[mmaimcal]Phase Repeatability of the Attenuators

[Mark Holdaway]

> 
> i think you've got the timescales too short here.  isn't the limit
> on switching bands something like 5-10 seconds (larry, correct me
> if i'm wrong here)?  certainly the settle time is more than 1 second,
> so you'll have to have many seconds on calibrator, and some multiple
> of that on source.

I believe 1.5 seconds is stated.   If it is 10 seconds, there is
little point in fast switching.

BTW:  Darrel points out that the attenuator's phase won't scale with
frequency, which means

	* we MUST measure it (and other phase components
	which don't scale)independently of the fast switching
	observations and make a note of it and apply those phases
	before scaling the 90 GHz cal phase

	* the applicable number is more like 6.4 deg  (though
	if this is the case, I would be more stringent and
	say sqrt(16^2 - 15^2) = 5.5

	-M