Log entry time 13:54:52 on June 17, 2001

Entry number 63559

  There has been some recent confusion about charge asymmetry and feedback
(some of which I have only just understood myself). Here are some
notes and observations. Apologies to those of you who already know
some/all of this already, and some of this is in previous halog
entries:

1) There is no feedback working at present, except the manual feedback
  provided by one of us, using the "Helicity Correlated Feedback Controls"  
  GUI. How this GUI works, at present, is this:

   a) you type in a value (in ppm). This is translated into a *change*
       in the value of the angle of the rotatable half wave plate, as soon
       as you click "timestamp". The present calibration appears to be
       32 ppm/step of the stepper motor that rotates the plate. The 
       value of this stepper is stored in EPICS in the parity DAQ 
       datastream, and can be looked at by doing the following:
         grep -a psub_aa_pos /adaql2/data2/parity/parity01_803.dat
      to get the value readout for run 803 (with each EPICS event).

   b) Note that this means that if you have the value "130 ppm" typed
     into the GUI and you "timestamp" it twice, then it will correct
     for a 260 ppm asymmetry (i.e. it will move the plate by 8 steps,
     not 4 steps). 

   c) Note that whatever clock sets the "timestamp" is about 3 hours 
     behind. 

   d) Note that the actual asymmetry dependence on the halfwave plate
      step was last measured (B. Michaels Halog entry 63195, June 15th) 
      was 26 ppm/step - if this still holds that we should "undercorrect"
      by 26/32 = 81% or so.  

2) The parity DAQ-based analysis of the charge asymmetry ("apar") is 
  not giving correct values of the asymmetry, for reasons that are 
  not yet known. The spectrometer DAQ-based version "qasy" appears
  to work, and to agree between right & left spectrometers pretty
  well, as well with the "Dixon 1-min avg asymmetry" provided by
  the MCC. This later can be watched via the "Hall A Charge Asymmetry"
  GUI (charge_asym.adl) - I suggest we watch this.

3) The "qasy" analysis is running in the background at the end of every
  run; the results of recent runs can be read by typing
      chkfeedback s
  You can also run the program by typing:
      goqasy   (to get to the correct directory)
      qasy   (and follow the directions to use the asyfit.macro to
              get Gaussian fit results) 
            
  Beware that if you don't enter the run number correctly, the script
  doesn't complain but uses some stale data for the fit... so make
  sure that the data file it is reading is the one you want!

4) the "chkfeedback" summary results don't exactly agree with the 
  results I get when I run qasy and the fit macro by hand myself; 
  eg. for run (20617/1617/817): 
  
  (left arm)  20617 - I get  97, 82, 65, 84 ppm respectively from the 4 BCMs,
               average = 82 ppm 
             
  (right arm)  1617 - I get 75, 79, 86, 72 respectively from the 4 BCMs
                average = 78 ppm                     "

    "chkfeedback s" gives   130 +- 41 ppm

   I don't understand the difference, but it rarely seems to get this
   large; maybe different cuts on when the beam drops out?

 5) There are "natural" drifts in the charge asymmetry of order 400 ppm
  or so seen in the data (i.e when the rotating halfwave plate is kept
  fixed), over time periods of a few runs or so, i.e. compare the
  results of 
      run 20816/1516/816   69 +- 36 ppm 
      run 20818/1518/818  383 +- 42 ppm
    so we do need to be vigilant in nulling it out. However, it also 
  means that asymmetries of less than about 30 ppm can't be nulled
  with the present setup since this is less than the discrete step
  size.

(D.S. Armstrong)