Alfa compression by radars using pdev

23mar08

Links:
Data processing
Table of the data sets
06mar08: sitting az=340, za=18
11mar08 sitting at az 340, za=18
12mar08 sitting at az 340, za=17
13mar08 zastrips az=180, za=2->19.5->2. Radars: remy ,aerostat, no faa.
18mar08 sitting at az=340,za=18. rotate alfa.
18mar08 za strips at az=360 and 180. za 19.5->2->19.5. faa,remy, no Aerostat
Table of za locations for jumps when at the meridian.
Conclusions

INTRO: (top)

Jumps in Alfa's power levels with long recovery times (greater than seconds) has been seen for some time (see alfa pwr jumps). During the painting project alfa was brought down to the lab to be worked on. After reinstalling alfa in the dome, test were done with the pdev spectrometer to measure the power jumps. The setup was:

Send the 300 MHz of alfa into 2 170 MHz bands of the pdev spectrometer. The first IF was centered on 1375 MHz (first lo= 1375+250=1625.). One of the 170 MHz bands was centered at 1300 MHz and while the other was centered at 1450 MHz.
The pdev spectrometer was typically configured as:

512 or 128 Length ffts, 170MHz Bw, 32 Bit output data, pshift=0x177. ASHIFT, DSHIFT was setup so there was not saturation.
Data was sampled at 200 usec rate for many 12 sec cycles.

The telescope was parked at a particular position (say 340,18 were faa is strongest) or it was driven at a constant rate (either in za or az).

Processing the data:

The idl routines masrdrsat and masrdrsatplt were used to process and plot the data.
For each scan the total power for the 1450 MHz band was computed and then normalized to the median value.
The peak spectral value for the 3 radars was recorded each 200 usecs. The search was done for 1 MHz or 3 channels (whichever was longer) about the radar frequencies. PolA and PolB were searched simultaneously. The frequencies used were:

FAA:1330,1350 MHz
REMY:1270,1290 MHz
AeroStat: 1241.74,1256.5,1246.2,1261.25 MHz

Often the aerostat was not running but we still got a signal since the punta Salinas radar also uses some of these frequencies. You can tell the punta Salinas radar by the longer blanking (3versus 1.4 secs) and the extra short blanking for the san juan airport radar (5.5 seconds after the end of the AO blanking).

Data covering the entire time was plotted:

Each beam is plotted in a different color.

The 1450 (clean) band was smoothed to 20 milliseconds. Offsets were included for plotting.
For each radar the peak power over 20 milliseconds was plotted in db above Tsys.

Higher resolution about any compression was also plotted:

The 1450 band usually showed a jump in some beams as well as positive going spikes that are intemods/harmonics of the faa radar.
The radar peak powers are also shown during the compression event.

Dynamic range.

The maximum dynamic range for the radar peaks varied from 35 to 45 db above Tsys. This is the ratio of the Peak radar power in 1 channel vs the average power in a channel. To get the increase in total power divide this increase by the number of channels (usually 128 for 1 single 170 MHz band). This doesn't take into account multiple radar frequencies (eg 1330,1350) or that there are 2 170 MHz bands.
For early datasets, the compression is probably occurring in the pdev mixer chassis. This depended on the attenuation and input level used to the pdev device.

The datasets: (top)

Terminology:

PS punta Salinas (at the aerostat frequencies)
AERO Aerostat
FAA faa airport radar
REMY Remy radar

