[mmaimcal] Correlator reconfiguration time

[Harvey Liszt]

Brian--

You asked:
> 
> What is it about OTF that requires a fast correlator reconfiguration? Going
> to an off? 

No, just the abiltity to dump data multiple times once the device is set up.
As I said, the GBT correlator sets up (globally reconfigures) fairly slowly 
but rapidly refreshes in a fixed mode.

Here's how we crippled the 140' for surveying:

For OTF, data have to be dumped at intervals well below a telescope beam 
crossing time.  At the 140', one needs at the very most a few sec integration 
per independent beam to get down to the confusion limit for galactic hydrogen.
So on each row of an OTF map, you want to spend perhaps 1 s crossing a beam:
in making an OTF map the total amount of integration on each independent beam 
is increased because adjacent rows are spaced well below the beam size.  So 
the correlator dump time has to be << 1 s.  But the 140's vintage 1982 auto-
correlator, run by an onboard Varian microprocessor to save a few $$$ and
NEVER reprogrammed since, can't complete 1 cycle (set up, take data, and 
dump) in less than 20-30 s and it can't dump data without going through this
cycle.  It is perhaps two orders of magnitude shy of the basic requirement 
for OTF surveying of galactic hydrogen.

> What observations for the MMA do you think will require the
> fastest reconfiguration?

Trying to beat down the quasi- 1/f noise in HEMTS, or sky noise, in
total power mapping.  And, if there is a switching secondary, something
related to chopping.

We should probably have a contest.  I can't claim familiarity with all
the possible modes that might demand fast dump times, but they always
turn out to be less than what you planned for.  For the GBT, the new
standard continuum back end was originally designed to have a ~ 1 s 
cycle time.  Try using that to switch out 1/f noise.

regards, Harvey