we have been using the show_deep_asym_all_*.C macros as a daily (if not
per shift) measure of polarization. as such, for the method that we've
been using it, there simply isn't much data. increasing the number of
bins in the fit region worsens the bin-by-bin statistical errors which
worsens the chi^2. that is primarily why the macro has been setup to fit
on three bins and not a higher number of bins. in my experience, the
show_deep_asym_all_*.C macros are good for measuring relative day-to-day
changes in polarization but not so ideal for measuring the actual
polarization. after all, the show_deep_asym_*.C macros are designed for a
QUICK, not thorough, polarization determination.
if you want an overall realistic measure of the polarization, you
must collect data for multiple days and then either 1) augment the
show_deep_asym_all_*.C similar to what it sounds like vitaliy did or else
2) run the full polarization programs. currently, i am the only one
running the full polarization programs b/c they are more complicated than
the quick show_deep_asym_all_*.C macro. i shall write up a explanation on
how to use the full programs.
aaron
On Tue, 22 Mar 2005, Electronic Log Book wrote:
> Operator: vziskin
>
> The shift started with few empty target runs. After this, Genya set up the ABS and we were on our way to polarized running. However, the sextupole (???) broke shortly after and most of the shift was occupied by fixing it. Got the beam back at around 23:30 and running smoothly thereafter. L7 trips on the average twice a fill. The max current is set at 200 mA. Larry tried to go higher by the background was bad. Dave is keeping it at 200 mA for now. The lifetime is quite good, ~28 min.
> Since, there was no beam most of my shift I got a chance to look at the polarization for the past four days. First of all, I was the one who changed the polarization script in first place. The reason for this was that the fit is being done in the momentum transfer range of 0.1 <Q^{2} <0.35. However, the limits of the histogram were set between 0.05 and 0.65. The total number of bins was set to 6. This means that the fit was actualy done with respect to two bins at most. If one get a good \chi^{2} with this fit, it is only by accident. Since the cuts demand a cerenkov on the electron side there is no reason to plot events beyond Q^{2}=0.35. Also, anything below Q^{2}=0.1 is outside of our physical acceptance and more importantly outside of the "Arenhovel grid" acceptance. So the fit should be limited to the range in the momentum transfer of 0.1 to 0.35 (GeV/c)^{2}. Since this is the fit range, we may as well increase the number of bins in the range of our fit to!
6!
> . This was the reasoning for the original change. Sorry for not communicating this right away. But I believe that this was the procedure that we used last year.
> After applying this change I looked at the polarization from all the runs taken after the leak in the dissociator was fixed. In total we collected 34kC. The product of the beam and target polarizations is (see attached figure)
> Left : 0.51825 +- 0.019495
> Right : 0.455586 +- 0.01311
>
> Clearly, the two sectors are different. This difference could be attributed to the "kinematical corrections", target angle, lack of full BLASTMC analysis, etc. Correct me if I'm wrong but we saw the left/right polarization differences last year. And I believe that the difference is still persitent at high Q^{2} even when we applied the full analysis.
>
> Regardless, the target appears polarized, though the polarization may not be as high (~15-20% lower) as last year.
>
>
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:32 EST