12G-SDI Physical Layer Analysis using the Ultra 4K Tool Box

Transcription

1 12G-SDI Physical Layer Analysis using the Ultra 4K Tool Box Authors: Alan Wheable FISTC, MITOL Senior Technical Author, Alex Huntley MEng Senior Consultant and Andy McMinn BEng Senior Consultant at Omnitek This White Paper covers the use of the Advanced Physical Analysis features of the Omnitek Ultra 4K Tool Box to perform physical layer analysis of 12G-SDI signals which differs from the techniques already adopted by the industry to measure SD-SDI, HD-SDI and 3G-SDI signals. Here we will explain the differences between traditional physical layer analysis and those that now need to be adopted for 6G-SDI and 12G-SDI to ensure repeatable and comparable results to those typically only available on very high-end oscilloscopes such as the Teledyne LeCroy SDA 820Zi-A. When performing physical layer analysis there are two distinct areas of interest. These are the signal quality, viewed and measured using the Eye Diagram display, and signal jitter viewed and measured using jitter meters and jitter spectrum analysis tools. Traditionally users expect to see a recognisable Eye Diagram that can be used to provide a visual level of signal quality. Ideally to ensure that the displayed Eye Diagram is without jitter the most appropriate waveform sampling and capture method must be employed. One of the significant challenges of 12G-SDI signal measurement is how to accurately and repeatedly measure waveform amplitude because without this measurement values of rise time and fall time cannot be made. Due to the high 3rd and 5th harmonic frequencies of the 6GHz fundamental frequency (relative to the bandwidth of the BNC and cable) the signal actually sent by currently available transmitters is effectively a distorted sine wave. Making accurate amplitude measurements of signals like this depends on the measurement technique used. Jitter measurement has always been a significant part of physical layer analysis as it ultimately determines whether or not that there will be errors in the data. The first challenge is to ensure that the jitter produced by a piece of equipment is within SMPTE limits and the second is how to identify what is causing out of specification jitter, which can often be determined by the frequency band in which it occurs. The generation of known levels of jitter can be used to measure the susceptibility of equipment under test to jitter. The Ultra 4K Tool Box allows jitter of a user-defined frequency and amplitude to be introduced into the generated SDI output signal. This White Paper discusses these aspects of physical layer analysis using the Omnitek Ultra 4K Tool Box in comparison to the Teledyne LeCroy SDA 820Zi-A high-end oscilloscope and shows that the results are comparable and well within acceptable tolerances. Copyright Omnitek 2015

2 Contents 1 Making 12G-SDI Signal Measurements The Challenges of 6G-SDI and 12G-SDI The Importance of Eye Diagram Shape Factors that affect signal shape Common Measurement Practice Eye Diagram Measurements Eye Diagram Capture Amplitude Measurement Rise and Fall time Measurement Eye Diagram Comparison of Omnitek Ultra and Teledyne LeCroy Jitter Measurement Understanding Jitter Measurement Definition of Signal Jitter Jitter Measurement Method Noise Floor and Low Jitter Mode Rx Jitter Tolerance Jitter Transfer Intrinsic Jitter and Output Jitter Jitter Meters Instrument Jitter Spectrum Instrument Jitter Histogram Instrument Jitter Waveform Instrument Jitter Comparison of Omnitek Ultra and Teledyne LeCroy Output Signal Conditioning Output Jitter Generation Output Signal Amplitude and Slew Rate Conclusion Page 2 of 24

