Practical Cryogenic Receiver Front Ends for Commercial Wireless Applications

Transcription

1 Practical Cryogenic Receiver Front Ends for Commercial Wireless Applications Balam A. Willemsen Santa Barbara, CA USA TH4E-3

2 Outline Introduction Background Cryocoolers and Reliability Design for Low Cost Filter Trimming Conclusions Slide 2

3 Motivation Cryogenic electronics have been demonstrated in specific applications - Radio astronomy, military, satellite, - All are low volume, not cost driven Can it be made practical for commercial applications? - i.e. reliable, low cost and reproducible Slide 3

4 Effect of CRFE on Receiver Noise In the presence of out-of-band interference SuperLink Enhances Uplink by Reducing In-Band Noise in the BTS Receiver Slide 4

5 Summary of trial results 10 U.S. trials in 850 MHz CDMA networks CRFE consistently significantly reduced handset transmit power Category Handset transmit power change (db) # Average Urban Suburban Rural OEM A OEM B OEM C Slide 5

6 CRFE Field Trial Data Handset Power Reduction Before After Mobile Tx Power < -5.0 dbm 5.0 to 0.0 dbm 0.0 to 5.0 dbm > 5.0 dbm Median Mobile Tx Power dbm dbm Tx Power = db Slide 6

7 STI Sapphire Cryocooler Benefits and features Hermetically sealed - No maintenance required Small and compact Ibs. (2.8 kg) in. OD x 11.8 in. (92mm OD x 300mm) Efficient ~100 W/5 W lift Units in field running 24/7 > 6,000 Cumulative run-time > 230,000,000 hours MTBF > 1,000,000 hours Slide 7

8 CRFE field reliability Measured system MTBF > 500,000 hours Cumulative Hours 200,000, ,000, ,000, ,000, ,000,000 75,000,000 50,000,000 25,000,000 24/7 always on operation NO maintenance 0 01/ /2000 Cumulative Hours Num Shipped 01/ / / / / / / / Number Of Units Shipped Slide 8

9 SuperLink Rx Components Microprocessor Controller Cryocooler Dewar HTS Filter Chip Filters & LNA s Slide 9

10 Relative Costs of major CRFE sub-components Significant cost reduction requires effort in all areas System, 13% Cold Circuits, 30% Cooler, 24% Cryopackage, 33% Slide 10

11 Design for Low Cost Major effort to reduce cryopackage cost from Reduce part count - Cold Stage + 3 microenclosures replaced with one piece microenclosure. Simplify parts - One sided machining - Reduce number of threaded components Use spring clips where possible Eliminate tuning screws Slide 11

12 Inside the CRFE dewar Slide 12

13 CRFE microenclosure x/dewar Slide 13

14 Computer Assisted Filter Trimming (CAFT) 2004 dewar included 60 tuning elements and required ~ 1 hour to tune by hand Introduced CAFT approach - Incorporate tuning features into filter layout Works very well for standalone filters Mismatches from LNA, switches, cables, in rf chain degrade return loss Extend to trim filters in-situ to tune full rf chain Slide 14

15 Typical Trimmable Filter Complex 10 resonator filter with 3 cross couplings G. Tsuzuki et al. IEEE MTT-S 2005 Slide 15

16 Simplify model of rf chain Z1,θ1 M G Z2,θ2 Abstract physical circuit as a mismatched transmission line and idealized gain block - Adds five non-filter parameters to the model Trim filter only to optimize entire rf chain - Measure once, trim once Slide 16

17 Dewar Trimming Process Assemble microenclosure in dewar - (filters, LNAs, cryocables, switches, ) Cool down and collect dewar S-parameters Perform parameter extraction on dewar S- parameters Determine filter trimming recipe Apply filter trimming recipe to set of filters in dewar Cool down and collect final dewar data Weld dewar for final system integration Slide 17

18 Typical performance of one CRFE rf path Measured data Slide 18

19 Conclusions Long-life, high-reliability cryocoolers enable the use of CRFEs commercial wireless networks - 24/7 maintenance-free operation - The cryocooler alone does not dominate either system cost or reliability All components must also be designed for lower cost Computer-aided filter trimming (CAFT) enables: - Lower cost: Simplification, reduced labor - Higher performance Slide 19

20 Acknowledgement The author wishes to acknowledge the assistance and support of everyone at STI, in particular R. B. Hammond, N. O. Fenzi, E. R. Soares, S. M. Bilski, G. Tsuzuki, A. Karandikar and A. Fiedler for many useful discussions regarding this presentation Slide 20