[evlatests] Intriguing pulsed emission in S-band

[Dan Mertely]

Hi Rick.  It sounds like the ABQ NexRad radar, but I'll
have to dig into my records and find the exact frequencies.
The 4.5 sec period is likely the rotational frequency of
the radar, not the actual pulse repetition period (which
would be much shorter).  Within each burst you saw should
be a multitude of short pulses (in the uS range) separated
by 20 mS or so.  Since they electronically scan in EL while
rotating in AZ, the strength of the signal will seem to
vary from rotation to rotation, depending on where they
catch us in the sweep.  -Mert

Rick Perley wrote:
>      Trolling through a number of S-band datasets have revealed 
> worrisome issues with the switched power (although the visibilities in 
> nearly all cases look fine).  We suspect that the strong RFI found in 
> this band is somehow to blame, so Ken ran a short test yesterday to help 
> diagnosis. 
>     We didn't find any of the switched power troubles we were looking 
> for.  But this relative small database has allowed a better look into an 
> issue which I've seen in all S-band data taken so far. 
> 
>     In subband 6 (frequencies 2.628 through 2.756 GHz), both the sum and 
> difference powers from the switched power system show a regular pulse.  
> The effect is about 10% of the total power in that subband.  More 
> careful review reveals the following:
> 
>     1) It is a *broadband* phenomenon.  There is no sign of any extra 
> narrow-band power in the spectral plots for this subband.  (resolution = 
> 2 MHz).   There is no sign of this pulsation in subband#5, but there is 
> a weak detection of it in subband #7. 
>     2) Although it is seen on (almost) all antennas, and on both 
> polarizations, there is a clear spatial preference for antennas near the 
> center and down the east arm.  The pulsation is nearly undetectable on 
> the antennas at/near the ends of the north and west arms.
>     3) The pulse period is 4.6 seconds.  The pulse is (almost) never 
> seen in two adjacent time records (time resolution is 1.0 seconds). 
>     4) The pulse has a width comparable to our PDif switching period 
> (100 msec, I think).  This is seen in the PDif values, with have a 
> smoothly sinusoidal variation (from positive to negative), with a period 
> of about 45 seconds.  (A little more thought, using the two 
> periodicities listed above, should allow specification of the pulse 
> duration). 
> 
>     I cannot see this pulsation in the visibilities -- presumably due to 
> the wideband nature, and the (presumably) large delay.  I haven't yet 
> found it in the autocorrelation spectra, but am going to look more 
> carefully. 
> 
>     Any ideas as to the origin?
> 
>    
> _______________________________________________
> evlatests mailing list
> evlatests at listmgr.cv.nrao.edu
> http://listmgr.cv.nrao.edu/mailman/listinfo/evlatests