[evlatests] FRING aborting, solved (from an AIPS perspective)

[Steven T. Myers]

Michael,

In the SDM SpectralWindow.xml table for this data I see:

         <spectralWindowId>SpectralWindow_0</spectralWindowId>
         <basebandName>BB_1</basebandName>
         <netSideband>LSB</netSideband>
         <numChan>256</numChan>
         <refFreq>8.672E9</refFreq>
         <sidebandProcessingMode>NONE</sidebandProcessingMode>
         <totBandwidth>1.28E8</totBandwidth>
         <windowFunction>UNIFORM</windowFunction>
         <chanFreqStart>8.672E9</chanFreqStart>
         <chanFreqStep>500000.0</chanFreqStep>
         <chanWidth>500000.0</chanWidth>
         <correlationBit>BITS_2x2</correlationBit>
         <effectiveBw>500000.0</effectiveBw>
         <name>Subband:1</name>
         <oversampling>false</oversampling>
         <quantization>false</quantization>
         <resolution>500000.0</resolution>

I don't know what was agreed with Michel (and/or Francois) if anything but is 
the intent of SDM to MS conversion that the sign of channel order be determined 
by the netSideband value or by the sign of the chanFreqStep (I would have 
expected the latter).  In any event, our filler is probably just using the step 
(which Rich says is always a positive number in the SDM) and not flipping it for 
LSB.  I'm not sure what ALMA intends either.

In my quick test changing chanFreqStep to -500000.0 in the SDM gave the right 
channel on import into CASA (havent checked that this works on export to AIPS).

If you think we should be using the netSideband then we'll file a JIRA for 
Michel on this.  I will also ask Robert and Michel what they are expecting.

   -Steve

On Wed, 14 Oct 2009, Rick Perley wrote:

>    Eric quickly discovered why FRING aborted on the X-band data:
>
>    The FQ table's entries for the channel increment were reversed in
> sign.  For X-band, this has to be negative.  For the X-band database I
> received it was positive.  (The problem arises because the sub-band
> centers decrease in frequency as the subband number increases.  The
> frequency index should also decrease (i.e., be negative), but were
> listed as positive, or frequency increasing.  This did bad things to a
> global solution.)
>
>    Eric flipped the sign of the frequency increment index in AIPS,
> after which FRING worked normally, both on the original data, and on the
> FLOPMed version.
>
>    The problem arises upstream from AIPS.  Steve is investigating ...
>
>
> _______________________________________________
> evlatests mailing list
> evlatests at listmgr.cv.nrao.edu
> http://listmgr.cv.nrao.edu/mailman/listinfo/evlatests
>

-------------------------------------------------------------------------
|:| Steven T. Myers                      |:|  Tenured Astronomer       |:|
|:| National Radio Astronomy Observatory |:|  Ph:  (575) 835-7294      |:|
|:| P.O. Box O, Socorro, NM 87801        |:|  FAX: (575) 835-7027      |:|
|:| http://www.aoc.nrao.edu/~smyers      |:|  smyers at nrao.edu          |:|
-------------------------------------------------------------------------