[mmaimcal]Re: [Almasci] Antenna Acceleration

[Mark Holdaway]

> So I looked at this again and found that in both Mark's numbers and mine we
> were approximating the integrals by sums which means that the velocity
> effectively refers to the point half way through the previous step, etc. 
> Allowing for this fixes the discrepancy I noticed last night.  Here are the
> plots in a pdf.  I hope they are reasonably self-explanatory.
> 
> Best Richard

Richard,

Thank you for this verification!

There was some concern at our ALMA/NA Imaging & Calibration meeting
yesterday that the ALMA ACU might not be able to reproduce such an
antenna motion profile.  So, while you aren't a servo expert,
you are saying that according to the way YOU understand the ALMA ACU,
there will be no problem reproducing such an antenna turnaround.
(This is actually quite remarkable as there are just over 6 of the
0.048 s intervals in this turnaround.)

I should point out that we don't really care exactly where the antenna is
pointing during the turnaround -- in my simulations, the turnaround is
completely "off source", and by the time you are "on source" again, you
are up to speed at constant velocity.  So, the errors in the second plot
are completely inconsequential -- in addition to being terribly small. 
(Some efficiency gain can be obtained by beginning the turn around before
you are off source, but I have ignored that.)

Anyway, the tiny errors you get would indicate to me that the ACU -- as
understood by Richard Hills -- will be able to accurately execute
reasonable but complicated OTF whirly-type paths such as you have proposed
in the past, should we need them. 

Take care,

  -Mark