16apr06

# 16apr06 Initial report of comb.

plots:
750 Khz comb (.gif)
an average spectra of 40 seconds (.ps)  (.pdf)
39 Khz comb (.gif):
an average spectra of 40 seconds (.ps)  (.pdf) :

Two combs were seen in the 327 receiver in data taken on 13apr06. One comb is at 750 Khz while the second comb is at 39.45 Khz

#### The 750 Khz comb:

The first image shows the 750 Khz comb (.gif):
• The dynamic spectra have 2048 channels over 25 Mhz with hanning smoothing. The channel width is 24 Khz. The black vertical bars at the bottom are drawn at the comb frequencies. The comb lies  at 314.25 + n*.75 Mhz.
The strength of the birdie is seen in an average spectra of 40 seconds (.ps)  (.pdf) :
• The top two plots show 314 to 327Mhz and then 327 to 340 Mhz.
• The bottom two plots blow up the vertical scale. The red dotted lines are placed every  375 Khz. The 750 comb (2*375) is a lot stronger, but you can still see birdies every 375 Khz. So the comb is probably a square wave at 375 Khz and the 750 Khz are just the odd harmonics that are strongest.

#### The 39.45 Khz comb

The image shows the 39 Khz comb (.gif):
• The dynamic spectra have 2048 channels over 390 Khz with hanning smoothing. The channel width is 380 hz. The black vertical lines at the bottom are drawn at the comb frequencies. The comb lies at:  335.797 + n*.03945 Mhz.
The strength of the birdie is seen in an average spectra of 40 seconds (.ps)  (.pdf) :
• The top plot shows the 390 khz band.  The amplitude of the comb is about equal to Tsys in a 380 Hz channel (tsys at 327 is about 120 K). The red dots are at the comb frequencies of:
• 335.797+ .03945*n
• The second plot is a blowup in amplitude. There does not appear to be any birdies at twice the comb frequency.
• The third plot is a blowup in frequency. The comb is taking a little more that 3  channels (hanning smoothing should spread it over two channels). So we may be starting to resolve the birdie with a 380 Hz channel.

#### When did the combs appear?:

The 750 Khz comb:
• 05apr06 19:22 x111 data . the comb was not present
• 06apr06 18:51 p1639 pulsar data. The comb was not present.
• 11apr06: 7:53   x111 data. The comb was present.
• 15apr06: 16:00 p1693 data. The comb was present.

The 39 Khz comb:
It was seen on 13apr06 in the high resolution data. The x111 data does not have high enough resolution to see this comb.

#### Summary:

• A 750 Khz comb was first seen in the 327 receiver on 11apr06. It was not present on 06apr06.
• The comb gets up to about 1.5 Tsys in a 24 Khz channel.
• The comb is located at 314.25 + n*.75 Mhz.
• The 750 Khz is probably the odd harmonics of a square wave .The fundamental is probably .375 Khz.
• Tsys for the 327 receiver jumped up on 11apr06. It was not high at 8:30 am. It was high at  16:30. This does not quite match the 750 Khz comb since the 750 Khz comb first appeared at 7:56 am on 11apr06.
• The 750 Khz comb was also seen in the 430 dome receiver on 11apr06.
• A 39.45 Khz comb was first seen on 11apr06. Previous data taken mar,apr06 did not have high enough resolution to see this comb.
• The 39.45 Khz comb has an amplitude about equal to Tsys in a 380 hz channel.
• The comb might be starting to be resolved in a 380 hz channel.
processing:x101/060413/327_combs.pro

