[evlatests] Strange differential bandpasses

[Ken Sowinski]

On Fri, 27 Aug 2021, Rick Perley via evlatests wrote:

> Drew, et al.:
>
> Perfect!  Period shown in that C-band plot is about 100 MHz, which 
> corresponds pretty closely to the expected value corresponding to the 
> length of the horn.  Good evidence for a mismatch between the window and 
> the OMT….

The change in the sinusoiid should correlate with temperature.
Might the reflections have been exacerbated by moisture from
all the recent rain?




> I really doubt that the 4-band dipoles are in any way involved here. 
> Period is all wrong.
>
> Rick
>
> Sent from my iPad
>
>> On Aug 27, 2021, at 8:49 AM, Drew Medlin <dmedlin at aoc.nrao.edu> wrote:
>>
>> 
>> Hi all,
>>
>> Definitely seen in other bands beyond  L and S. I would say the magnitude of this is variable: sometimes L will show more than S band.  I've attached a plot from the pipeline here showing calibrated amp. vs. freq. for a C band observation taken this month here, and wlll try to find  some higher frequency examples and send them if I can find them.
>>
>> 20B-057.sb40024247.eb40024991.59433.01627496528.ms-J1254+1141-bb15-PHASE-amp_vs_freq-LL_RR.png = C band, 8bit ("PHASE" in png name is because it's the phase calibrator)
>>
>> <20B-057.sb40024247.eb40024991.59433.01627496528.ms-J1254+1141-bb15-PHASE-amp_vs_freq-LL_RR.png>
>>
>>
>>
>>
>> Cheers,
>> Drew
>>
>>> On Aug 27, 2021, at 6:56 AM, Rick Perley <rperley at nrao.edu> wrote:
>>>
>>> Drew, et al.:
>>>
>>> It has been suggested that the resonance is due to reflections associated with the length of the S-band horn.
>>> Similar effects are seen in L-band data, I believe.  I can’t fill those data into AIPS (which can’t handle mixed-mode SBs).  It would be useful to know if the resonances seen at L-band have a frequency period associated with the length of the L-band horn — it should be significantly less than that seen at S-band.
>>> Also, it would be useful to know if this effect has been seen at any band other than S and L.
>>>
>>> Rick
>>>
>>> Sent from my iPad
>>>
>>>>> On Aug 26, 2021, at 3:48 PM, Drew Medlin <dmedlin at aoc.nrao.edu> wrote:
>>>>>
>>>> Thanks for processing this observation, Rick. This is something we've been seeing for years in the pipeline results, but have not had a full answer about the cause.
>>>>
>>>>
>>>> Cheers,
>>>> Drew
>>>>
>>>> Drew Medlin
>>>> dmedlin at aoc.nrao.edu
>>>>
>>>> Scientific Data Analyst - VLA Scientific User Support Group
>>>> National Radio Astronomy Observatory
>>>>
>>>>> On Aug 26, 2021, at 3:41 PM, Rick Perley via evlatests <evlatests at listmgr.nrao.edu> wrote:
>>>>>
>>>>> One of our DAs, Edward Starr, has shown me some strange bandpass effects, whereby properly calibrated data, after application of calibration, shown undulating sinusoids with frequency.  There appear to be (at least) four separate observations affected by this, all taken in July or August.
>>>>>
>>>>> Only one of these is accessible to AIPS at the present time.  I loaded this one (21A-033, S-band, observed 23 July 2021) into AIPS to investigate.
>>>>>
>>>>> What I found is most peculiar:  The bandpasses of nearly all antennas systematically change over the course of the observation (13:30 to 17:35 IAT).  The effect is very easily seen if one uses the initial scan of 3C48 to calibrate the bandpasses, then generate 'differential' bandpass solutions on the subsequent data (following normal calibration).  By the end of the run, the observation of 3C84 shows bandpass ripples up to 2.5% in amplitude, and 2 degrees in phase.
>>>>>
>>>>> The character of the change in bandpass is exactly as Edward originally described -- a nearly perfect sinusoid, which steadily grows in amplitude over time.  The period is very close to 43 MHz, which corresponds to 3.5 meters in free space, if due to a standing wave.  The characteristics shown in the plot make it virtually certain that this is a beat phenomenon due to a reflected signal.
>>>>>
>>>>> I have attached an example -- the most spectacular antenna is ea05.  This is the differential bandpass, from 3C84, observed at the end of the run, using the bandpass from 3C48, taken at the beginning.
>>>>>
>>>>> All antennas look like this, but none have amplitudes as large as this one.  The 'beat' pattern is the same in all antennas on which the effect can be discerned.  Other antennas which have the largest effect are ea08, 12, 20, and 26.  There is no spatial relationship of these -- they are evenly spread about the array. This is not an elevation effect, as the growth of the pattern in the phase calibrator (J0204+1514) neatly fits the final observation of 3C84.  The last observation of J0204+1514 was taken at an elevation of 32, the subsequent observation of 3C84 was taken at elevation 54.  The original calibration, on 3C48, was at elevation 79.
>>>>>
>>>>> The two polarizations give identical results -- both in amplitude and phase, including the frequency location of the peaks and troughs.
>>>>>
>>>>> I've seen sinusoids before, but nothing quite like this case.
>>>>>
>>>>> Ideas?
>>>>>
>>>>>
>>>>>
>>>>> <S-ea05.png>_______________________________________________
>>>>> evlatests mailing list
>>>>> evlatests at listmgr.nrao.edu
>>>>> https://listmgr.nrao.edu/mailman/listinfo/evlatests
>>>>
>>
>