Alfa glitches from dewar monitoring

27apr12: data with alfa cover on.

• alfa centered at 1450 Mhz.
• Mock spectrometers setup to take 86 Mhz bandwidth spectra (polA,B) on all 7 beams.
• 20 seconds of 1 millisecond sampling was done
• This was done with the alfa dewar monitoring running and then repeated with the dewar monitoring turned off.
  
• 30 seconds of 65 Usec samples was also done with the monitoring on (to compare with pulsar search setups).

• Compute the total power for each 86 Mhz spectra taken
• compute the spectra for the 64 usec data:
• abs(fft(totalPowerSamples)).

• 27apr12: alfa cover on:
• Dewar monitoring on:
  
• 1 millisecond total power samples (.ps) (.pdf):
• Top frame is polA, bottom frame is polB
• each colored strip is a different beam. Offsets have been added for display.
• The glitches every 5 seconds is the dewar monitoring reading the bias voltages,currents for the 3 amp stages.
  
• The higher frequency glitches inside each group is reading a single amp stage bias voltage or current.
  
• 65 usec spectra of total power (.ps) (.pdf)
• The vertical scale is Tsys/2 (I threw away the symmetric part of the transform without multiplying by 2.).. but Tsys with the cover on is less than looking at the sky.
  
• the spectra show power at 20.82 Hz. This is the rate that the individual amp stages are read. You can easily see 4 harmonics of this.
• The 20.82 Hz power is spread by the 5Hz sampling. It is also spread because the timing is set by a computer program that is generating the read requests. It is not tied to the station clock.
• Dewar monitoring off:
• 1 millisecond total power samples (.ps) (.pdf):
• The glitches every 5 seconds are now gone.
• beam 6 was jumping around because the its biases hadn't been adjusted yet.
  
• 65 usec spectra of total power (.ps) (.pdf)
• Page 1: 0 - 100 Hz.
• the 20.82 Hz power is now gone. There are still few spikes present
• Page 2: blowup showing 0-60 Hz.
• The remaining birdies are at:
• 15 Hz.
  
• 17.65Hz with 1 or two harmonics
• 24.83 Hz
  
• there is also something around 2.6Hz. We need to take a longer data set to look at this.
  

31may12: P2030 search data.

    P2030 pulsar search data from 31may12 was examined to see how strong the dewar glitches were.Since this was regular observing the alfa was looking at the sky.

• Datafile used: p2030.20120531.B1907+10.b0s0g0.00000.fits
• 132 seconds with 5 second cal at the start
• 1450 cfr with 172 Mhz bandwidth0, 65 usec sampling.
  
• B1907+10: 
• Period : 0.2836 secs
• dm: 148.4
  
• flux .7 mJy @ 21cm.
• The total power over 172Mhz was computed for each spectral sample (65 usec sampling).

• Page 1: time domain. power vs time.
• Top: this is the entire 132 seconds.
• The jump at the beginning is the cal
• the glitches every 5 seconds come from the dewar monitoring.
• The red vertical lines show the data for the next plot.
• 2nd: blowup 13 to 21 seconds.
• You can see two dewar monitoring groups spaced by 5 seconds.
• 3rd: blowup 13.85 to 14.95 seconds.
• This is one group of measurements from the dewar monitoring.
• Bottom: blowup 14. to 14.2 seconds.
• The individual reads can be seen. These are spaced by .04756 secs or 21.027 Hz.
• :Page 2: spectra of total power time series.
• a 2^20 length transform was done (excluding the cal ). This is about 68 seconds of data.
• The spikes with green lines under them are the pulsar and its harmonics.
• The red vertical lines show the dewar monitor power. This agrees with the data taken with the cover on alfa.

Summary

• The alfa dewar monitoring is creating  periodic power at 20.827 Hz and harmonics of this.
• It occurs in all beams and both pols.
  
• The dewar monitoring timing is set by a computer program so it is not locked to the station clock.
• A group of samples is taken every 5.x seconds.
• We need to disable the alfa dewar monitoring during alfa pulsar observations.
• This will take some programming since the same program is used to set the hi,lo cal.
• We also need to check if  spectral line observations are being affected by these glitches.
• There is other periodic power that needs to be located and removed:
• 15Hz, 17.65 Hz, 24.83Hz,, and probably 2.65 Hz.