Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.naic.edu/~phil/lbw/scansep10/scanlbw_sep10.html
Дата изменения: Thu Sep 30 19:30:54 2010
Дата индексирования: Fri Feb 11 14:31:14 2011
Кодировка:
Поисковые слова: р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п

lbw

Checking lbw stability, rfi sep10

30sep10

PolB instabilities were reported in lband wide (lbw) in jul10. In sep10 while warming the dewar, the window blew out. The receiver was brought down to the lab and the window was replaced. The receiver was reinstalled 25sep10 and cooled down.
Tests were done to check the stability of the receiver (and also to look at the rfi):

27sep10: 300 second test covering all of lband using the mock spectrometer
29sep10: 4200 second test using the interim correlator and the setup of a2335 (who originally reported the problem).

27sep10: 300 second integration over entire band:

A 300 second test was done with the receiver on 27sep10. The setup was:

The mock spectrometer was used to cover the entire band.
4 bands of 172 MHz bandwidth and 8192 channels (20KHz resolution) were used.

The cfr for the bands were: 1200,1350,1500,and 1650 MHz.

3000 .1 second integrations were taken.
The telescope sat at az=285, za=9 degrees.
The data was taken around 17:00 ast. There was some lightening in the distance.

Processing the data:

dynamic spectra were made for each band (polB)
the average spectra was computed for each band.
the rms/mean by channel was computed for each band.
An rfi mask was generated by finding all rms/mean channels larger than 3 sigma
The rfi mask was used to compute the total power for each band.

Data plots,images:

PolB images of the 3 bands

each spectra has been normalized to the median spectrum to flatten the image.
Since the telescope was stationary, continuum sources drift through the beam.
The vertical scale is seconds (with .1 sec resolution)
I've labeled some of the more obvious rfi.

1200 MHz dynamic spectra (.gif)

the 4 radars 1232-1256 are from punta salinas. You can see their blanking when they point at us.
1227 is gps l2 (with low level spread).
1248 : low level rfi is glonass.

1350 MHz dynamic spectra (.gif)

1330,1350 radar is faa.

1370 radar is 1330 mixer image.

1290 radar: punta borinquen

1410 is mixer image.

narrow in time, horizontal lines (28,90,..) are lightening

1500 MHz dynamic spectra (.gif)

1525-1550 inmarsat sat to ground
1575 and associated low level rfi.. gps L1

1650 MHz dynamic spectra (.gif)

1575 gps L1
1598-1608 - glonass satellites
1618-1628 - iridium

5 minute average spectra and rms/mean (.ps) (.pdf)

top: 5 minute average
bottom: rms/mean by channel

The green line is the expected rms/mean from the radiometer equation.

The rms/mean is greater than the expected value because of the continuum sources.

The dc channel birdie is not stable (since it's rms is larger).
Some of the rms/mean rfi is polarized.. 1302,1303,

total power vs time (.ps) (.pdf)

the rfi was excluded from the total power accumulation.
top: total power vs time polA, polB
bottom: TotalPwrA/totalPwrB vs time
You can see the compression every 12 seconds from the radars.
There is no large drifting as seen back in jul10.

processing: x101/100927/chklbw.pro

29sep10: 4200 second test around 1385 MHz.

On 29sep10 a 4200 second test was done. The setup was:

lband wide, linear mode, 1280 hipass filter and 1230-1470 filter in place
telescope sitting at 272 and 8 degrees za
interim correlator 9 level sampling, 1 second spectra, 25 Mhz bw, 1024, centered at 1385.
take data for 4200 seconds.

Processing the data:

rms/mean computed over the 4200 seconds and then a mask was created using a linear fit to rms/mean.
total power was computed using the above mask.
the spectrum of the total power was also computed.

The plots show the total power and its spectrum (.ps) (.pdf):

page 1 top: total power vs RA.

black is polb, red is polb.
the horizontal scale shows how the sky drifted by for the 4200 seconds.

page 1 bottom: ratio pola/polb vs time

horizontal scale is not in secondsl
You see variations in the ratio at the beginning.
the variations are probably coming from the sky.

The telescope was pointing close to the north polar spur which is polarized.

Page 2 top: spectrum of total power.

black is polA, red is polB. polB has been offset for display.
The max freq is .5 hz (from the 1 second sampling).
polA seems to have more variation in the .1 to .3 hz region than polb.
the green line at .2 Hz would be the alias of the 1.2 hz crosshead frequency. nothing is seen.

Page 2 bottom: blowup of horizontal scale.

On this scale both pols look the same.

Summary:

no large drifts or jumps are seen.
polA looks looks a little noiser in the .1 to .3 hz range.
there is no power variation at the crosshead frequency

processing: x101/100929/chklbw.pro

home_~phil