3 1 Making 12G-SDI Signal Measurements 1.1 The Challenges of 6G-SDI and 12G-SDI With the advent of SD-SDI, HD-SDI and 3G-SDI the measurement of Eye shape and jitter became an everyday task to ensure that signals were correct and transmission was maintained. Measurements were made at the transmitting equipment end to ensure that the best Eye Diagram was available at the receiving equipment. With SD-SDI, HD-SDI and 3G-SDI signals the Eye Diagram can be seen to be basically a square wave with harmonic distortions allowing the amplitude, rise time, fall time, overshoot and undershoot to be measured relatively easily and accurately (see section 2.1). For 12G-SDI with a fundamental frequency of 6GHz, the high harmonic frequencies (ie a 3 rd harmonic of 18GHz and a 5 th harmonic of 30GHz) mean that the signal shape at the transmitting device is effectively a sine wave. HD-SDI (1.5G) Eye Diagram Typical 12G-SDI Eye Diagram 100% 100% 80% 50% Overshoot Rise Time Fall Time 80% 50% Rise Time Fall Time 20% Undershoot 20% 0% 0% (rise time 270ps and fall time 270ps) (rise time 35ps and fall time 35ps) Figure 1.1 Comparison of HD-SDI and 12G-SDI Eye Diagrams One challenge when trying to make objective measurements of the signal quality is the amount of jitter on the signal. To allow measurements to be made purely on the waveform shape the jitter is largely removed from the displayed waveform. In the case of the Omnitek Ultra 4K Tool Box this is done using a highly agile clock recovery system that allows compensation for any signal jitter up to 20MHz and removes this from Eye Diagram display. 1.2 The Importance of Eye Diagram Shape The SDI signal Eye Diagram shape is important because it determines the total cable length that the signal can be sent (without cable compensation or repeater) to a receiving device for successful decoding. With 6G-SDI and 12G-SDI this cable length is proportionally shorter than with HD-SDI or 3G-SDI (typically 220m for HD-SDI, 180m for 3G-SDI, 90m for 6G-SDI and 60m for 12G- SDI). The Eye Diagram shape is an indication of the quality and return loss of the transmitting equipment s output circuitry and manufacturers want to ensure that the quality is as good as possible and well within SMPTE specified limits. To ensure compliance with SMPTE specified limits, all waveform measurements need to be made using a 1m length cable connected directly to the output of the transmitting device while the device is transmitting 100% colour bars. In practice analysing the physical layer on longer cables can only give an approximate representation of the signal quality. Page 3 of 24

4 1.3 Factors that affect signal shape As mentioned earlier, the design of the signal generator output circuitry has a significant effect on the shape of the 12G-SDI signal. Factors include the driver frequency response / slew rate, insertion loss (caused by the type of PCB substrate type), return loss (caused by the chosen BNC connector and tracking) and cross talk between other inputs and outputs (caused by proximity to other PCB tracks and connectors). From both simulation and practical experience, Omnitek have found that Edge-Mount BNCs produce the best results. Figure 1.2 Omnitek s experimental 12G-SDI BNC PCB Cable quality, type and length will all have an effect on the signal shape because of the attenuation of high frequencies, but over a 1m length (SMPTE RP 184) these should not be a significant impact. Throughout this document the examples shown use Belden 1694 A cable. In practice infrastructure components including connector types, patch panels, cable joiners, T- Pieces, termination pads and loop through connections will also have a significant effect on the 12G-SDI signal shape. 1.4 Common Measurement Practice SMPTE recommendations consistently define the measurement practices for SDI signal waveform shape and jitter but quite often these are the ideal theoretical methods. In practice broadcast equipment manufacturers, systems integrators and users employ a whole range of methods to ensure that the equipment stands up to the day to day challenges in broadcast. The SMPTE RP 184 and RP 192 expect SDI waveform and jitter to be measured using the recommend method (ie using a 1m cable and 100% colour bars). In practice, however, manufacturers, installers, system integrators and operators will be checking signal quality using their own preferred test pattern or even a live source over the longest cable lengths that they can get away with. Broadcast equipment manufacturers will typically ensure that their products easily meet the SMPTE standards using 1m good quality cables and 100% colour bars. They will also measure the performance over 10m, 30m, 50m, etc using dynamically changing test patterns, live images and also pathological test patterns to determine the limits and boundaries of their products. For products that transmit SDI signals, the shape of the waveform and the level of intrinsic (output) jitter effectively determines the length of cable over which their signal can be successfully sent. For products that receive SDI signals, the tolerance of low signal, poor signal shape and jitter effectively determines the length of cable over which their product can correctly receive and successfully decode the SDI signal. System integrators, equipment installers and maintainers will often use whatever signal is available at the time to prove that equipment works correctly and will often be using Eye Diagram and jitter measurement to establish operational limits (primarily cable length and type that can be used). There is an expectation that there is a visible Eye Diagram and acceptable jitter tolerance under all circumstances. With the adoption of 6G-SDI and 12G-SDI interfaces, the testing will become more stringent and new practices will evolve. Page 4 of 24

