[evlatests] Calibrating 3-bit data

[Bryan Butler]

it'd be nice to have that documented properly, to avoid 'lore' building 
up around it.  'a few %' vs. '5 %' vs. '10 %', at which bands, under 
which conditions, etc.  problem is, you start telling observers they can 
get to 'a few %' accuracy without observing a flux density scale 
calibrator, and they'll believe you, and could get bit by it.  so, i 
suggest a memo, author list dhawan and perley.  ;).

	-bryan


Rick Perley wrote, On 1/30/12 8:37 AM:
> Vivek's statement arises from the AIPS calibration corrections
> ('CALIB'), on data which have been corrected by the switched power
> tables. He and I have always noted that these correction factors are
> within a few percent of 1.0, where 'a few' is typically 5%. (There are
> usually a few antennas which are quite deviant -- usually due to
> problems in the SY tables). It's certainly true that the high
> frequencies are worse, but that's easy to explain in terms of
> extraordinary atmospheric absorption, bad pointing, winds, etc.
> If you have trouble accepting these statements, I have some 'flux
> density' data you can try this out yourself ...
>
> Rick
>
> Bryan Butler wrote:
>> i've never seen a definitive, and *repeated*, demonstration of this.
>> i'd guess it's more like 10% at best, on average (you might get lucky
>> and get better on occasion). at least at K-band and above. it might be
>> slightly better at the lower bands. maybe you have data that shows it,
>> but i'd like to see it.
>>
>> -bryan
>>
>>
>> Vivek Dhawan wrote, On 1/29/12 17:34 PM:
>>> The 8-bit data have been shown to calibrate to
>>> a few %, compared to the classic method with 3C286 etc.
>> _______________________________________________
>> evlatests mailing list
>> evlatests at listmgr.cv.nrao.edu
>> http://listmgr.cv.nrao.edu/mailman/listinfo/evlatests