Transceiver Design and Preliminary Evaluation

1
Transceiver Design and Preliminary Evaluation

• John Musson

2
General Requirements List
3
Field Receiver Requirements
4
AM vs PM Detection

• Gradient measurement (and drive) is AM ? linear
• Detectable at or just above noise floor for
  synchronous (or near-synchronous) detection 9
• BW is determined by risetime, and fixed.
• Phase measurement (and drive) is PM ? non-linear
• An additional 12-15 dB is required to overcome
  Threshold Effect 9
• True for I/Q as well!
• BW is determined by risetimes, but also by
  position and trajectory on constellation 11 it
  can be time-varying.

Figures courtesy RF Design Magazine, Jan. 2007
5
Field Receiver Specifications
6
VSWR and Mismatch Errors (?)

• In a true 50 O system, we seek to quantify
  minimize mismatch errors.
• A receiver VSWR of 1.11 assures lt 0.5p-p
  accuracy (/- 0.04 dB) for load mismatches up to
  1.21.
• Our probe cables are NOT 50 O..
• However, our calibration fixture IS
• 1.21 is relatively easy (and cheap!) to realize.
  ?1.11?
• We retain LLRF module accuracy and repeatability.
  

7
Field Receiver Specifications (cont.)
8
ADC Performance

• Effective Dynamic Range
• EDR -1.25 6.02b 10log fs
• b of bits, fs sample frequency
• 1 Hz BW
• DR gt Analog, and LSB MDS
• Noise Figure can be assigned
• Function of sample rate and of bits
• S/N degradation from sample clock jitter 7
• Reference Frerking, M., Digital Signal
  Processing in Communication Systems

9
ADC
10
Transmitter
11
Latency Budget
12
(No Transcript)
13
Design Strategy

• Utilize available strong Probe Input signal to
  maintain good (SN)/N, while balancing
  non-linearities to control THD.
• Use as few front-end components as possible
  avoid narrowband passives, as well as excessive
  actives. Most importantly, minimize thermal drift
  effects
• Side with nature and use high-performance
  devices!
• Add thermal monitoring for anything
  unanticipated.
• Achieve as much using analog as is feasible
• Generic front-end is re-useable

14
Mussons Vision
15
Block Diagram
Test Signal (loopback)
System Cost 1000.00
Pdiss 3.0W
16
Pout - 14 dBm Pnoise - 124 dBm IIP3 340
dBm S/N 90 dB C/I 360 dB
VSWR 1.11
Pin 20 dBm BW 100 kHz
(S/N derived for 0dBm)
Isol gt 70 dB
-10 dBm loopback for self-test
17
Pout - 26 dBm Pnoise - 124 dBm IIP3 61
dBm S/N 78 dB C/I 82 dB
Pout 11 dBm Pnoise - 83 dBm IIP3 54.7
dBm S/N 74 dB C/I 70 dB
Pout - 3 dBm Pnoise - 96 dBm IIP3 56
dBm S/N 75 dB C/I 73 dB
Gain - 9 dB NF 50 dB
18
Tx
19
Rx 2,3,4

• Identical to Field Probe channel, but lacking
• High-IP3/Level 17 mixer
• Replace with MCL-type Level 7 mixer to preserve
  LO drive levels. NF should be unaffected.
• Degradation in IIP3 expected
• IIP3 47 dBm
• C/I Pin 20 dBm 54 dB (0.2 THD, 0.7)
• Quality KL IF filter
• Employ discrete component filter to save cost.
  Performance degradation should only involve IL
  and increased phase vs temperature variations.

20
I/O Connector, Physical Layout
Signal Guarding
TTL
21
Our First Prototype 2-Channel Receiver!
22
Tangential Sensitivity / NF

• -81 dBm - -77dBm range for 10 kHz BW (PTI Xtal
  filter). SN/N difficult to establish since
  converter has loss, not gain. Spectrum Analyzer
  used as a tuned 70 MHz detector.
• Assuming 100 kHz BW (per spec.) we have -70 dBm
  sensitivity, achieving the 68 dB S/N (amplitude
  control) for lower input of 0 dBm. We also almost
  make 72 dB for phase control requirement.Will
  easily re-gain w/ digital filtering.
• NF 54 dB (50 dB calculated)

Representative data.actual results may vary!!
Literature suggests 3-8 dB of SN/N present 6.
Weve assumed 0 dB for safety..
23
Rogue!
Slope 1.012
System Gain -12.17 dB
24
Two-Tone IMD Test for IIP3
Improve Two Tone, Third Order Testing, Mini
Circuits Tech Note
25
Pin 20 dBm
2-Tone Input (1497 MHz)
100 kHz
!!
26
Pout 8 dBm
2-Tone Output (1497 MHz)
100 kHz
27
(No Transcript)
28
Environmental Chamber
29
-0.213 degrees / C
-0.38 ps / C
30
Misc.

