[d2n-analysis-talk] H2 BB cer

[posik at jlab.org]

Hi Brad,

   So I am going to look at the BB cerenkov for the H2 elastic events and
try to get an efficiency as we discussed in our last meeting. However I
have a few questions.

  I would want to determine the efficiency by starting with good protons
in the LHRS. I then look at how many of these good proton events had an
electron event by looking at Bigbite. In addition the the proton cut I
also use a BB one track cut and a BB z vertex cut. So I now have a
"good" electron sample.
Should I also add a shower+preshower cut to this?

 Now using this electron sample, I will apply a TDC cut to each PMT and
see how many electrons are left.

 Finally I will apply an ADC + TDC cut to each pmt, and see how many of
the electrons are left. However if I make a cut on the ADC won't this
result in an inaccurate efficiency because of the ADC problem on the
beamline side(PMTs 1-10)?

Also I am just about done with the second round of BB MWDC calibrations,
it seems like the resolution on the residuals improved ~10-15um. So we now
have a residual resolution range of ~200-280um depending on which chamber
we are looking at. For comparison transversity had a resolution range of
~200-250um. I did not see much of a change in the number of tracks though.

One other thing that I wanted to check on is how I am doing my track
comparison. There is a variable defined in the BB opticsE06010 class
called vzflag. This takes 2 values, 1 or 0. If it is 1 then the optics are
valid and if it is 0 the optics are invalid. This is supposed to remove
the BB fringe field by removing the BB edges(See vz.png, Black is bendx vs
bendy and the red is bendx vs bendy with vzflag=1).

vzflag is defined by selecting portions of the variable bendx (middle
plane x) and bendy (middle plane y), also defined in the opticsE06010
class.

When evaluating the number of tracks I have like in last weeks meeting , I
have been doing the ratio of events (BB 1track cut+ T2 + vzflag)/(T2) ~14%

But I am wondering if I should be doing a ratio like (BB 1track cut + T2 +
vzflag)/(T2+ vzflag) ~90% ?

By the way when I use the vzflag cut all BB.tr.n==0 tracks are removed, I
think because bendx and bendy both require a reconstructed track.

Thanks
matt
-------------- next part --------------
A non-text attachment was scrubbed...
Name: vz.png
Type: image/png
Size: 279178 bytes
Desc: not available
Url : https://mailman.jlab.org/pipermail/d2n-analysis-talk/attachments/20100517/d44ca738/attachment-0001.png