Data Sets with links to daily descrption and plots
date	fileNumber Link to plots	nchan	time secs	Jumps beams	radars present	az/za	alfa angle	Notes
06mar08	000	512	24	-	faa,ps	285/10		pshift 0x155,dshift=0, accum overflow
	100			3,4	faa,ps,remy	340/18		overflow
	200			0,2,3,4	faa,ps,remy			overflow
	300			2,3,4	faa,ps,remy			dshift=5 no overflow
	400			0,2,3,4,5	faa,ps,remy			dshift=5
11mar08	000	512	24	3	faa,ps,remy	340/18	19	pshift:0x1ff small jumps.
	100			3	faa,ps,remy		19
	200			3	faa,ps,remy		0
12mar08	000	512	24	-	faa,ps,remy,aero	340/17		This was not at 340,18, not jumps seen
12mar08	100	512	24	-	faa,ps,remy	340/17		This was not at 340,18, not jumps seen
13mar08	zastrip.0..2	128	875	-	aero,remy	180/(2..19.5)	19	dshift=0 some accum overflows. 02deg/sec
13mar08	zastrip.100..102	128	875	-	aero,remy	180/(19.5..2)	19	dshift=1, no overflows. -.02 deg/sec
18mar08 rotate alfa	sit.000	128	24	3,4	faa,ps,remy	340/18	19	see how saturation, radar strength changes with alfa rotation angle. filenames: sit340_18.20080318...
	sit.100			0,1,2,3,4	faa,ps,remy		0
	sit.200			1,3,4,5	faa,ps,remy		40
	sit.300			0,1,3,4,5	faa,ps,remy		60
18mar08 zaStrips	zaN.000..002	128	875		faa,ps,remy	360/(19.5..2)	19	za strips move at .02 degs/sec. Filenames: zaswingN, zaswingS
	zaN.100.102				faa,ps,remy	360/(2..19.5)
	zaS.000.002				faa,ps,remy	180/(19.5..2)
	zaS.100.102				faa,ps,remy	180/(2..19.5)

06mar08 sitting az=340, za=18 (top)

5 scans were taken on 06mar08. There was overflow in the pdev device for the first 3 until dshift was set to 5.

File:00000. az=285,za=10 (.pdf) there is no jump in the power at this az,za.
File:00100. az=340, za=18 (.pdf) there are two jumps spaced by 12 seconds (.2 and 12.2 seconds)
File:00200. az=340, za=18 (.pdf) there are two jumps spaced by 12 seconds (7 and 19 seconds)
File:00300. az=340, za=18 (.pdf) there are two jumps spaced by 12 seconds (0 and 12 seconds)
File:00400. az=340, za=18 (.pdf) there are two jumps spaced by 12 seconds (2 and 14 seconds)

Comments:

Alfa still has a problem with a power jump and long recovery time when the faa radar points at AO.
There was no tsys increase and then long recovery for az=285,za=10. It was present for all of the az=340,za=18 positions.
At az=340,za=18 beam 3 always had the jump and slow recovery. Beams 2 and 4 sometimes had it. This is probably a geometry problem. Moving in az,za would probably cause another beam to have the problem.
The spacing of the 2 millisecond dips lined up with the faa ipps.
The 1330,1350 radars are causing intermods/harmonics in the 1450 MHz band at 135.02,1474.94, and 1480.33

processing: 080306/rdrsat.pro

11mar08 sitting at az 340, za=18. (top)

Fixed the saturation levels in pdev. Took two scans at alfa angle = 19 degrees and then 1 scan at alfa angle=0.

file:00000. az=340,za=18 alfa angle=19
file:00100. az=340,za=18 alfa angle=19. repeat with same setup
file:00200. az=340,za=18 alfa angle=0

processing: 080311/rdrsat.pro

12mar08 sitting at az 340, za=17 (top)

Sat at 340,17 rather than 340,18 and saw no jumps. Didn't record the alfa angle.

file:00000. az=340,za=17
file:00100. az=340,za=17

processing: 080312/rdrsat.pro

13mar08 zastrips az=180, za=2->19.5->2. Radars: remy ,aerostat, no faa. (top)

Two za strips were done at az=180. 2->19.5 degrees followed by 19.5->2 degrees. The dome moved at .02 degrees/second. The 1450 band data was smoothed and then decimated to .2 seconds. The radar bands had peak holds done over the same time step. The radars rotate at 12 seconds. In this time the za moved by .02*12 = .24 degrees. The repeatability will be limited by this value ( the za position of compression going up will be a little different than the za position of the compression going down.
During the data taking the Remy and aerostat radars were running. The faa radar was off. No compression was seen in the
entire dataset.
The plots are:

file.00000. az=180 za=2->19.5 degrees.
file.00100. az=180 za=19.5 -> 2 degrees.

