.t. voltage 87 Burns, switching-off 122 Cathode Coupled Multivibrator Timebase 57 Cathode-heater voltage, efficiency diode 91 Cathode ray tube 1 rectangular 11, 12 sizes and shapes 10 "Cavitrap" line output valve 110 Constant flyback time line output stage 105 ratio line output stage 105, 108 Control of diode current (efficiency diode) 87 Correction of non-linearity 69, 87 Critical anode voltage of fhyratron 47 Critical damping 78 Crossover, of cathode ray tube 5 Current timebase 117 single valve 118 two valve 119 Damping of Line Deflecting Coils 78 Damped oscillation 78 D.C. component 18 restorer 20 Deflecting coils 13 field, high impedance 67 low impedance 68 Deflection, of beam 9 electrostatic 9 magnetic 9 Desaturated line output transformer 103 Differentiating circuit 34, 41 Differentiation of sawtooth waveform 70 Diode current, control of 87 Diode synchronizing separator 21 Distortion correcting circuits 69 Double limiter 38 Driven timebase 47 INDEX 122 Efficiency Diode 85 circuit 81,112 plus boost circuit 84 with high heater cathode insulation 92 of line circuit, importance of 81 scan circuits 81 E.H.T. supply 50Hz 93 line output or flyback 93,115 r.f. oscillator 93 voltage Electron beam in electrostatic 93 field 5, 9 gun magnetic field 2, 9 velocity Electrostatic deflection 9 focusing - 5 lens 2 anode 6 3 anode 6 4 anode 6 Energy, stored in L and C 77 Equivalent circuit of deflecting coils field 65 line 75 Equivalent circuit of transformer 68 Feedback, in timebase 50 Ferroxdure 4 Flyback blackout circuit 55, 73, 108 e.h.t. 93 advantages and disadvantages 94 Fluorescent screen 15 Focusing of cathode ray tube 2 unit 3 Field flyback blackout circuit 55, 73 output stage 65, 74 transformer 68 synchronizing separator 34 Grid of Cathode Ray Tube 2 Hard Valve Timebases 50 High heater-cathode voltage diode 92 High impedance field deflecting coils 67 Hold control of timebase 50 Integration of Sawtooth Waveform 70 circuit 36 in screen grid 38 Interlace 35 filter 44 l l

128 Ion 7 burn 7 trap 7 adjustment of 8 Ionization 7 Leakage Flux 68, 99 reactance 68, 99 Lens, electrostatic 5 magnetic 2 Limiter 36 Linearity 48, 54, 65 control 89 correcting coils (line) 90 Line output stage 75,111 operation of line 103 valve, operation of 80 transformer, desaturation 103 synchronizing separator 33 Low impedance field deflecting coils 68 Magnetic Deflection 9 focusing 2 lens 2 Metrosil 59, 91 Multivibrator timebase 59 Negative Feedback Correcting Qrcuit 71 ion 7 Noise gated synchronizing separator 27 Non-linearity 87 Output stage, field line Overshoot INDEX ,74 73, Partial Differentiating Circuit 41 with auto, limiter 44 Pentode synchronizing separator 23 Picture synchronizing separator 18, 32 Positive ion P-tube 7, 1/ Pulling on whites 25 Pulse shape, effect of 33 Pulse width discriminator 45 Rectifier Operation, E.H.T. Circuit Rectangular cathode ray tube Resonant circuit R.F. oscillator e.h.t. supply Saturated Linearity Control 89 Sawtooth current generator 117 single valve 118 two valve 119 Screen of cathode ray tube 15 S correction 98 Secondary emission 15 Self-running timebase 47 Shadowing of cathode ray tube 15 Smoothing capacitor, e.h.t. 94 Stabilization of timebase 60, 63 (line) 100 Stored energy in capacitor 77 inductor 77 Squegging oscillator 52 timebase 52 Stripping of cathode 48 Supply voltage stabilization 63 Switching-off burns 122 Synchronizing of timebase 48 pulses, effect of differentiating 34 separator 18, 32 auto bias type 22 diode 21 effect of noise 24 noise gated 27 pentode 23 Temperature Stabilization 62 Tetrode cathode ray tube 5 Thermistor 60 Third harmonic tuning 98 Thyratron timebase 47 valve 47 Thyrite 61 Timebase 46, 57 blocking oscillator 51, 58 cathode coupled multivibrator 57 driven 47 hard valve 50 multivibrator 56, 59 self-running 47 squegging oscillator 51 synchronizing 48 thyratron 47 Toroidal deflecting coil yoke 14 Transformer, equivalent circuit 68 Transistor, field output stage 74 line output stage 111 picture synchronizing separator 32 timebase 57 Triggering action 49 Triode cathode ray tube 4

129 124 TELEVISION SERVICING Voltage Dependent Resistor Waveforms across Field (VDR) 61 Deflecting Coils 65 line as deflecting rectifier coils 75 Voltages in Line Output Circuit 75 Wehnelt cylinder 2 Voltage timebase 46, 57 Width control 95

130 Vol. 1 Vol. 2 COLOUR TELEVISION WITH PARTICULAR REFERENCE TO THE PAL SYSTEM G. N. Patchett, B.Sc.(Eng.), Ph.D., C.Eng. 2nd edition 50/- (paperback) ELECTRICAL ENGINEERING FOR ORDINARY NATIONAL CERTIFICATE G. N. Patchett, B.Sc.(Eng.), Ph.D., C.Eng. Current Electricity 8/6 Vol. 3 Magnetism and Electrostatics 7/6 Vol. 5 RADIO SERVICING Alternating Current Theory 12/6 Basic Electronics 9j- Vol. I Basic Electrotechnology 8/6 Vol. 3 Final Radio Theory 10/- Vol. 2 Intermediate Radio Theory 12/6 Vol. 4 TELEVISION SERVICING G. N. Patchett, B.Sc.(Eng.), Ph.D., C.Eng. Fault Finding 9/6 Vol. I 14/- Vol. 2 15/- Vol. 3 8/6 Vol. 4 15/- THE OSCILLOSCOPE BOOK 7/- E. N. Bradley ELECTRONIC NOVELTY DESIGNS 8/6 I. J. Kampel HOW TO GET THE BEST OUT OF YOUR TAPE RECORDER Percival J. Guy PRACTICAL TAPE RECORDING 7/6 Percival J. Guy SERVICING TRANSISTOR RECEIVERS 15/- F. R. Pettit USING AN OSCILLOSCOPE 14/- D. W. Easterling RADIO SERVICING PROBLEMS 9/- W. A. L. Smith, M.I.Prod.E. WELDING SCIENCE Vol. 1 5/- J. Gardner, F.R.S.A., M.Inst.Met., A.Inst. W. Vol. 2 7/6 ro DUAL-STANDARD AND 625-LINE TELEVISION RECEIVERS 8/6 Gordon J. King, Assoc.I.E.R.E., M.T.S. FAULT LOCATION EXERCISES IN RADIO AND TELEVISION SERVICING Vol. 1 Vol. 2 8/6 Vol. 3 10/- K. J. Bohlman, A.M.Inst.E. THE DESIGNER'S GUIDE TO BRITISH TRANSISTORS I. J. Kampel 23/- (paperback) PAL-D COLOUR RECEIVER: QUESTIONS AND ANSWERS 13/6 K. J. Bohlman, A.M.Inst.E, MECHANICS' COURSE IN RADIO, TELEVISION AND ELECTRONICS (433) G. N, Patchett, B.Sc.(Eng.), Ph.D., C.Eng. Electronic Systems {Pari I: First Year) 151- F.lec ironic Systems (Part tl: Second Year) 19/-