[daip] TY table problem

[Craig Walker]

On 2002.10.28 08:34 Jim Ulvestad wrote:
> This looks like a result of the same problem we discussed
> previously, whereby the station monitor data don't report
> the source change until a second or two after it actually
> happens.  This causes some ambiguity in which source is
> being observed at which stations at the ends of scans, and
> I could see it causing something like this.
> 
> What happens if you run QUACK with option 'TAIL'?  This
> won't do anything to the TY table, but it will flag the
> offending data and may prevent unwanted interpolations.
> 
> jim


Approximately a minute of data are affected in each of 2 scans
on only some antennas.  Running QUACK to drop the last minute 
of every scan because of a problem in two scans is rather too severe.

In the end, I used TBOUT, emacs, and TBIN to fix the TY table.
Rather a pain, but it seems to have worked.  

Back when I wrote the original TSM, I jumped through a lot of 
hoops to deal with time order problems and I think I was pretty 
successful.  In those days we could have whole hours of repeated 
data!  Why can't the current software deal with problems of a 
few seconds?  In fact, it looks like it does ok except occasionally
on the first scan of a correlator job.

Craig