# 18apr06 comb is coming from the sband console in dome  (top)

Data was taken on 18apr06 to see if the 750 kHz and 39 kHz comb were coming from the sband console in the dome. We used the 327 receiver. The setup was:
• center a 25 Mhz band (2048 channels) at 327 Mhz. Center a second band with 390 Khz bandwidth (2048 channels) at 336.92 Mhz.
• Take 500 1 second records. During this time open, close the doors to the sband cabinet. Also turn off the power to the cabinet.
The first image shows the dynamic spectra for the 25 Mhz band (.gif) :
• Horizontal  dashed lines mark the approximate time when things were changed:
• 0-70 seconds : cabinet is on, doors are open.
• 70-170: cabinet is off, doors are open.
• 170-245: cabinet is on, doors are open.
• 245-310: cabinet is on, doors are closed.
• 310-397: cabinet is off, doors are closed.
• 397-467: cabinet is on, doors are closed.
• 467-500: cabinet is off, doors are closed.
• You can see a 750 Khz comb as well as a 375 Khz comb (that is weaker).
• Closing the doors makes the comb smaller but it does not go away
The second image has the dynamic spectra for the 390 Khz band (.gif):
• The horizontal dashed lines are the same as the 25 Mhz band.
• There is a 40 Khz comb when the cabinet is on.
• Each 40 Khz comb element is a triple spaced by 2.8 Khz.
• The 40 kHz comb is drifting some in freq. (probably caused by our turning the power on and off).
Average spectra were made during the door open, door closed, and power off times: The plots show the strength of the combs (.ps) (.pdf):
• Page 1: 315 to 327 Mhz (25 Mhz band). Top is door open, middle is door closed, bottom is cabinet off. Red dashed lines are spaced every 375 Khz. The strongest birdie is at 321.5 Mhz. It is 120*Tsys with the door open and 30*Tsys with the door closed, These spectra have  a 24 Khz channel width (after hanning smoothing).
• Page 2: 327 to 340 Mhz (25 Mhz band). The strongest birdie is at 330 Mhz: 55*Tsys with door open,  6*Tsys with the door closed. The birdie at 336 Mhz is not coming from the sband cabinet since it is still present when the power is turned off.
• Page 3: 390 Khz band centered at 335.92 Mhz. Dotted green lines are plotted at the 40 Khz comb. There is a triple at each comb element spaced by 2.8 Khz. The comb frequency is about 40.55 Khz but is changed a little with time. The 2.8 Khz spacing is also approximate. The strongest comb element (in this band)  is at 335.9164 Mhz. It is 2100*Tsys with the door open, 60*Tsys with the door closed. These spectra have a 380 Hz channel width.
After inspecting the inside of the sband cabinet, we found that some of the internal metal covers had not been screwed into place (there were lots of screws).  After securing these covers we took some more data.
The dynamic spectra show the 25 Mhz band after the internal covers were secured (.gif):
• The horizontal lines show when things occurred:
• 0-77: cabinet on, doors closed
• 77-166: cabinet off, doors closed.
• 166 : turned off portable spectrum analyzer (sbt250) they were using in the dome.
• 184: cabinet back on.
• You don't see any change in the spectra at 77 seconds when the cabinet was turned off or at 184 seconds when the cabinet was turned back on.
• You do see a change in the spectra when the portable spectrum analyzer they were using in the dome was shut off at 164 seconds. It radiates enough to get into the 327 system. Some of the frequencies were:
• 315.8 Mhz
• 318.5 Mhz

#### Conclusion:

• The sband console makes a comb at:
• 375 Khz with the 750 Khz odd harmonics the strongest.
• 40.55 Khz (this could drift down to 39.45 Khz measured on 16apr06). This comb has a triplet at each comb element spaced by 2.8 Khz.
• With the cabinet doors closed but the internal covers not secured with their screws, the birdies can get up to:
• 750 Khz comb: at 321.5 Mhz it is 30*Tsys in a 24 Khz channel. This is the peak in a 25 Mhz band centered at 327 Mhz.
• 40 Khz comb: at 335.9164 Mhz it is 60*Tsys in a 380 Hz channel. This is the peak in a 390 Khz band centered at 335.92 Mhz (it is not the strongest peak in the 25 Mhz band).
• Correctly securing the internal covers in the sband cabinet with their screws makes the combs go away in a 30 second average with a 24 Khz channel width.
• The portable spectrum analyzer used in the sband room in the dome causes rfi that can be seen in the 327 Mhz receiver.
processing: x101/060418/chksbcab.pro

<-- page up
home_~phil