14jun12: glitches gone with bias mon
The alfa dewar monitoring was causing
glitches in the receiver. They would show up as multiples of
20 Hz in the p2030 data.
On 14jun12 a test was done to see if the glitches
would go away if we stopped reading the bias values, or we had to
stop all alfa mon reads (temperature, cals, etc.) to get rid of the
glitches. The setup was:
- alfa cover was off
- mock spectrometers 172 Mhz, 128 channels, 65 usec dumps. Used
the 1450 Mhz band.
- Software was written to turn off just the alfa bias
monitoring. The temperatures, amp status, supply voltages were
still being read.
- take 300 seconds of data with bias monitoring off, other
- take 300 seconds of data with all monitoring off.
The data processing included:
The plots show the results of the test:
- Input each file and compute the total power for each
sample and beam.
- normalize each total power time series to its median value.
- Do a 4 million point fft on the total power data.
- Plot the spectra of the total power vs freq.
total power spectra with
bias mon off and all monitoring off
|Bias Reading off
|all readings off
- 500 Hz bw:
- 60 Hz, 120Hz, 240Hz, 360Hz are strong independent of the
bias monitoring state.
- The 120,240,260 are harmonics of 120 Hz. This could be
detected 60 Hz. Probably from inside each mock box.
Bms 5 and 6 are the cleanest.
- 350 Hz. This is from the FAA radar.
- 5 Hz bw.
- bm 0 polA shows the crosshead frequency of 1.2 Hz.
These can be compared to the plots taken with the bias monitoring on
Total power spectra all
monitoring on (24may12)
- The 20 Hz glitches from the alfa dewar monitoring is coming
from the bias monitoring. Read the other values (temp,power
supplies, etc) do not contribute to these glitches.
- bm0 a sees the crosshead frequency.
- 60 Hz and harmonics of 120 Hz are seen. they are probably in
the mock boxes. bms 5 and 6 are the cleanest.