Daily summary:

The remy radar and the aerostat do not cause the jumps with long recovery times in these za strips.
The aerostat signal is strongest at higher za . Peaks occur near:

za=17.5, 18.1 for az=180 degrees.

processing: x101/080313/rdrsatza.pro

18mar08 sitting at az=340,za=18. rotate alfa. (top)

Data was taken at 4 settings of the alfa rotation angle: 19,0,40 and 60 degrees to see how the jumps moved from beam to beam. When rotating from 0 to 60 degrees, beams N -> N-1 so you might expect to see the jumps move from beam to beam. A 60 degree rotation moves 1 beam to the next, but the waveguide probes are not at the same angle as the previous beam.

rotAngle 0 .. jumps beams 0,1,2,3,4.. 2 and 4 the strongest
rotAngle 19 ..jumps in beams 3,4
rotAngle 40 .. jumps in beams 1,3,4,5
rotAngle 60 .. jumps in beams 0,1 3,4 5 3 and 5 the strongest

The the rotation from 0 to 60 degrees mapped the strongest compressions 2,4 -> 3, 5 which was expected.
processing: x101/080318/rdrsatrot.pro

18mar08 za strips at az=360 and 180. za 19.5->2->19.5. faa,remy, no Aerostat (top)

Four: za strips were done:

az=360. za 19.5->2->19.5
az=180. za 19.5->2->19.5

    The dome moved at .02 degrees/second. The 1450 band data was smoothed and then decimated to .2 seconds. The radar bands had peak holds done over the same time step. The radars rotate at 12 seconds. In this time the za moved by .02*12 = .24 degrees. The repeatability will be limited by this value ( the za position of compression going up will be a little different than the za position of the compression going down.
    Jumps in the total power (with long recovery times > secs) were seen. The data is plotted vs za. When moving down in za, the jumps are from the right to left in the plots.When moving up the jumps go from small to larger za.
    During the data taking the Remy , punta Salinas, and FAA radar were running. The Aerostat radar was off. Peaks in the aerostat plots are from the punta Salinas radar (which shares the same frequencies).
The plots are:

rdrsatzaN.00000. az=360 za=19.5->2 degrees.
rdrsatzaN.00100. az=360 za=2 -> 19.5 degrees.
rdrsatzaS.00000. az=180 za=19.5->2 degrees.
rdrsatzaS.00100. az=180 za=2 -> 19.5 degrees.

Compression was seen in the up down directions when the azimuth was at 360. It was not seen when the za swings were at an azimuth of 180. All of the jumps lined up with just the FAA radar.

Za locations for saturation during za strips
za Deg	az Deg	direction	Beams with Saturation	maximum strength
3.0	360	up	5	<1%
5.45	360	down	3,4	1%
5.55	360	up	3,4,5	1%
6.15	360	down	4	<1%
6.29	360	up	3,4	2%
6.4	360	down	3,4,5	2%
13.87	360	down	3,4	1%
13.97	360	up	3.45	2%
17.23	360	down	3,4	4%
17.35	360	up	2,3,4,5	1%

Daily Summary:

Jumps with long compression were see during za strips (2->19.5) when the azimuth was positioned at 360 Degrees.
No jumps were seen when the azimuth was at 180 degrees.
Jumps were repeatable when moving up and moving down (within the .24 degrees of the 12 second radar rotation). This says that the problem is not being caused only be a glint from a plane. It is there all the time (as long at the radar points at us).
The jumps are being caused by the FAA radar. They are not being caused by the remy or punta Salinas radar.
When the jumps occur, the faa peak power is not a narrow function of za when moving at .02 deg/sec.
The Aerostat radar was not running during this test.

processing: x101/080318/rdrsatza.pro

Conclusions: (top)

The faa radar causes the power jumps with long recovery times (> seconds).
The jumps are not seen from the punta salinas, remy, or aerostat.

We may need more sampling of the aerostat at different az's.

All of the radars produce compression with fast recovery.
Along the meridian the jumps only occur when az=360. az=180 does not show any jumps. (a table of the jump za's).
Maximum jumps are about 4% of Tsys.
The jumps are seen close to the same za when moving up in za and when moving down (separated by 15 minutes).

The power is stationary in space and time. It is not only coming from a glint off of a passing airplane.
When moving in za the jumps can last for > 1 rotation of the radar. At .02 deg/sec this is > .24 seconds.

When sitting at a fixed az,za and rotating the alfa turret by 60 degrees, the jumps move from 1 beam to the next.
When using short enough

page up
home_~phil