[mmaimcal] Vertex Antenna turn around time

[Mark Holdaway]

> > around time.  So, Vertex is within a factor of 8 of meeting Mark's
> > simulated break-even point.
> > 


Its more complicated than a factor of 8.

The antenna slews at 0.5 times the rate.
The turnaround time is 4 times as long.

There are two sources of noise: thermal (determined by the
duty cycle) and atmospheric emission fluctuations.
Larger sources and higher frequencies are dominated by the atmospheric
fluctuations.  For smaller sources and lower frequencies, the
two terms are more comparable or the duty cycle dominates.

Now: the duty cycle is 

	source_size/slew_rate
       ----------------------------------------
	source_size/slew_rate  + turnaroundtime

which is certainly not a factor of 8 higher now (ie, going slower
HELPS you to get a larger duty cycle).  Also, you need to take the sqrt of
this to go to sensitivity.

The turnaround time is not a big deal for larger sources, but the
slew rate will reduce OTF's ability to subtract the variable atmospheric
emission.  It is not a full factor of 2 worse, as the root structure
function goes like t^{0.6} or similar...  so it is sort of
sqrt(2) worse here. 


This is complicated, as there are many cases to consider, none of
which i have time for.

But it seems that while BS and OTF used to be comparably good
even for sources one beam acorss, now BS will be superior to
OTF for the smallest sources.  A back of the envelope calculation
might suggest BS wins for the smallest sources, and the crossover
point where ORF begins to win is when the source gets to be
somehwere like 2-4 beams accross, ie, a good size for BS to work.


	-Mark