• Input VSWR 1.41
• LO Input Impedance 1.51 (with representative
  atten.)
• Output Impedance 100 Ohms (diffl, w/ balun
  transformer)
• System Gain -12 dB
• Initially, ¼-wave stub less than stellar.
  Re-sizing resulted in 1.61 with load (not
  mixer). 201 at 70 MHz. Discrete diplexer works
  to give 20 dB additional isolation (10 dB each
  leg) with 2 dB IL (currently in use on Injector
  upgrade). We expect much more isolation, and less
  IL from stub.
• Isolation LO-RF, RF-LO -62 dB
• Power Consumption 375 mW per channel

31
Test Summary
32
Test Summary (cont.)
33
Digital Filtering

• The proposed 56 MHz sample clock for ADC provides
  1401 over sampling
• 10 dB S/N improvement expected from CIC
  decimation FIR, etc.(gt20 dB is theoretically
  possible!) 7
• Overall latency ltlt 1us

34
Summary How did we do?!

• Good
• Outstanding amplitude stability vs temp
• Not discernable (lt0.02dB )with HP8507A over 25 C
  range
• Linearity and S/N per model
• Additional gains from digital filtering
• Phase comparable to Injector LLRF upgrade
• Loopback and diagnostic capability
• Bad
• Re-visit SmartLoad, ¼-wave stub, stuffing
• Refine layout microstrip calcs
• Good construction practices
• Ugly.TBD!
• Possible additions/changes
• Discrete tx 70 MHz filter
• In short, concentrate on layout and testing

35
References

• Introduction to Radar Systems, 2nd Edition,
  Merrill Skolnik, McGraw-Hill, New York, NY, 1980
  (ISBN 0-07-057909-1)
• Local Oscillator Phase Noise and its Effect on
  Receiver Performance, John Grebenkemper,
  Watkins-Johnson Tech Note, Vol. 8 No. 6 Nov/Dec
  1981.
• Sensitivity Analysis of Radio Architectures
  Employing Sample and Hold Techniques, M.C.
  Lawton, Hewlett Packard Laboratories Tech Note.
• Aperture Uncertainty and ADC System
  Performance, Brad Brannon, Analog Devices
  Application Note AN-501, September, 2000.
• Understanding the Effects of Clock Jitter and
  Phase Noise on Sampled Systems, Brad Bannon, EDN
  Magazine, December, 2004.
• Receiver Dynamic Range Parts 1 2, Robert E.
  Watson, Watkins Johnson Tech Note, Vol. 14 No. 2,
  Mar/Apr 1987
• Digital Signal Processing in Communications
  Systems, M. Frerking, Chapman and Hall, New York,
  NY., 1994 (ISBN 0-442-01616-6)
• Communications Receivers, 2nd Edition, U. Rohde
  et al., McGraw-Hill, New York, NY, 1997 (ISBN
  0-07-053608-2)
• Digital and Analog Communication Systems, 3rd
  Ed., L. W. Couch, Macmillan, New York, NY. 1990
  (ISBN 0-02-325391-6)
• Fundamentals of RF and Microwave Power
  Measurement (AN 64-1) , Hewlett-Packard Tech
  Note
• Multimode RF Transceiver Advances WEDGE System,
  RF Design Magazine, Jan. 2007 pp. 44-49

36
Fundamentals of RF and Microwave Noise Figure
Measurements, HP Tech Note 57-1 Noise Figure
Measurement Accuracy- The Y- Factor Method, HP
Tech Note 57-2
Radio Astronomy, J. Kraus, Cygnus-Quasar, 1988
37

Introduction to Radio Frequency Design, W.
Hayward, ARRL, 1994
38
Transceiver Calibration Fixture
39
Transceiver Calibration Fixture (cont.)
40
Transceiver Calibration Fixture (cont.)
41
My Useful Math

• To 290 K (IEEE)
• KTo -174 dBm
• NF T/290 1
• F 10 log NF
• IIP3
• Pintermod 3Ptone 2PIIP3
• NFnet

• SFDR3 2/3 (IIP3 174 F -10log BW)
• SFDR2 ½ (IIP2 174 F -10 log BW)
• Pphase noise Punwanted 10log BW Prx phase
  noise

42
Receiver Bandwidth

• SDR has 2 associated bandwidths
• Analog
• Minimum element in Front End
• IF / Digital
• Generally the narrowest, set by IIR / FIR
• DR Calculations should use the analog BW
• SNR should use narrow/digital BW
• BW determined largely by sensitivity (KTB) and
  latency (Group Delay)
• Ex. JLAB LLRF Rx uses a 8 MHz BPF exhibiting 100
  ns of latency

43
HP 8508A Gold Standard
Frequency 300 kHz 2GHz _at_ 1.5 GHz, AB100mV
Absolute Amplitude Accuracy /- 1dB
Relative Amplitude Accuracy /- 0.6dB
Absolute Phase Accuracy /- 4 degrees
Relative Phase Accuracy /- 0.4 degrees
http//www.caip.rutgers.edu/kahrs/books/sampling.
html
44
I/O Signal List
45
BOM Power Dissipation