5 2 Eye Diagram Measurements 2.1 Eye Diagram Capture When looking at the Eye Diagram of an ideal HD-SDI signal with low levels of jitter it is very easy to manually assess the waveform shape, signal amplitude, rise time, fall time, overshoot and undershoot. So this jitter free display has become the accepted norm. 100% 80% 50% 20% 0% Overshoot Undershoot Rise Time Fall Time Figure 2.1 Typical HD-SDI Eye Diagram Users expect to see a recognisable Eye Diagram that can be used to provide a visual level of signal quality. To allow measurements to be made purely on the Eye Diagram the jitter is removed because even jitter within SMPTE specified limits for 12G will effectively close the Eye making visual measurement difficult. There are basically two different approaches to capturing the Eye Diagram from the SDI signal. These are real-time capture (employed by Real Time Oscilloscopes) and signal subsampling (employed by Sampling Oscilloscopes). Real Time Oscilloscopes build the Eye Diagram from individual line traces from the SDI signal whereas Sampling Oscilloscopes build the Eye Diagram from individual pixel samples. These two different approaches are used to produce a pseudo live display. Real Time Oscilloscope display (Teledyne LeCroy SDA 820Zi-A) Sampling Oscilloscope display (Omnitek Ultra 4K Tool Box) (measured rise time 34.8ps, fall time 34.6ps) Figure 2.2 Eye Diagram Comparision (measured rise time 33.9ps, fall time 34.5ps) The proprietary Jitter Tracking technique used by the Ultra stabilizes the Eye Diagram to give comparable results to that of high-end oscilloscopes as discussed in section 2 and section 3. Note that a common misbelief with Eye Diagram capture is that an instantaneous (so called live display) is better than that of a captured or integrated Eye Diagram. In practice the time period is irrelevant as the Eye Diagram displayed is always actually the same and this is the important factor when making the critical measurements of signal amplitude, rise time and fall time. Live Displays tend to show artefacts of the sub sampling process such as a sparkles which have no benefit when measuring Eye Pattern Parameters and can give erroneous results if there is jitter on the input signal. Page 5 of 24

6 Depending on the setup, Real Time oscilloscopes can produce very different looking eye patterns. First they take a large number of samples of the data stream. A software PLL (phased locked loop) then fits a clock to the sampled data stream in order to fit the samples into the eye pattern. The parameters of that software-based PLL can make a significant difference to the appearance of the eye pattern. The optimum PPL bandwidth and frequency which gives the best Eye Diagram is often referred to as the Golden PLL The effect on the Eye Diagram when the Real Time oscilloscope is using a low bandwidth PLL, which does not track jitter, can be seen in the following example where the 12G-SDI signal contains a 0.1 UI jitter at 500kHz. Real Time Oscilloscope display with a low bandwidth PLL Omnitek Ultra 4K Tool Box with its high bandwidth PLL Figure 2.3 Effect of low bandwidth PLL The effect on the Eye Diagram when using a high bandwidth PPL, giving comparable results to the Ultra can be seen below with the same 12G-SDI signal containing a 0.1 UI jitter at 500kHz. Real Time Oscilloscope display with a high bandwidth PLL Omnitek Ultra 4K Tool Box with its high bandwidth PLL Figure 2.4 Effect of high bandwidth PLL Page 6 of 24

7 2.2 Amplitude Measurement For 12G-SDI the first challenge is how to correctly measure the amplitude of a sine wave. Here there are a number of different ways of methods, each of which will can produce slightly different measurements for rise time and fall time. Statistical analysis of the waveform amplitude provides the most reliable and consistent measurement: Mode (100%) 20% Eye Aperture 100% Mean (20%) Figure 2.5 Mean and Mode measurement Mean Measurement determines the amplitude using the weighted mean (centre of gravity) of the amplitude values from a 20% sample centred about the eye. This is the default measurement method used by most high-end oscilloscope manufacturers. Mode Measurement determines the amplitude of the signal using the most common value (or peak value) from all the amplitude values. The mode and mean measurements can produce different values. This is most evident when not all of the transitions in the bit stream reach full amplitude. This is very typical for 12G-SDI signals and therefore the mean will typically read smaller than the mode in this case. A mean measurement will reflect different amplitudes of high and low frequency content and is arguably the more representative value. Page 7 of 24

8 2.3 Rise and Fall time Measurement The rise and fall times of the 12G-SDI signal need to be within SMPTE ST and RP 184 specified limits ( The rise and fall times, determined between the 20% and 80% amplitude points shall be no greater than 45ps and shall not differ by more than 18ps ). 100% 80% Rise Time Fall Time 50% 50% Sample Point 20% Figure 2.6 Eye Diagram Measurements The rise time is calculated from the rising slope (leading edge) of the signal between the 20% and the 80% amplitude levels. Likewise the fall time is calculated from the falling slope (trailing edge) of the signal between the 80% and the 20% amplitude levels. Note that if rise and fall times are poor the high frequency components of the Eye Diagram will not reach full amplitude causing the Eye to close vertically and there is an increased risk that data is sampled incorrectly. If there is any difference between rise and fall time speeds this can introduce duty-cycle dependent jitter. Both of these affect the total distance that the SDI signal can be successfully transmitted. The amplitude, rise time and fall time values are measured automatically by the Ultra 4K Tool Box and displayed in the Eye Waveform window. 0% Figure 2.7 Automatic Eye Diagram measurement Page 8 of 24

