NPS » NPS software » NPS analysis » NPS DVCS

No TL live time and trigger efficiency correction applied

Applied TL live time and trigger efficiency correction

Compared the results with 1.5GeV threshold and corrected the random events subtraction with pre-scale factor

It turns out that the previous big negative slope was due to the wrong value of branch "NPS.cal.nclust" in the root file which means I didn't loop all the clusters for each event.

I did some test and found that when the first 0 appears in the "NPS.cal.clusE" array, all the rest values in that array are 0. This allows me to calculate the real cluster number myself easily.

I used the real cluster number in my analysis script and finally got a much better result than before. The fluctuation of the charge normalized DVCS yield is only around 5% now (Emax algorithm).

I also tried the "only 1 cluster > 1.2" algorithm which means I only consider the one good cluster event as a DVCS event. The fluctuation of the charge normalized DVCS yield is also around 5% if we ignore the strange points at low beam current.

1. Analyzed run 3104 and 3105 (one photon trigger runs), the slope is 0.0029. Previous result for KinC_x50_4 is -0.0074.
2. With the help of Mark, we found that the "NPS.cal.nclust" is correct. But for some events, there could be more none-zero values in the clusE and clusT array than "NPS.cal.nclust".
Those unexpected none-zero values are absolutely the same as previous event (but it happens randomly).
3. Previously I thought the "NPS.cal.nclust" branch is wrong so I used all none-zero values for each event to do the analysis again. Surprisingly, I got a much better result than before.
The fluctuation of the charge normalized DVCS yield is only around 5% (Emax algorithm). And the slope is very small (0.0007).

I'm a little confused why double counting some clusters can help improve the result and what it implies.

Updates for the analysis meeting on Dec 04

This time we have much better results for pi0 and dvcs charge normalized yield.
But it's a little strange why it's not a flat distribution for LH2 when the energy threshold is 2.0 GeV. It implies there is still some thing we are not that clear. I'll try to figure it out.

Added one more page compared with "Updates_Dec06.pdf". The new page shows the gain loss effect on the charge normalized DVCS yield.

Charge normalized yield result from Kinc_x50_2 to Kinc_x50_1'

Some of those kinematics showed a decreasing yield with the run number.

Charge normalized yield results from Kinc_x50_2 to Kinc_x50_1'

Some of those kinematics showed a decreasing yield with the run number.

Checked the new runs for Kinc_x36_5'

New yield results for Kinc_x36_5' with new calibration coefficients

Results for KinC_x60_4a and KinC_x60_4b with EDTM live time correction

Calibrated KinC_x60_4a results and slope table which indicates NPS calo dead-time

LD2 and LH2 runs in the same yield plot

Applied tighter beam current cut but no stable condition was required

Applied tighter beam current cut and also stable condition was required (abandon +/- 3 scaler readouts around beam trips)