[evlatests] Bandpass Ripple in ea28, RCP, L-band

[Paul Demorest]

Drew thinks this ripple is still visible in recent (ie, past ~week) 
L-band pipeline results.  Not sure whether it's improved since Sep 4..

-Paul

On 2019-10-04 15:12, Dan Mertely via evlatests wrote:
> OK, it sounds like all of the data were taken
> before our latest attempt at mitigation.  It
> will be interesting to see if the ripple is
> less noticeable now.  -Mert
> 
> On 10/4/2019 3:04 PM, Rick Perley via evlatests wrote:
>> The C configuration data were taken 15 Dec, 2018.
>> 
>> The B configuration data were taken 18 Apr, 2019
>> 
>> The A configuration data were taken 16 Aug, 2019
>> 
>> Rick
>> 
>> On 10/4/19 3:01 PM, Dan Mertely via evlatests wrote:
>>> Hi Rick. Could you tell me the dates of the observations?
>>> We worked on the W8mon back on 4 September, but haven't
>>> heard whether our connector cleaning fixed the problem.
>>> -Mert
>>> 
>>> On 10/4/2019 10:28 AM, Rick Perley via evlatests wrote:
>>>>      I'd like to return to this topic, with some more information.
>>>> 
>>>>      There is no doubt that the W8 RFI monitor is the guilty party. 
>>>> The new information is this:
>>>> 
>>>>      The data came in three configurations:  For the first, ea24 was 
>>>> on W8.  For the second and third, ea28 was on W8.  In reviewing the 
>>>> calibration, I find:F
>>>> 
>>>>      a)The rapid oscillation noted before is on these antennas -- 
>>>> and only these antennas.
>>>> 
>>>>      b) The cross-polarization signals (RL and LR correlations) are 
>>>> *completely ruined* on all baselines connected to ea24 for the first 
>>>> configuration, and ea28 for the other two.
>>>> 
>>>>      The reason for this cross-polarization problem is that the 
>>>> fine-scale ripple is both extremely large, and extremely time 
>>>> variable. So there is no way we can realistically 'calibrate it 
>>>> out'.      No doubt the parallel-hand signal (for the polarization 
>>>> used for that monitor) is similarly affected by the time 
>>>> variability.
>>>> 
>>>>      The fact that this problem was easily visible on two different 
>>>> antennas (which occupied the same pad -- W8), on three observations 
>>>> spanning many months, strongly suggests that many other observations 
>>>> have been affected.
>>>> 
>>>>      Hence my questions:
>>>> 
>>>>      A) Has the current problem (with ea28) been repaired?
>>>> 
>>>>      B) Do we really need this monitor?  Unless the reflection 
>>>> associated with this monitor is reduced by at least 20dB, the 
>>>> observational data cannot be used.
>>>> 
>>>> 
>>>> On 9/3/19 2:42 PM, jjackson via evlatests wrote:
>>>>> Wes and I just had a chat about this and our chief suspect would be 
>>>>> the W8 RFI monitor cable that is connected to L-band on this 
>>>>> antenna. The length is just too long for anything else to make 
>>>>> sense on a single polarization.  30 meters would be about right for 
>>>>> the length of the cable running down the antenna.  We'll have 
>>>>> someone take a look at it.
>>>>> 
>>>>> Jim
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> On 9/3/2019 11:01 AM, Rick Perley via evlatests wrote:
>>>>>>      Bandpass solutions from my Cygnus A observations, taken in 
>>>>>> April and August this year show a strong bandpass ripple in RCP 
>>>>>> only, for antenna 28 and L-band.  The period of the ripple is 
>>>>>> about 4.6 MHz, corresponding to 30 meters if the cause is a 
>>>>>> reflection. The problem was not visible in data taken in December 
>>>>>> last year.
>>>>>> 
>>>>>>      The ripple pattern is strongly time dependent (both in its 
>>>>>> amplitude and phase), so is very difficult to calibrate out.