[SBS_sim] tracking improvement ?

[Kalyan Allada]

Dear all,

I was playing around with Ole's tracking code and found that the code 
currently uses first 3 samples of the 10 samples that are in the 
simulated data(?). This has some consequences:

1. First 3 samples are still in the raising edge of the signal (so there 
is no slope change ..peak not identified)
2. The piece of code that calculates the charge integral assumes that 
the 3 samples are from peak region. So the integral may not be correct 
if all of them are on the rising edge (or falling edge).

I modifed the code to accept the 3 samples from the peak region (by 
looking at the no background runs) and also added an additional 
condition to reject the signals based on the ratios: consider 3 samples 
a, b and c.  For good events the ratios  (a/b <1 && c/b <1).

In the plot below you can see the tracking efficiency obtained from g15 
dataset for old method and new method(after my changes to the code). 
There is definitely some improvement as we approach full background. I 
don't know why the absolute numbers of tracking efficiencies are small, 
but this is what I got from the Ole's replay_xy.C script in the CVS. 
Maybe there are some cuts which I don't understand. We can discuss about 
it more tomorrow.

Kalyan






-------------- next part --------------
A non-text attachment was scrubbed...
Name: trk_eff.gif
Type: image/gif
Size: 6465 bytes
Desc: not available
Url : https://mailman.jlab.org/pipermail/sbs_sim/attachments/20110124/bee96479/attachment.gif