[mmaimcal] Re: [alma-sw-ssr] Band switching times revisited...

[mel wright 456]

Hello Steve,

	thanks for collecting this info together. I had
seen Darrel's response before when asking about frequ
switching. I don't think there is anything terrible here;
it is, as you say, just one of the "features" we must work
with. It may not be as bad as it sounds. when we freq switch
at hatcreek we turn the LO's on and off. that part seems to
work OK (but maybe not well enough for ALMA standards !)
Anyway, clearly there will be subarrays for simultaneous
obsevations in several bands. The observation preparation
tool needs to know about Rx band change overheads, just as it
needs slew time overheads, in order to build efficient
observing scripts.  For dynamic scheduling, I
think of it as a supply and demand equation. The system
has a certain status, including currently hot Rx's. The
stack of observing objects (yet another reason to think
of SBs in logical groups !) are the demand side of the
equation, including required Rx's. The OO priorities
and requirements can build up demand for a change in
status (like swap hot Rx bands).

cheers,

	Melvyn