Gps Rfi aug08

04sep09

Project a2010 (alfalfa) saw some interference near 1346.7 MHz while taking data during aug08. There setup was:

The alfa receiver was used centered at 1385 MHz without the 100 MHz rf filter installed.
Data was taken with the wapps using 100 MHz bandwidth and 1 second dumps.
They did drift scans each nite.
Data was taken in the early morning between 0 and 5 am ast
The rfi was seen at 1346.7 on: 22-25aug08, 27-29aug08. It was not seen on 22aug08 (i didn't look at data prior to 20aug08).
To characterize the rfi, the peak value +/- 1MHz about 1346.7 was taken as the location of the rfi in time..

The table below shows when the rfi was seen:

DATE	Rfi present	az/za	File number/recNum with rfi	AST secMidnite beam 2 peak
22aug08	yes	359.6/12.68	.00010/470	8272
23aug08	yes	359.6/12.92	.00009/270	7999
24aug08	yes	359.6/13.53	.00011/515
25aug08	yes	359.6/13.77	.00011/234
26aug08	not seen	-	-
27aug08	yes	359.6/14.02	.00009/179
28aug08	yes	359.6/14.26	.00010/407
29aug08	yes		.00007/245

The first plots shows dynamic spectra for 22aug08 file 10 polA (.gif) (beams 0:3) (beams 4:6):

Beams 0-3 show the rfi at 1346.7 near record 470. There are other narrow lines that also occur at this time.
Beams 4-6 have rfi near 1342 and 1352..

The 2nd plots show 20 1 second spectra for 22aug08 near the peak for beam2 (.ps) (.pdf):

each plot is a different beam (bm0-3 on page 1, bm4-6 on page 2)
The red line is near the peak for the 1346.7 rfi for beam 2.
Beam 3 also has the rfi but occurring about 12 second earlier.
The narrow spikes in beams 4-6 occur during the 1346.7 rfi in beams 0-3.

The 3rd plot shows when the 1346.7 peaks occurred for each day (.ps) (.pdf):

Page 1: This is the ra/dec offsets relative to beam 0 for the alfa beams when the az=360 and the rotation angle is 19.1. Something drifting into beam2 would appear in beam 0 about 20 seconds later.
Pages 2-3 (polA,polB): For each day the peak power near 1346.7 is plotted versus AST seconds from midnight (note that 26aug08 is missing since no rfi was seen).

273 seconds/day is added to each day after 22aug08. This did the best job of lining up the pulses.
In all cases, the rfi arrives in beam3 then beam2, then beam1. Looking at page 1, shows that the object is moving from low declination to high declination (rises from the earths equator).
A 24 hour sidereal rate would be 3min 56 seconds=236 seconds less that 24 hour solar time. The measured time is 37 seconds shorter than a 24 hour sidereal rate.
The za is moved by .25 degrees in za per day (to lower declination) this will cause the period to be less that 86400-236

1 orbit in 24 hours is .25 Amin/sec.. so the .25 degrees za would decrease the difference to 236-60=296 seconds.
An orbit in 12 hours is .5 Amin/sec .. so the .25 degrees za would decrease the difference by 236-30=266 seconds.This is close to the 273 sec less periodicity we measure (also remember that a satellite's orbit could be inclined to the equator).

Page 4: This has the ra/dec for the rfi peak for each day for beams 2 and 3. The stepping is declination is because of the za motion each day. The bottom plot shows the ast secsFromMidnite for each peak.

Conclusion:

Se need a satellite that rises from the equator and has a sidereal rate orbit with a 12 hour period. Turns out this is what the gps satellites have.
Using the gps orbit calculator at http://www.calsky.com./cs.cgi to see if any satellites were close to us..

date	time Ast	gps Satellite	srcAz/za	ra/dec	AO srcAz/za
22aug08	2:17:42	prn26/sv26 BIIA	180.5,12.7	23:54/5:35	180.4/12.68
23aug08	2:13:09	prn26/sv26 BIIA	180.5/13	23:54/5:21	180.4/12.92

So on these nites gps satellite 26 flew very close to our beam.
Gps will emit 1227.6, 1381.05, and 1575.42..I tried but did not find a simple way to mix any of these or there harmonics to 1346.7...So i don't know if the satellite had out of band emission or we created the frequencies in our system.

Processing: x101/080824/rfia2010.pro

page up
home ~phil