[evlatests] C-Band Sensitivity Troubles

[Rick Perley]

    I neglected to list the antennas whose sensitivities are the worst 
at C-band:

    They are (starting with the worst):

    RCP:  19, 11, 21, 24,26,1 (Tsys varying from 130 to 70 K)
    LCP:  19, 17, 21, 11, 1, 24, 23, 26 (same Tsys range). 

    I also neglected to note that the procedure described earlier gave 
sensible results only for the A and C IFs.  For B and D, some EVLA 
antennas gave very different Tsys values -- which are always lower.  In 
all cases, these are the most recently upgraded EVLA antennas.  As Cyg A 
effectively doubles the system temperature at this band, the simplest 
explanation is that we're running the power levels too high (or too 
low), giving a non-linear response when we move to Cyg A. 



Charles Kutz wrote:
> It was not an inch of water! Although there were a few drops on the 
> window inside the feed.
>
> In order to maintain some modicum of order and accuracy in information 
> flow, I would highly prefer that information flowed from the 
> appropriate channels.
>
> I was waiting on the full debrief from the FE crew this afternoon 
> prior to disseminating their findings.
>
> Chuck
>
>
>
>
> Dan Mertely wrote:
>> Hi Rick.  The FE guys just found an inch of water in the
>> K band feed on one of the bad antennas.  Can you tell us
>> which of the C band antennas showed the worst Tsys?  They
>> may be able to look down the feed to the window of that C
>> band & see if we have a similar situation.  (You mentioned
>> the best performers, but didn't say which one was worst.)
>>
>> -Mert
>>
>>
>>
>> Rick Perley wrote:
>>>     I used a few minutes of maintenance time this morning to roughly 
>>> calibrate the Tcals for the antennas at C-band, and hence get an 
>>> estimate of the system temperatures.  This was done by observing 
>>> Cygnus A ( a strong source of known flux density), noting the 
>>> reported rise in system temperature, and adjusting this by the 
>>> expected rise.
>>>     Cyg A provides about 495 Jy at 6cm.  If we make the assumption 
>>> that the efficiency of the antenna at this frequency is same for all 
>>> antennas, and is equal to 0.55, then the expected rise in Tsys will 
>>> be about 50K.  Although there will be some variation in antenna 
>>> efficiencies, these are most unlikely to be greater than a few 
>>> percent -- a far smaller error than the observed spread in Tsys.  So 
>>> I expect we should be able to calibrate the Tcals to perhaps 5% -- 
>>> certainly sufficient to judge whether the observed poor sensitivity 
>>> at C-band is due to high Tsys.
>>>
>>>     The results of this exercise are as expected (sadly).  After 
>>> correction by this procedure, the 'cold sky' system temperatures for 
>>> all antennas (EVLA and VLA) are typically 40 K to 80K, and correlate 
>>> very well with the observed sensitivities (as derived from 
>>> correlator coefficients, which are independent of the measured 
>>> Tsys).     Some details:
>>>
>>>     1) There is no difference in the mean Tsys for EVLA and VLA 
>>> antennas -- about 60K.
>>>     2) The lowest Tsys values are from EVLA antennas 14, 16, 18, and 
>>> 4 -- about 35K in both RCP and LCP.  Antenna 13 (which also has 
>>> decent sensitivity) did not fringe in these tests, and gave a zero 
>>> degree increment on Cyg A.  .   In 2005, Bob Hayward and I measured 
>>> antenna 13's Tsys (by hot/cold load tests) to be 24K, with an 
>>> efficiency of about 0.55.  Presuming 13 is similar to the others, 
>>> the Tsys appears to have degraded by at least 10K since then -- or 
>>> the efficiency to have dropped to about 0.40.
>>>     3) All other EVLA antennas have Tsys values higher than 50K -- 
>>> that's twice the expected (and required) values!!!
>>>
>>>     4) The three VLA antennas with remarkably good sensitivities 
>>> have the lowest system temperatures amongst the VLA antennas -- 20, 
>>> 22 and 27 all have Tsys values about 45K.
>>>     Although not a precise substitute for proper measurement of Tsys 
>>> (via correct values of Tcal), these high Tsys values are very 
>>> unlikely to be caused by deviant system efficiencies.  The strong 
>>> (but tentative) indication is that there is something seriously 
>>> amiss with our C-band EVLA receivers.
>>>
>>> _______________________________________________
>>> evlatests mailing list
>>> evlatests at listmgr.cv.nrao.edu
>>> http://listmgr.cv.nrao.edu/mailman/listinfo/evlatests
>>
>> _______________________________________________
>> evlatests mailing list
>> evlatests at listmgr.cv.nrao.edu
>> http://listmgr.cv.nrao.edu/mailman/listinfo/evlatests