9 2.4 Eye Diagram Comparison of Omnitek Ultra and Teledyne LeCroy In this direct comparison between the Omnitek Ultra 4K Tool Box and the Teledyne LeCroy SDA 820Zi-A, a 12G-SDI signal was generated by the Omnitek Ultra 4K Tool Box and connected in turn to the SDI inputs of the 2 different devices first using a 1 meter Belden 1694 A cable then secondly using a 10 meter Belden 1694 A cable using the Mean amplitude measurement method. Figure 2.8 Omnitek Ultra 12G-SDI input with Belden 1694 A 1m Cable The Ultra measured the signal amplitude as mv, the rise time is 33.9ps and the fall time is 34.5 ps. Figure 2.9 Teledyne LeCroy SDA 820Zi-A 12G-SDI input with Belden 1694 A 1m Cable The SDA 820Zi-A measured the signal amplitude as mv, the rise time is 34.8ps and the fall time is 34.6ps. Page 9 of 24

10 Figure 2.10 Omnitek Ultra 12G-SDI input with Belden 1694 A 10m Cable Figure 2.11 Teledyne LeCroy SDA 820Zi-A 12G-SDI input with Benden 1694 A 10m Cable The difference in measurements between these two different products is well within expected limits and even between different units of the same model as shown in the following table. Omnitek Ultra 4K Tool Box Teledyne LeCroy SDA 820Zi-A Difference 1m Belden 1694A 1m Belden 1694A Signal Amplitude 801.7mV 794.6mV 7.1mV (0.88%) Rise Time 33.9ps 34.8ps 0.9ps (2.58%) Fall Time 34.5ps 34.6ps 0.1ps (0.28%) 10m Belden 1694A 10m Belden 1694A Signal Amplitude 538.4mV 546mV 7.6mV (1.39%) Rise Time 42.6ps 43.9ps 1.3ps (2.96%) Fall Time 44.7ps 42.4ps 2.3ps (5.14%) Page 10 of 24

11 3 Jitter Measurement 3.1 Understanding Jitter Measurement Definition of Signal Jitter Jitter is a significant, undesired issue with any communication link and can be defined as the temporal deviation from a periodic signal with ideal duty cycle. This is known as Time Interval Error (TIE). Ideal Clock Jitter Clock Difference Jitter Figure 3.1 Clock Jitter Diagram For 12G-SDI, Jitter from 10Hz (f1 Timing Jitter lower band edge ) to 1.2GHz (f4 - Upper band edge) is classified as Timing Jitter and jitter above 100kHz (f3 Alignment Jitter lower band edge ) is classified as Alignment Jitter. Jitter measurement ranges Clock extractor jitter transfer functions 0 db 0 db Timing jitter measurement range 0 db 0 db Alignment jitter measurement range f1 f3 f4 f1 f3 Jitter Frequency Jitter Frequency Figure 3.2 Jitter measurement ranges from SMPTE RP 192 Jitter tolerance is detailed in SMPTE RP 184, RP 192 and in the corresponding video format specification (SMPTE for 12G-SDI interfaces). See the Ultra Rx Jitter Tolerance section. Page 11 of 24

12 There are two fundamental categories of jitter, these are Random Jitter (RJ) and Deterministic Jitter (DJ). Random Jitter is noise that cannot be predicted as it has no discernible predictable pattern and due to its random nature will have a Gaussian distribution over time as can be seen in the following image. Figure 3.3 Gaussian distribution of Random Jitter Deterministic Jitter is repeatable and the peak-to-peak value of this is therefore predicable. This jitter category is subdivided into: Periodic Jitter (PJ) - which repeats in a cyclic manner Data-Dependent Jitter (DDJ) which is related to the bit sequences in a data stream Duty Cycle Distortion Jitter (DCD) which is related to the difference in rise and fall times of the signal and often the result of low slew rate signals In the following image a 4UI amplitude, 50kHz sinusoidal periodic jitter can be seen with its charactoristic (Dual-Dirac) distribution. Figure 3.4 Distribution of Deterministic plus Random Jitter Page 12 of 24

