�� . �� : http://www.naic.edu/~phil/pnt/makingthemodel.html
�� : Mon May 21 20:03:37 2012
�� : Tue Oct 2 06:52:20 2012
��:
�� : � � � � � � � � � � � � � � � � � � �

pointing

Steps in making the pointing model

27jan12

The steps in making a pointing model are listed below. I've tried to give an estimate of how much time this takes. All estimates assume:

no equipment failures
no rain
distomats continue to work (through any fog)
temperatures are low enough that we don't lose tension in the tiedowns at high za when pointing at a tower.

Steps in making a model:

make sure the weight distribution on the platform is stable (no planned moves of heavy objects).
Optional:

survey the platform from ao9
reshim the elevation rails if the ao9 survey warrants it

Find the tiedown relative positions that remove any residual tilt in the platform

The tilt sensors on the rotary floor are used to measure this
do a floor spin of the rotary floor to measure the floor residual tilt (takes a few minutes)
Do 2 or 4 az swings at slew speed while measuring with the tilt sensors

1 swing is 270 to 530 and takes about 15 minutes.
this can be done during the day as long as the temperature does not cause a relative tilt in the platform during each 15 minute spin
Time: about 1-2 hours

Compute the tiedown changes needed to remove the residual tilt.

fit the 1az term getting amplitude and phase. this tells which tiedowns need to be moved.

Redo the az swings with the new tiedown positions. verify that the tilt has decreased (there is a 50% chance the correction is put in with a negative sign!).

If in a hurry this can be on the same day. It's probably better to do the check on a following day to give a chance to look over the results.

When this is completed, we will know the tiedown positions to level the platform to about 10 asecs (that's about as good as we can do with the tilt sensors). The rest will have to be removed with the model.

Find the tiedown average positions that puts the receivers in focus.

A focus curve is done on J2253+161

While tracking the source rise to set, the tiedowns are pulled through a range of about 14 inches (taking 14 minutes)
the dome is moved up and down (once every 2 minutes) and the turret is swept back and forth in a sine wave. These let us measure the peak gain every 2 minutes. we get 7 measurements for each up or down of the tiedowns.
We need a source close to 18 deg dec so we cover a large za range.
We need a strong point source as sband because we want to measure the peak even when we are out of focus (so we can determine the best za location).
We don't want to lose tension in the tiedowns while moving between td=20" and td=6" (maybe we can go 21 to 7 ").. so it can't be very hot out. When J2253+161 moves into the after noon this can be a problem (i'll probably need to find another strong source).
Time: about 2.5 hours.

Take data for the model (see model17 dec07 example)

Use the updated tiedown positions and the old model.
track 40 sources spaced 2 degrees in declination to cover the entire dish (using heiles calibration scans).
Time: this takes about 80 hours (averaging 2 hours/source with wait time..)

We need to have at least a 12 hour coverage of lst so we have enough sources to choose from. The scheduling for the last sources may not be too efficient since we may have to wait for some of the last sources to rise.
We can run into the morning and start earlier in the evening depending on the temperature.

in Feb you can probably run till 10am.
in the evening you may be able to start around 16:00 or 18:00 (depending on how hot a day it is).
You probably want to make sure that the high and low dec sources are taken when it is colder.

Compute the model:

This can take a day to sort things out..

Transfer the model taken with sband to the other feeds:

We need to determine the feed offsets on the floor for each of the other receivers.
track 6-8 sources with sbn and the new model.
On the next days track 1 of these sources with each of the other receivers to measure the offset.

lbw,sbw,sbh,cb,cbh,xb give 6 receivers. with 6/night takes 2 nites
800,327,430 need to also be done but they are not as critical.

time: 2 nites

Files that need to be updated:

If the tiedown reference position has changed:

~phil/vw/tdAutoTrk.h .. put in new reference. make sure temperature at ref is the same.
/home/phil/vw/datatk/pnt/Prog

make
Then need to reboot point crate (or stop pointing programs, then < loadpntprog in vxWorks in /home/online/vw/load directory)

/home/online/Tcl/Proc/pnt

tieref.proc .. put new reference position in here
tietrk.proc .. make sure offsets are 0.

/share/megs/phil/svn/idl/td

tdparms.pro - update new reference
../make install

Hopefully this is it..

Summary of time:

The time line needs to go in order (maybe residual tilt and focus curve could be interchanged).

	time	Notes
measure platform residual tilt	2hours on 2 different days	can probably be done on 2 cool afternoon2
Measure the focus curve on J2251+161	3 hours Possibly two days	but needs to be done when it's cool. (this is a problem in feb since the source transits around 14:00) Might want to try this i 1 day, then find another source.. so this should not be scheduled immediately before the model datataking..
take data on 20 sources	4 12hour days	about 40 hours on source.. plus time to wait for sources to rise Assuming 12hours/nite 17:00 to 7am 40 hours is 3.33 days so 4 days 12 hours a day will probably work when you include the time for the source. If the temperature is cool we can start earlier and end later. this could speed this up.
compute the model	1 day off telescope
transfer model to other receivers:	12 hours 12 hours*2days tot:3 nights	track 6-8 sources sbn retrack with the other 9 receivers (maybe leave out 327).

home_~phil