Intro

    The sparameters for cal input to dewar output were measured by  felix on  10feb23. This had the dewar as delivered by hamdi.
     The s11 measurements showed reflections from the 2 sections of cal cable that went from cal input to cal coupler. These cables were standard sma cables. These measurements also included a 10db pad at the input to the cal coupler (at 20K).

    Felix/carlos replaced the 2 sections of sma cable with a single stainless steel cable. This should remove the closer reflections (from the first section of cable). It should also help with the dewar temp since the stainless steel cables would conduct less heat. They also removed the 10db pad that was in front  of the cal coupler. If we had reflections  from the input to the cal coupler, you would expect S11 to increase (by at least 10db).

    On 28mar23 felix remeasured the s-parameters in the lab with the dewar cold (after the stainless steel cal cable was installed, and the 10db pad removed). The setup was:

• The receiver was in the lab and cooled.
• absorber was placed in front of the window.
• the field fox was used to do the 2 port measurement,.
• s1 input : cal sma input on dewar
• s2 output: lna sma output
  
• the s-parameters  were measured between 2 and 14 GHz. 1001 channels were used.
• the reference plane was the cal sma input and the lna sma output.
  

---

The S-parameters:

    Plots show the

• The s-parameters: s21 s11,s22,s12  
  
• The acf computed from each sparameter spectra
• The acf was computed as:  abs(fft(spc))[0:lenfft/2-1] (since it is symmetric)
• I then divided by the integral of lag[0] to the max lag.
• The lag axis was:
• max Bw:14-2 GHz = 12GHz
• shift rate: 1/12GHz (.083 ns resolution)
• max lagdelay: .083*500=41 ns.
  
• The signal path is:
• Cal  input
• steel cal cable
  
• no 10db pad
  
• coupled input (20db )
• highpass filter
• lna
• output
  

    The plots show the measured s-parameters (.ps) (.pdf):

• Black is polX, red is polY
• Page 1: S21 : gain: input cal port, output lna output
• Top S21 gain:
• there is cable loss + 20db of attenuation from coupler before getting to the lna
  
• the cumulative gain is
  
• polX: 14db at 2.5 ghz to 5db at 14 GHz
  
• polY: 11db at 2.5 ghz to 5db at 14GHz
• The previous measurement was:
• PolX:7db at 2.5Ghz to 0Db at 14Ghz
• PolY:3db at 2.5Ghz  to 0db at 14Ghz
• The measured gain increased by 7db
  
• If the 10db pad was 10db when cooled, then you'd expect the gain to increase by 10db.
• It may be that the longer stainless steel cable has 3db more loss that the two shorter sma cables (or the 10db was not 10db when cooled).
  
• the amp gain (measured by hamdi)
  
• polX 40db at 2.5 GHz to 36 db at 14Ghz
• polY  35db +/- 1db 2.5 to 14 GHz
• Bottom: acf of S21
• there is ripple power at 5 to 7 ns.
• Page 2: S11: reflections at cal input.
• Top: spectra of S11
• The level stays below -12db for the entire band.
  
• this is 3 to 4 db  better than the previous measurements with the sma cable installed.
  
• Bottom: acf of S11
• there is a single strong ripple at 9 ns.
  
• the previous measurement had multiple ripples because there were 2 sections of cable.
  
• The acf has been normalized to the total power.
• PolX ripple is about 2.5 times stronger than the polY.
  
• a 9ns ripple with cable velocity of .7 would be a distance of .945 meters (37 inches) which is about the length of the stainless steel cable.
• With the 10db pad (in front of the coupler) removed, you would expect S11 to increase by at least 10db if the previous reflection was from the input of the coupler.
• The measured values looks like the previous reflection may have been from the input of the 10db pad rather than the coupler input.
• the 10db pad was a standard mini-circuits attenuator (not a special cryogenic one).
  
• Page 3: S22: reflections at lna output
• Top: spectra of S22
• values range from -20db to -10db.
• Bottom:acf of s22
• ripples similar to S11 with a few added frequencies,.
• Page 4: S12: lna output to cal input.
• Top spectra
• < -50db
• Bottom: acf
• no ripples stand out
• Page 5: plot ACF's for all parameters on 1 page.
• I plotted the Acf on 1 page to compare the freq ripples
• I converted the x axis to distance using .15 m/ns
• these distances assume the velocity is c. The should be multiplied by the velocity in the cables (say .7)
• S11 and S22 show some of the same ripples.
  

   

SUMMARY:

• The s-parameters from the cal input to the lna output were measured.
• What changed when the stainless steel cable replaced the 2 sections of sma cable and the 10db pad at the input to the cal  coupler was removed.
  
• The multiple ripples seen with the 2 sections of sma cable are now a single ripple at 9ns.
• The 37" distance is about the length of the single stainless steel cable,
• S21 increased by 7db rather than the 10db you would expect by removing the 10db pad
• either the pad was not 10db when cooled
• or the loss in the longer stainless steel cable is larger than the 2 pieces of the previous sma calbles
• or both..
• S11 did not get larger when the 10db pad was removed.
• Reflections must have been coming from the input to the attenuator rather than the input to the cal coupler.
  

processing: x101/230328/sparmswbrcv.pro