25jun13: wiggle polA,B cables at dewar,postamps

• Wiggle polA,B cables at output of postamp chassis.
• Wiggle cables between dewar and postamp chassis
• Tap on the dewar.
  

• xband receiver centered at 8800 Mhz
• 2-12 ghz mixer
• interim correlator. 25 Mhz  bandwidth,  9 level sampling  both pols, and 1 second sampling
  

• The black trace is polA, the red trace is polB.
• without touching anything, polA would have jumps  > .5% in the total power.
• Page 1:
  
• Top frame: 400 seconds of data.
• The black trace is polA, the red trace is polB
• The dashed red lines show where the polB cable was wiggled.
• 2ndFrame: 900 to 1400 seconds.Wiggle polA cable, tap on dewar
  
• The dashed and solid red lines show where the polA cable was wiggled
• 984 secs: output of postamp chassis
• 1009 wiggle cable between postamp chassis and dewar (postamp chassis side)
• 1087 wiggle cable between postamp chassis and dewar (dewar side).
• 1114 solid line. stop wiggling dewar side of cable.
• the blue lines show where the dewar was tapped on.
• 1230 secs, hit dewar once
• 1250-1270 seconds tap on dewar once a second.
• PolA value was jumping up to .5% but the jumps were not correlated with the wiggling or tapping
• 3rd frame: large jump in polA while dewars,cables left alone.
• a 1% jump in PolA total power at 2880 seconds. We were not wiggling or tapping at this time.
• bottom frame: wiggle polB cable.
  
• The polB cable between dewar and postamp chassis was wiggled at 3546 and then at 3563.
  
• At 3546 the signal jumped down by 2% and then returned.
• At 3563 the signal went down by about 1.3 % and remained there.
• Page 2: remove the circulator for polB that is next to the dewar.
• the output of the dewar has a circulator and then the cable that goes from the dewar to the postamp chassis. For these tests, the polB circulator was removed.
• 
  

Summary 25jun13:

• polA has intermittent jumps of up to 1%. These jumps don't seem to correlate with wiggling the cables or tapping on the dewar.
• polB  has trouble when the cable between  the dewar and the postamp is  wiggled.
• Tapping on the dewar did not seem to make a difference.
  

02jul13: swap cables at output of dewar

    On 02jul13 we swapped the polA, B outputs from the dewar to see how the instabilities changed.
The diagram below shows the arrangement for the straight through and cable reversed tests:

file:///home/phil/public_html

• xband receiver. center freq 8800 Mhz. 2-12Ghz mixer
• interim correlator, 25 Mhz bandwidth, 9 level sampling, 1 second sampling
• The telescope was sitting at az=285, za=4
  

• Top frame: straight through test.
• black is Dewar output polA through polA cable
• red is dewar output polB through polB cable
• the black lines (dewarPolA, Cable polA) has the jumps
• The vertical black lines show where we wiggled the polA cable at the if/lo cabinet. There was no change from the wiggling
• The vertical red lines show where we wiggled the polB cable at the if/lo cabinent. There was no change.
  
• Middle frame: Cables reversed test.
• The black line is polA cable coming from dewar polB output
• it  has two jumps, but it does not have the instablities
• The red lines is polB cable coming from dewar polA output.
• You can see that instabilities.
• So the problem stayed with the dewar polA output. This means that the problem is in the dewar or the bias box
• Bottom frame. Blowup cables reversed test.
• The dashed,solid vertical lines show where we wiggled the cables.
• Red vertical lines: we wiggled polB cable (coming from dewar PolA output). There were no jumps when we wiggles the cable.
• Black vertical lines: we wiggled polA cable (coming from dewar polB output). There was a jump in the power when we wiggled this cable.
• The polB  jumps when we wiggle the cable happened when both the polA and the polB cable were attached to polB dewar output.
• This means that the polB jumps are probably coming from the feed through connector of polB in the dewar.
  

Processing:x101/130702/xbtest.pro

02jul13 PolA instability, high time resolution data, bias monitoring

    The rf cables were returned to their normal positions. I then took some high time resolution data with the mock spectrometer. The setup was:

• xband receiver centered at 8800 Mhz. 2-12 Ghz mixer used
• Mock spectrometer with : 172Mhz bandwidth, 4096 channels, 16 bits, 1 millisecond sampling.
• Data was taken for 60 seconds.

• Page 1:total power time series
• Top frame: total power vs time for 60 seconds.
• Black is polA, red is polB
• Bottom frame: blowup in time of polA total power vs time
• There are discrete jumps in the total power output. The size of the jump is about 1.7% of Tsys. The jump lasts from 50 to a few hundred milliseconds.
• Page 2: spectrum of the total power time series.
• Top frame: 0 to 80 Hz.
• Black is polA, red is polB
• The frequency resolution is 16 millihz.
• PolA has a lot more power out to about 10 hz..
• There is a small spike at 60 hz.
  
• Bottom frame: blowup in frequency showing 0-10 hz.
• The largest spike is at 1.8 hz

• Xband has one amplifier.
•  The plots have
•  1 page for the  currents and one for the voltages
• The black trace is polA, the red trace is polB
  
• 28jun13-02jul13 (.pdf)
• Starting midday on 01jul13 the polA dewar current started showing extra noise.
  
• PolB looks clean
• Other jumps may be from the sband transmitter.
• The currents look clean (except for possible sband jumps).
  
• 02jun13-02jul13 (.pdf)
• you can see the jumps in the current starting 01jul13.
  

05jul13: drive xband dewar with cbh bias box.

• xband receiver centered at 8800 Mhz.
• the cbh bias box was used to drive the amps in xband
• the telescope was sitting at az=285, za=2.5 deg
• The mock spectrometer took data with: 172 Mhz bandwidth (centered at 8800 Mhz), 1024 channels, 16 bit spectra, 1 millisecond sampling, for 60 seconds.
  

• Page 1:total power time series
• Top frame: total power vs time for 60 seconds with 1 millisecond resolution.
  
• Black is polA, red is polB
• Bottom frame shows the same data smoothed to 100 milliseconds
• No jumps are seen in polAor polB.
  
• Page 2: spectrum of the total power time series.
• Top frame: 0 to 80 Hz.
• Black is polA, red is polB
• The frequency resolution is 16 millihz.
• There is a small spike at 60 hz. 
• polA and b show the same spectrum
  
• Bottom frame: blowup in frequency showing 0-10 hz.
• No extra power seen in pola.

Summary:

• Instabilities were seen in both polA and polB of xband. They are two separate problems.
• The problem was reported by project A2675.
  
• PolA:
  
• The problem is in the dewar or the bias chassis.
• The bias currents for polA show lots of noise starting 01jul13.
• The polA jumping was also present on 25jun13 but the bias currents for polA don't show a lot of noise
• The polA jumps were occurring more often on 02jul13 than 25jun13 so it may be that the bias monitoring on 25jun13 missed the jumps (the bias monitoring is only measured once every 30 seconds).
  
• With 1 sec time resolution, the amplitudes of the jumps looked  random.
• Using  the 1 millisecond data, we see
•  that the amplitudes of the jumps are always about 1.7% of Tsys.
•  The duration of the jumps varies from 10's of milliseconds to a few hundred milliseconds.
• With 1 millisecond dumps we did not resolve the transition time
• resolution:
• Using the bias box from cband high made the polA jumps go away. So the problem was in the xband bias box.
  
• PolB:
• Jumps occur when we wiggle the cable attached to the polB output of the dewar
• The jumps occur when the polA or polB cable is used to connect to polB output of the dewar
• The problem must be in polB feed though in the dewar.
  

home_~phil