13 3.1.2 Jitter Measurement Method SMPTE RP 192 defines a number of different methods that can be used to measure jitter. The Omnitek Ultra employs method 6.2 Jitter measurement by means of a clock extractor which permits an Upper band edge of less than 1/10th of the clock frequency (1.2GHz for 12G-SDI). Note: The wide-band clock recovery block with a bandwidth of f4 is difficult if not impossible to implement if the specification puts f4 at a significant fraction of the bit rate. This difficulty is inherent in the nature of the scrambling used in SDI signals and is especially difficult when processing the EAV/SAV headers and the pathological signals. Because of this limit on the upper frequency for the clock recovery, some methods based on recovered clock as per figure 5 often provide only a partial result for timing and alignment jitter. As permitted by SMPTE RP 192, Omnitek, like many test and measurement equipment manufacturers, have adopted the same 5MHz Upper band edge for measurements made using the Jitter Meters. The Jitter Spectrum and Jitter Waveform instruments, however, display jitter up to 20MHz. Real Time Oscilloscopes, such as the Teledyne LeCroy SDA 820Zi-A, employ method 6.3 Jitter measurement by means of a Digital Sampling Oscilloscope and a Software (SW) based clock extractor which allow them to measure jitter at frequencies above 20MHz. Historically test and measurement manufacturers have measured jitter using a peak to peak method which measures the jitter within a specific time frame. In practice this may be based on a 0.5 second time span capturing 5,000,000 independent samples. Peak-to-peak measurement is a poorly defined measurement as it includes all random jitter and deterministic jitter with equal weight and each measurement taken over time can be significantly different. SMPTE RP 184 recommends that the Probabilistic Peak to Peak (p-p-p) method is used as this gives the probability for a given logic transition to differ from the ideal. This is stated as Jitter at 1 in 10 -x and is an estimate of the most likely p-p jitter in unit intervals when measured over 10 -x independent jitter samples. The p-p-p value of jitter is determined using a Gaussian probability density function (PDF) analysis of sampled data. This becomes a more accurate representation over time but for most practical purposes the time period is limited to 10 -x jitter samples. Refer to SMPTE RP 184 Annex C. For 12G-SDI jitter analysis the Ultra takes a set of 40,000,000 independent jitter samples each second and measures peak-to-peak jitter. The sampling and averaging / damping of the Ultra means that the peak to peak jitter measurement reported is very close to a p-p-p at 1 in This approach gives repeatable and comparable results to that of Probabilistic Peak to Peak measurement. Colour bars is recommended in SMPTE RP 184 as a non-stressing test pattern for jitter measurement. Some may measure jitter using a pathological stressing test pattern as a source, which in most cases will just be measuring the jitter introduced in the test equipment s input circuitry caused by trying to lock to a stressing signal and will cause erroneous results. Page 13 of 24

14 3.1.3 Noise Floor and Low Jitter Mode Noise Floor is the inherent noise introduced by receiving equipment and the measuring equipment itself. The significant difference in approach taken between high-end oscilloscopes and small form factor products means that high-end oscilloscopes have a very low Noise Floor. To ensure that the Noise Floor of the Ultra remains low, the system should be configured correctly. Single Clock this should be set in the Configuration menu. Disconnect Analyser Instruments this is done by removing the link from the SDI Inputs panel to the Analyser / Convertor panel in the Connections menu Disconnect Generator ideally this should be done by removing the link from the Generator panel to the HDMI Output, Display Port Output or SDI Outputs panel in the Connections menu. If the Generator is used the input and output video formats should also be the same. Free Run ideally the Ultra s system locking reference should be set to Free Run to avoid jitter introduced from the reference signal itself and the system s locking circuitry. Genlock should be set to the Eye input unless the Ultra s own output is being analysed. Note that the Noise Floor of the Ultra is lower than previous generations of test and measurement equipment so that the measured jitter on HD-SDI and 3G-SDI is now measured to a higher level of accuracy on the Ultra Rx Jitter Tolerance All receiving equipment needs to be tolerant to a specified level of jitter at different frequencies. Test and measurement equipment should be tolerant to a higher level of jitter, compared to other receiving equipment, so that jitter can be measured correctly and the measured transmitting equipment jitter is not affected by the receiver s input jitter. All generating equipment must only produce levels of jitter that are within specified SMPTE levels to ensure that all receiving equipment can decode the signal correctly without errors. To ensure that jitter generated by equipment is within specified SMPTE limits, it needs to be checked. Typically this is done during the design and manufacturing stages as well as when equipment is being commissioned. The following is the SMPTE RP 184 and RP 192 jitter tolerance template. The values are defined in SMPTE for 12G-SDI interfaces. Sinusoidal Input Jitter Amplitude A1 A2 f1 f2 Jitter Frequency Figure 3.5 Jitter Tolerances from SMPTE RP 192 and SMPTE f3-20db/decade slope f4 f1 = 10Hz f2 = 3.75kHz (calculated f2 = f3/(a1/a2) f3 = 100kHz f4 = 1.2 GHz (ie 1/10th clock rate) A1 = 8 UI (168psec) Timing Jitter A2 = 0.3 UI (25.2psec) Alignment Jitter Page 14 of 24

