Main INDEX
Monthly INDEX
PREV
NEXT
User name Pitt
Log entry time 21:36:33 on November 6,2002
Entry number 51089
keyword=Proposed run plan for owl shift
Proposed run plan for owl shift tonight
________________________________________
NOTE: There is also a request to the analysis shift near the end of
this message.
This plan should be done at the same current as last night; in fact the
conditions should be as close as possible to what they were last night.
Basically do the same program as last night with the following exceptions:
For tonight, we will have the feedback program do the charge feedback
automatically, but the position feedback should be done manually.
1. When you start up the feedback, you should select BCM=2 (so it
uses BCM2 for the charge feedback) and BPM=2 (so it uses BPM G0B; the
last stripline).
2. For the runtype on the DAQ, you can put it into ts_scaler mode. We
don't need all the other information.
3. When you start up the feedback, disable position feedback and only
enable charge feedback. Probably 10 minutes is a reasonable update
time for the charge feedback, but adjust it as you think appropriate
given the size of the errors.
4. For the position differences, you will keep track of them using the
G0 realtime monitor. Instructions for this should be somewhere in the
manuals, or you can ask the previous shift crew how to run it. You
will want to look at the X and Y position difference histograms for
G0B (I'm not sure what they are called in realtime).
5. Start up the realtime and take data for about 30 minutes; compute
the postion differences in x and y (and their errors) and note them
in a log file.
6. I have written a code that will compute the desired change in Vx and
Vy. It also takes account of the fact that there is a significant
coupling between the position differences and the IA voltage. Everything
is found in /home/gzero/users/pitt/feedback. Before you use it the
first time you must set the correct calibration constants in
cal_matrix.dat. Help on this is in the README file.
Then just type pos_change, and it will ask you for the measured
x and y position differences. It will return the needed CHANGES
to V_X, V_Y, and IA voltages. The returned values should be ADDED
to the current values.
7. After the changes are made you can kill realtime and restart it
(or there is also a clear button but Wolfgang told me he hadn't tried it
before). Then wait another 30 minutes and make another change. Leave
the main DAQ run going the whole time.
To the analysis crew:
The following two types of plots would be useful:
1. Postion differences (X and Y) and charge asymmetry versus time within
a run. Each value should be binned in intervals of about 5 minutes (2250
quartets. I think Jason basically already has such a script written.
2. Integrated values of the above quantities versus time. This is different
from step 1. The plots should still have a point plotted every 5 minutes
or so, but each point will be the average of that quantity back to
t=0. So for charge asymmetry you would have:
t value
5 min charge asymmetry average from t=0 to 5 mins
10 min charge asymmetry averaged from t=0 to 10 mins
15 min charge asymmetry average from t=0 to 15 mins
If the feedback is working this should clearly illustrate that it is homing
in more precisely on zero as time progresses.