18sep08: bias for 3a/3b changed. Look at p2030 data
21sep08: bias for all beams changed. Look at a2010
data (1 second integrations).
Large pulses (radars) were causing long recovery times in
alfa. Adjusting the bias voltage and
caused alfa to recover much quicker. The
decision was made to change the bias V,I of alfa to these new
values. Eventually we hope to install leds on the stage 3 amps to fix
the problem (this will take a few months of down time).
The bias voltages and currents modifications
- 18sep08: daytime.. beam 3a,3b Vd went 5.0 V -> 5.92V, I
went 19mA -> 17mA.
- 19sep08: daytime.. all other beams were adjusted to Vd 5.92 and I
18sep08: p2030 data after beam 3a,3b bias
The p2030 data taken 18sep08 after the 3a,3b
modifications was analyzed to check the stability of the system. The
- The wapps in 3 level mode take 100 MHz bw with 64usec sampling
and 256 lags.
- alfa is centred at 1440 Mhz with the 100 Mhz 1390-1490 filter int.
- The total power was computed and then averaged to .1 seconds.
P2030 sums polarizations so only PolA+PolB was available.
- Data was plotted for each scan (usually 270 seconds).
The plots show total power vs time for the 18sep08
p2030 run (.ps
- The plots are total power (polA+polB) vs time in units of Tsys.
- Each beam is color coded. Beam 3 is dark blue
- Beam 3 has a drift of 1.5% in 270 seconds in the first
scan. Beam 4 has a smaller drift. All other beams are relatively stable.
- Scans 2-16 all show beam 3 as stable as the other beams that were
- beam 6 has some jumps in it. Beam 6b is dead and p2030 is adding
polA + polB. So maybe polB is causing this problem. Maybe we should go
back to ganesh's idea of inserting stable noise rather than
the output of the amp??
- Changing the bias voltage on beam 3a,3b did not affect the
stability at the 100Mhz*.1 second level.
- beam 6 with polA,B summed is having trouble.
21sep08: a2010 alfalfa data after all beams
had their bias voltage modified. (top)
The a2010 data taken on 21sep08 was analyzed to
check the stability of all the beams after the bias changes in alfa.
The processing was:
- a2010 takes 100 Mhz of data with the wapps centered near 1385 Mhz
with 1 second integrations.
- The total power for each 10 minute scan was computed. Frequencies
with rfi were not included in the computation of the total power.
The plots show the total
power vs time for the 11 scans (.ps)
The total power vs time looks similar to the monitor data
has been published each month for a2010. There are no jumps or slow
drifts (just sources).