Intro:

The setup:

• Rdev setup :
  
• 140 MHz clock. Use bandwidth 10,20,35 MHz
  
• 10 MHz bandwidth:
• dec_f=3, upshift output (shift_fir)=5
  
• txSamples=5500, dataDelay=6250, data samples=20000
• i/o rate: 10.2 Mbytes/sec
  
• 20 MHz bandwidth:
• dec_f=2, upshift output (shift_fir)=5
  
• txSamples=11000, dataDelay=12500, datasamples=40000
• i/o rate 20Mbytes/sec 35.7 Mbytes/sec
  
• 35 MHz bandwidth:
• dec_f=1, upshift output (shift_fir)=5
  
• txSamples=19250, dataDelay=21000 datasamples=70000
• all bands gave 550 usecs tx samples, 2 milliseconds data
• data ranges  would have been:
  
• Start Height: 93.75 km
• last Height:  393.75 Km
  
• Rdev setup for sine wave.
• 140MHz clock at  @-2bm.
  
• 10 millisecond ipp. rf pulse  fed into rdev from sps.
  
• sine wave= 30MHz+ 1160 hz, amp=0dbm -> splitter than tx,and height samplers
• This should have been + 11160 hz.. 1160 does not give very many cycles in tx window.
  
• took 3 files about 10-15 seconds each.
  
• files: sasdr_final.20130204.b0a.00000, 00100,00300 (10,20,35 MHz).
• Used upshift =5 for all bandwidths.
  
• sps:
  
• running 10 milliseond ipp
• Software:
• test run on pdevs0.
• scripts in ~thall
• pnet --conf=/home.local/thall/pnet.conf.fir10M --fpga=/pcmn/sasdr_dds.gz --extclk  --dump=10
• pnet --conf=/home.local/thall/pnet.conf.fir20M --fpga=/pcmn/sasdr_dds.gz --extclk  --dump=6
• pnet --conf=/home.local/thall/pnet.conf.fir35M --fpga=/pcmn/sasdr_dds.gz --extclk  --dump=50
  
  

The Sine wave test:

• Page 1:plot spectra, check sine wave frequency.
  
• top spectra 1 ipp tx sample
• 2nd spectra 1 ipp height samples
• 3rd blowup tx spectra about sine wave
• bottom: blow height spectra about sine wave.
• Page 2: measure phase offsets:
• top: sine wave Height[smp0] - tx[smp0] for each ipp. The red line shows the expected difference.
• 2nd:check phase difference for tx[smp0] 1 ipp to the next.  Check that it remains constant
• Bottom:check phase difference for height[smp0] 1 ipp to the next. Check that it remains constant.
• Page 3: plot voltages start of  1st ipp (plotted real portion of sine wave).
  
• The black line is the sampled data.
• The red line is a sine wave fit to the tx samples and the plotted for the tx and height samples. 
• Top: first 1200 useconds. The green,blue lines show where the txWindow stopped and the height window started.
• 2nd:start of tx samples. Shows that the sine wave fit follows the tx sample data.
• 3rd:end of tx samples. Shows that fit still works at end of tx sample data
• bottom; start of height samples.You see the jump in data and fit.
  

• 10 MHz bandwidth (.ps) (.pdf)
• Sine wave peak voltage:5000 counts
• Height[0]-tx[0] phase difference: -11.5 usecs
• page 4 shows data - sinewave fit.
  
• 20 MHz bandwidth (.ps) (.pdf)
• Sine wave peak voltage:2500 counts.
• Height[0]-tx[0] phase difference: -11 usecs
  
• 35 MHz bandwidth (.ps) (.pdf)
• Sine wave peak voltage:1500 counts
• Height[0]-tx[0] phase difference: -10.7 usecs
• spectra spurs height spectra down at least 80db
• spectral width around sine wave probably narrow data window.

• maximum input is .786 Volts peak to Peak. This is .28 volts rms.
• .28 Volts rms i  is log10(.28^2/2) + 30 = 2 dbm. So the max power would probably be a little higher than this.
• If the input sine wave was -4dbm ( 0dbm, split then 1db loss in the cable). this is 6db below the peak value.
• If the peak counts was 32767 then we should have had: 32767/sqrt(4.)=16384 counts. We had less than this:
• 10 MHz 5000 Vcounts-> Voltage 6.5 times low, 43 time low in power
• 20 MHz 2500 Vcounts-> voltage 13.2 times low, 172 times low in power
• 35 MHz 1500 Vcounts-> voltage 21.8 times low, 477 times low in power.
  
• Used upshift value of 5 (2^5) for output. New upshift values for  given bandwidths:
• 10 MHz: 5 + alog2(43)= 5 + 5.4  = 10
  
• 20 MHz: 5 + alog2(172)= 5 + 7   = 12
• 35 MHz  5 + alog2(477)= 5 + 8.9=14
  

Summary:

• The frequency of the sine wave is correct.
  
• Within 1 ipp, tx to data window sampling is stable for the duration of the test (10-15 secs).
  
• The tx to height phase is off by about 11 usecs.
  
• It changes a little depending on the bandwidth.
• the tx  and height starting phase is stable for all ipps.
• new output fir upshift bits should be:
  
• 10MHz - 10
• 20 MHz - 12
• 35 MHz - 14 

<- page up
home_~phil