15 ,000 UI Ultra 12G-SDI Physical Layer Analysis The following is a plot of the typical Jitter tolerance against SMPTE RP 184 and RP 192 jitter tolerance template over a 1m and a 20m length of Belden 1694 A cable with an f4 frequency of 5MHz G-SDI Jitter Tollerance Frequency in khz Jitter Tolerance (20m) SMPTE Requirement Jitter Tolerance (1m) Figure 3.6 Typical input jitter tolerance This graph plots the jitter amplitude, at which input CRC errors are detected, against frequency. This plot has been generated by introducing increasing levels of jitter amplitude using the Ultra s own jitter generator for frequency from 0Hz to 20MHz. When CRC errors are detected in the receiving equipment this determines the jitter tolerance at the specific frequency. Note that the amplitude of jitter, inserted by the Ultra jitter generator, is not great enough to cause errors for frequencies up to 130kHz or for frequencies above 330kHz. Page 15 of 24

16 3.1.5 Jitter Transfer Jitter Transfer refers to the jitter on the output of equipment that is the result of jitter applied on the equipment s input. This kind of jitter can also be the result of jitter from an analogue locking reference such as black & burst. The compliant jitter transfer function can be obtained from SMPTE RP 184, RP 192 and SMPTE for 12G-SDI interfaces as shown in Figure 3.2 Jitter Transfer can be measured practically by generating a known level of jitter to the input of the device under test, for each frequency from f1 (10Hz for 12G-SDI) to f3 (100kHz for 12G-SDI) and then measuring the jitter at the output of the device under test. A non-stressing test pattern, such as colour bars, should be used to avoid introducing data dependent jitter into the measurements. The device under test and the jitter generator should not be connected to any external locking reference as this may introduce Output Jitter. Jitter Generator Device Under Test Jitter Analyser Figure 3.7 Practical setup for Jitter Transfer measurement Known levels of jitter at different frequencies can be introduced into the SDI output of the Ultra as described in Section 4.1 and jitter can be measured at the output of the device under test using Ultra s Jitter Meters instrument as described in Section 3.1. The results obtained at each frequency should be compared to the output of the Jitter Generator to take into account the jitter of the Jitter Analyser itself Intrinsic Jitter and Output Jitter Output Jitter is the total jitter measured on the output of the equipment itself and includes Intrinsic Jitter introduced by the equipment s own output circuitry and Jitter Transfer inherited from a signal connected to the equipment s input. A device s Output Jitter can be measured by directly monitoring the output of the device using a jitter analyser as shown in Figure 3.7. The device s Intrinsic Jitter can be measured when there is no input or locking reference connected. Page 16 of 24

17 3.2 Jitter Meters Instrument The Jitter Meters instrument displays 5 simultaneous jitter meters which display jitter above 10Hz, 100Hz, 1kHz, 10kHz and 100kHz. These can be used to determine the frequency band in which jitter is occurring. Figure 3.8 Omnitek Ultra Jitter Meters Instrument The horizontal position and the colour of the meter diamond shape markers against a traffic light coloured graticule indicate the level of jitter within the filtered frequency band. When jitter is within specified limits the marker is white and when jitter is out of specified limits the marker is red. When the jitter is on the edge of being out of specified limits it turns amber. The Jitter Meters instrument uses high pass filters of 10Hz, 100Hz, 1kHz, 10kHz and 100kHz which are used to isolate jitter above these frequencies and remove jitter below them. These can be used to determine the frequency band in which jitter is occurring. 0dB 10Hz 100Hz 1KHz 10KHz 100KHz 5MHz 40dB / decade slope Figure 3.9 Jitter Meter high pass filters These filters have a sharp cut-off to ensure that jitter frequencies below the specific filter frequency are removed in accordance with SMPTE RP 184 and RP 192. In the diagram above the different filter responses are shown in different colours. Note that although the filters used with the Jitter Meters cut off the high frequency jitter, the Jitter Spectrum instrument will display all jitter frequencies up to 20MHz. This allows out of bound jitter to be analysed. Page 17 of 24

