[d2n-analysis-talk] E/p Correction & Effects on Pion Rejection Efficiencies

[Brad Sawatzky]

On Fri, 11 Dec 2009, David Flay wrote:

> I've made corrections to the E/p vs. p plot I showed on 11/17/09 so
> that E/p lies along a straight line now (see first plot).  Yesterday
> I've also made a plot of sigma/p vs. p, where sigma is obtained from a
> Gaussian fit to the E/p distribution (see second plot).  This fit is
> consistent with Patricia's work, however I am not too comfortable with
> the leading parameter being negative -- I'm currently looking into
> this.

What is your fit function for the pr_E_P_sig-p_12_11_09.png plot?

> Now, as for the effects on the efficiencies -- if you recall from
> 12/1/09, I showed the pion rejection efficiencies in both the GC and
> the PR.  They both tended to have a momentum dependence.  In
> particular, the E/p cut used in the PR in conjunction with a cut in
> the GC had a very pronounced effect.  This is the cut combination I
> investigated first.  My corrections to the E/p drift as a function of
> p seem, at the moment, to have little to no effect.  I'm not too sure
> as to why this is (see third plot).

When you get the chance, please generate plots that show (E/p) and
(E/p)_corrected overlaid on top of each other.  I'd like to see 5
separate plots, one for p0 = 0.6, 0.9, 1.2, 1.5, and 1.8 GeV/c.
(Or just 0.6, 1.2, and 1.5.  Don't spend a whole lot of time on this, I
just want to see the trend.)

> Finally, I revisited the Gas Cerenkov pion rejection efficiency as a
> function of cut position in the Cerenkov.  This as well seems to not
> change by much now that E/p has been corrected (to make sure I was
> testing the correction, I now select particles in the E/p (corrected)
> distribution, as opposed to the L.prl1.e vs. L.prl2.e plot).  However,
> I calculated the percent differences for these three data sets here,
> and on average the percent difference of p = 0.6 to p = 1.20 is
> ~0.25%, while the percent difference for p = 0.6 to p = 1.70 is
> ~0.68%.  These percent differences seem reasonable, but the effect
> looks more pronounced on the plot (see last plot).

I suppose that tells us that 
  a) the high GC cut ( > 700) leaves a pretty clean electron sample in
     the PR, and
  b) the clean electron sample generates a reasonably tight peak
     in the corrected E/p plot (making your efficiency largely
     independent of the E/p cut value).
I think this is a Good Thing(TM).

The fact that the overall efficiencies are 'lower' for 1.70 GeV (the
cleaner electron sample) than for 0.6 GeV (more pions) bugs me a bit.  I
feel like the explanation should be obvious, but it doesn't quite make
sense to me right now -- I guess I'll have to dig up your definition
of efficiency...

-- Brad

-- 
Brad Sawatzky, PhD <brads at jlab.org>  -<>-  Jefferson Lab / Hall C / C111
  Ph: 757-269-5947 -<>- Pager: 757-584-5947 -<>- Fax: 757-269-7848
The most exciting phrase to hear in science, the one that heralds new
  discoveries, is not "Eureka!" but "That's funny..."   -- Isaac Asimov