[daip] Plotting GMRT baseline position angles

[Andy Biggs]

Hi Eric. It turns out that that definition makes perfect sense. Somewhere along the way I got my clockwise and anti-clockwises mixed up.

That task you mention sounds quite useful. There's also the issue of the lower sideband having the frequency axis the wrong way round and so sharing the same reference frequency with the upper sideband. You also mentioned that FQID was set when it probably shouldn't have been.

Thanks,

Andy

On 23 Jan 2012, at 17:34, Eric Greisen wrote:

> Andy Biggs wrote:
>> Hi. In trying to diagnose a feature in my GMRT data, I've become a bit confused about how various AIPS tasks define the baseline position angle. This is partly because the help and explain files for VPLOT are inconsistent:
>> HELP - uv p.a. (deg N->E)
>> EXPL - uv p.a. (CCW from u-axis)
>> UVPLT is consistent, both saying N->E. However, the values plotted don't seem correct. For example, if I take a N-S baseline and plot the baseline position angle versus time, at transit the baseline claims to have a PA of -90 degrees, instead of the zero degrees that I would expect.
>> Regarding the feature in my data, I'm pretty sure that this is due to bandwidth smearing. I've removed all sources apart from one that is far from the phase centre and, when I average my data in frequency, the amplitude clearly decorrelates at all apart from a particular baseline position angle. The amplitude peak in VPLOT (and UVPLT) occurs at -55 degrees and my source lies at an angle of 55 degrees relative to the phase centre (as reported by IMDIST). If the bandwidth smearing explanation is correct, then I think that these two directions should be orthogonal to each other. I'm confused though:
>> * If the N->E explanation is correct, then the orthogonality is out by 20 degrees.
>> * If "CCW from u-axis" is correct, then they are exactly parallel.
>> As the GMRT has a different phase convention, could this be partly to blame?
>> I should also mention that the N-S baseline mentioned above will plot using BPARM=11,4 in UVPLT, but the equivalent plot in VPLOT fails with the following error:
>> localh> VPLOT1: Task VPLOT  (release of 31DEC12) begins
>> localh> VPLOT1: Multiple IFs not meaningful here
>> localh> VPLOT1: Doing no flagging this time
>> localh> VPLOT1: XYSC: AXIS 1 DEGENERATE
>> localh> VPLOT1: SCALVB: XYSCL ERROR  1
>> VPLOT will show amplitude, phase, weight, etc., but u, v, w and pa do not work.
>> Thanks,
>> Andy
>> _______________________________________________
>> Daip mailing list
>> Daip at listmgr.cv.nrao.edu
>> http://listmgr.cv.nrao.edu/mailman/listinfo/daip
> 
> The definition in the code for U-V position angle is ATAN2(V,U).  That would make the EXPLAIN file in UVPLT the only correct explanation.
> 
> GMRT uses opposite sign convention so the u and v both have opposite sign and so the position angle should differ by 180 degrees from what one would expect.
> 
> I do not know what your problem with VPLOT is, my GMRT file plots PA just fine.
> 
> I wonder if I should write a task FIXGM which would reverse all baselines (sign of u,v,w,phase and swap RL and LR) and then change the array name to e.g. TRMG.  That would reduce issues here...
> 
> Eric Greisen
>