18 3.3 Jitter Spectrum Instrument When analysing jitter, the Jitter Spectrum instrument can help determine the exact source or cause of jitter in the signal. This instrument performs a Fourier analysis of jitter frequencies in the signal being analysed. The results are displayed numerically and graphically for all frequency peaks as jitter amplitude against frequency from 5Hz to 20MHz. The Jitter Spectrum instrument automatically displays frequencies and amplitudes of the highest peaks. Figure 3.10 Omnitek Ultra Jitter Spectrum displaying a 2U amplitude, 500kHz jitter that has been inserted using the Ultra generator output The Jitter Spectrum instrument is very useful during the development of broadcast equipment as it will show jitter introduced by individual devices and different clocks within the equipment. The Fourier analysis will display the fundamental frequencies and harmonic frequencies of devices and system clocks. Within any design a number of different devices have specific power requirements. Buck Regulators and DC-DC convertors use a range of frequencies from 10s of khz to a few MHz. Within any FPGA design there will be a number of different clock frequencies in use (for example 150MHz). In addition to interference from these clocks, there is also the risk of beat frequencies being introduced. For example when the equipment supports both 60Hz and 59.94Hz there will corresponding 148.5MHz and MHz clocks which potentially can introduce a kHz beat frequency. Anywhere a clock is used there is a risk that jitter will be introduced by crosstalk and via power rails. The Jitter Spectrum instrument allows equipment to be tested on the bench and remedial action to be taken live. Page 18 of 24

19 3.4 Jitter Histogram Instrument The Jitter Histogram instrument displays the Probability Density function (PDF) of the jitter Relative Average Perturbation (RAP) as a function of frequency. Where the Probability Density Function defines the probability of a jitter frequency occurring within a range of frequencies determined by a normal distribution. Relative Average Perturbation is the average absolute difference between a period and the average of it and its two neighbours, divided by the average period. Figure 3.11 Omnitek Ultra Jitter Histogram displaying a typical Gaussian distribution for jitter of a 12G-SDI signal Figure 3.12 Teledyne LeCroy SDA 820Zi-A displaying a typical Gaussian distribution for jitter of a 12G-SDI signal Page 19 of 24

20 Figure 3.13 Omnitek Ultra Jitter Histogram displaying a classic bimodal (Dual- Dirac) distribution which is typical when the signal contains DJ (deterministic jitter) In this example a 4UI amplitude, 50kHz sinusoidal jitter has been inserted by the Ultra generator 3G-SDI output. Figure 3.14 Teledyne LeCroy SDA 820Zi-A displaying a classic bimodal (Dual- Dirac) distribution which is typical when the signal contains DJ (deterministic jitter) In this example a 4UI amplitude, 50kHz sinusoidal jitter has been inserted by the Ultra generator 3G-SDI output. The width of the histogram is determined by the range of the jitter on the signal being analysed. Ideally the histogram should be very narrow corresponding to a low value of SD (standard deviation). The depth of the valley between the two Dual-Dirac distribution peaks is an indication of the level of Random Jitter within the signal. Page 20 of 24

21 3.5 Jitter Waveform Instrument The Jitter Waveform instrument show the unfiltered jitter amplitude against time. This real time display is a useful tool to determine whether jitter occurs at a regular interval (ie Deterministic Jitter) or if it is random (ie Random Jitter). Figure 3.15 Omnitek Ultra Jitter Waveform displaying typical 12G-SDI signal jitter showing frame rate jitter. Figure 3.16 Omnitek Ultra Jitter Waveform displaying a 2UI, sinusoidal 20kHz jitter inserted by the Ultra generator 12G-SDI output The Ultra Jitter Waveform display shows all frequencies of jitter present in the SDI input signal up to 20MHz and therefore 100mS, 10mS, 1mS, 100uS and 10uS time-bases are provided to view specific frequency domains where jitter may be occurring. Page 21 of 24

22 Ultra Ultra Ultra 12G-SDI Physical Layer Analysis 3.6 Jitter Comparison of Omnitek Ultra and Teledyne LeCroy Jitter measurement using the Omnitek Ultra 4K Tool Box and the Teledyne LeCroy SDA 820Zi-A oscilloscope produces very similar results as shown in the following diagrams: Lecroy Figure 3.17 Linearity Plot of Ultra versus LeCroy Generated Ultra LeCroy UI UI UI UI UI UI UI UI UI UI UI UI UI UI UI UI UI UI UI The jitter measurement comparison was made using the Ultra Jitter Insertion function to insert specific levels of sinusoidal jitter at 500kHz as described in section 4.1. For each introduced level of jitter, corresponding measurements were made on another Omnitek Ultra 4K Tool Box 100kHz Jitter Meter (see section 3.2) and the Teledyne LeCroy SDA 820Zi-A oscilloscope using its cumulative distribution function (CDF) at 10-8 probability and using a band pass filter of 100kHz to 5MHz to correspond with the settings of the Ultra Ultra versus LeCroy Jitter Measurement Low Level Measurement Generated Ultra LeCroy UI UI UI UI UI UI UI UI UI Lecroy Figure 3.18 Low jitter level measurement The results of the comparison show a linear correlation between the values measured using the Ultra and the LeCroy. There are minor differences at low levels of inserted jitter which is indicative of the difference in Noise Floor between the two devices. Page 22 of 24

