[mmaimcal]Phase Repeatability of the Attenuators

[Mark Holdaway]

> 
> so, mark, you should account for at least 1.5 sec *between* the
> observations (this does not count on-source time, of course) in
> fast-switching - this should account for most of the antenna
> mechanical settling, and the LO settling.
> 
> 

The spec is 1.5 s, but Darrel had once said that it could be smaller.
In my Memo 403, I DID use the greater of 1.5 s and the slew + settle
time;  I found that 1.5 s dead time was generally well matched to
the slew + settle -- if we could do faster on the LO, it would
make only a marginal improvement.

I said "1 second", as sort of shorthand for "as fast as we can", which
turns out to be "1.5 seconds" when written in long hand.

	-M