[evla-sw-discuss] OSRO OPT - Sampler Outputs and Subbands

[Michael Rupen]

> In Tuesday's SSS meeting i promised i would ask some questions of a wider audience.  This
> is the 3rd of 3 emails.
> 
> We are trying to make the OSRO user interface in the OPT look as similar as possible to
> both the existing VLA correlator setup already in the OPT and the PST's presentation of the
> OSRO modes.  Scientists who attend SSS mtgs are helping us craft this; see latest mock-up
> at: http://www.aoc.nrao.edu/~dlyons/osro-mockup3.html
> 
> On the OSRO page Table 1 is used for full polarization setups in which there are two
> independently tunable subbands.  In the mock-up you'll see that for each subband we're
> allowing independent Doppler tracking.  This is made possible by placing each subband in a
> different sampler output pair (A/C vs B/D).
> 
> This works well except in Ka band.  In Ka the A/C IF goes only as low as about 32GHz.  This
> means w/ the approach of putting each subband in a different sampler output we cannot
> support certain requests (eg, subbands at 28GHz and 29GHz).  Of course, we could put such
> combinations in the B/D pair, but only one of these can then be Doppler tracked.
> 
> Since OSRO is pretty restrictive already, will we tell users that some combinations are not
> supported?


I do not think we should support multiple subbands within a single baseband 
(sampler).  This requires much more "smarts" of ObsPrep etc.: we have to be 
sure the subbands all fit within 128 MHz slots; we have to use the mixer in 
the StB if the subbands are not within even narrower slots; etc. etc.  Plus as 
Dave says we can't easily Dopper track two subbands within the same baseband.
We specifically designed OSRO to avoid dealing with all this.  It's 
unfortunate we didn't appreciate the problems with Ka band but I'd put that 
down as part of the shared risk.

          Michael