23 4 Output Signal Conditioning The 12G-SDI output (SDI4) on the Omnitek Ultra 4K Tool Box allows the generated output signal to be modified so that the susceptibility of the equipment under test can be checked. Typically the output jitter generation, signal amplitude and slew rate controls are used to emulate the performance of transmitting equipment and test the resilience of the receiving equipment when the signal has been sent over different cable lengths and cable types. These tools can be used to assess the jitter tolerance of receiving equipment at different jitter frequencies by increasing the jitter amplitude, for each frequency band, until CRC errors are detected in the receiving equipment. 4.1 Output Jitter Generation To ensure that equipment under test can tolerate known levels of jitter (jitter tolerance), the Jitter Generator allows jitter to be introduced into the generated SDI output signal at user-specified levels of UI (unit interval) and at user-specified frequencies. The introduced jitter is sinusoidal and is adjustable using the Connections tab by selecting the setup icon for the SDI output. Output Jitter is enabled by setting Enabled toggle switch to I. The Jitter frequency can be set between 0.001kHz and 3MHz using the numeric field and can be set in units and decimals of 1Hz. The amplitude of the jitter can be set in unit intervals or decimals of unit intervals. The maximum amplitude is 8UI for frequencies up to 120kHz. If the settings dictate a slew rate above the maximum, the actual rate achieved will be displayed. SMPTE RP 184, RP 192 and SMPTE ST detail the minimum receiver jitter tolerance that any piece of equipment receiving 12G-SDI should have. The introduction of known levels of jitter allows the jitter tolerance of the equipment under test to be checked against its specified tolerance levels. See the Ultra Rx Jitter Tolerance section for details of the Ultra s input jitter tolerance. 4.2 Output Signal Amplitude and Slew Rate To ensure that receiving equipment can still successfully decode the 12G-SDI stream with different signal amplitudes and slew rates these are adjustable using the Connections tab by selecting the setup icon for the 12G-SDI output (SDI4). Fast Slew Rate The signal amplitude can be set between 450mV and 840 mv. The signal slew rate can be set to Fast or Slow. Typical results are shown below. Slow Slew Rate Omnitek Ultra output slew rate set to fast (rise time 34.2ps, fall time 34ps) Omnitek Ultra output slew rate set to slow (rise time 37.7ps, fall time 37.4ps) Note that the slow slew rate increases the level of jitter (as shown by the thicker line of the Eye patter display) and therefore increases the risk of bit errors due to a narrower data sample area. In practice the higher the signal slew rate the larger the sample area which reduces the risk of bit errors. Page 23 of 24

24 5 Conclusion As can be seen in section 2 Eye Diagram Measurements the Omnitek Ultra 4K Tool Box compares very favourably with the Teledyne LeCroy SDA 820Zi-A high-end oscilloscope with high bandwidth PPL mode selected. It can also be seen in Section 3 Jitter Measurement that the Omnitek Ultra 4K Tool Box also compares favourably with the Teledyne LeCroy SDA 820Zi-A high-end oscilloscope for the jitter bandwidth measured by the Ultra. The results of the tests that have been performed, and demonstrated within this document, are well within acceptable limits for a test and measurement product of its class and as the first practical test and measurement equipment supporting 6G-SDI and 12G-SDI interfaces. The Ultra provides a practical and useable tool set available for equipment manufacturers and system integrators who are embarking on 4K / UHD and using 6G-SDI and 12G-SDI interfaces. It provides consistent and reliable results using dedicated instruments directly applicable to broadcast equipment manufacturers and installers. For further information about the Omnitek Ultra 4K Tool Box visit About Omnitek: Omnitek was formed in 2001 and specializes in video test and measurement equipment and in FPGA IP, development boards and design services for video-related products. In 2008 Omnitek was awarded a Queen s Award for Enterprise, for innovation. Published September 2015 Page 24 of 24