[Sane-analysis] Differing gen_run_.txt files and syncfilter executables

[O. A. Rondon]

Hi,

I have tried to do a systematic study of the BCM threshold set in
gscalers.param, following up on Jon's observation about different
helicity quantities depending on which BMC is gated. The point would be
to figure out if the analysis can be made insensitive to this choice,
and therefore more robust.

I replayed run 72614 for g_beam_on_threshold_cur = 25, 40, 50, 55 and 50
nA (always cutting on BCM1) and extracted the "Beam on" numbers from the
gen72614.txt report. I have posted the results as a text file [1] and
plots [2]

- the BCM's converge quickly for the threshold increasing above ~30 nA,
and they agree on beam on times for 60 nA, top plot-left axis [2]

- the corresponding loss of beam on time is very small above 40 nA, less
 than 2% (of total beam time), indicating that the statistics would not
suffer if the threshold were set to 60 nA, top plot-right axis[2]

- but there is a little mystery in the plot of charge difference vs
threshold: initially BCM1 records more charge (consistent with its
longer beam on time), the BCM's almost agree at 40 nA, but then BCM2
starts recording MORE charge as the threshold increases, while the beam
on time difference becomes zero, bottom plot. Any suggestions?

Finally, the gen72614.txt report for all cases always shows
BCM1 Q_tot [muC] = 301.782
BCM2 Q_tot [muC] = 301.565
which must be the charge without threshold cuts.

I am aware that the BCM's aren't accurate to better than ~5% at SANE's
currents.

Cheers,

Oscar

[1] http://www.jlab.org/~rondon/sane/analysis/beamon_55.txt

[2] http://www.jlab.org/~rondon/sane/analysis/beamon_55-sdc.pdf





Jonathan Mulholland wrote:
> Greetings,
> 
> There was some discussion about discrepancies in the report files output 
> by the replay at the last analysis meeting. The discrepancy in the 
> helicity info section of the report was caused by a change in the 
> gscalers.param file. The cvs has the syncfilter gated on BCM1, and I had 
> replayed some files with the syncfilter gated on BCM2--which was 
> responsible for the discrepancy.
>