Charge Asymmetry Feedback
Robert Michaels, rom@jlab.org, Jefferson
Lab Hall A, updated Aug 1, 2001
Feedback to control the helicity correlated
charge asymmetry runs automatically from Parity DAQ.
Thanks J. Grames, A. Vacheret, B. Bevins, P. Chevstov
TURNING ON AND OFF
To turn on or off feedback script, login
to an adaq linux box like
adaql1 as ``adaq'' and
execute one of the following. NOTE, this only turns
on/off the script called by Parity DAQ. If you turn
it off, you can still interact with the EPICS Gui
for Charge Asymmetry control.
HOW TO CHECK
The following procedures can check if feedback is working
and can measure the charge asymmetry.
- You can see the info get updated in the EPICS
Gui for Charge Asymmetry Feedback control.
Periodically a new value of Qasy, etc, gets written.
Warning, the time stamp is always shifted.
Warning, if you push time stamp by hand, you
force an over-correction. If the ``Rotatable
Waveplate Readback'' in this Gui changes, the feedback
is probably working (the whole chain).
- Check the EPICS variable for the rotatable waveplate being
controlled ``psub_pl_pos'' (this was changed at some
point in the run !) from server 129.57.255.4
- Execute the command ``chkfeedback'' from
the adaq account on Linux box like adaql1 to print out
recent results. You can also ``chkfeedback s'' for
spectrometer DAQ results.
- To measure the asymmetry ?
The ``do nothing'' method is to look at end-of-run
halog entries. You can also run ``apar'' or ``qasy''
codes to measure the charge asymmetry. The
apar code is the best to use, see Antonin Vacheret's page
www.jlab.org/~vacheret.
The ``qasy'' code uses spectrometer DAQ scalers --
login to adaql1 (linux) as ``adaq'',
type ``goqasy'' to go to the correct directory,
then type ``qasy''. After the first plot, type ".x asyfit.macro"
to obtain precise results which get printed on the window
where you typed this.
THINGS THAT CAN GO WRONG
- Feedback does nothing ? Maybe the server crashed
or MCC forgot to turn something on.
Or it is ``disabled'' in the ``Helicity Correl. Feedback Ctrl''
EPICS Gui.
It's a good idea to watch the asymmetry. Also
check the ``Rotatable Waveplate Readback'' (see the EPICS Gui).
Does it go
the right way ?
- Wrong slope of Asymmetry versus Waveplate-Set-Point
can lead to divergence.
A ``reversal waveplate'' can flip the sign of slope.
This has been automated, it should be ok, except that
this slope has been known to drift by factors of 2 in
a week, so it should be re-measured periodically.
- If you think the automatic feedback is wrong, you can
type ``feedback off'' and do it by hand using EPICS Gui.
MORE INFO
- Aug 1: A faster timescale of feedback is being
deployed. It runs in real time and applies feedback
every 10 minutes; also, the algorithm is more reliable
against beam trips. The user interface should be the
same as what I described above.
This page maintained by
Robert Michaels.
rom@jlab.org