[Sane-analysis] Bigcal Re-Calibration Progresss

[O. A. Rondon]

Hi Whit,

Thank you for the plots. I agree that the top-bottom split is mostly not
due to different backgrounds.

About the eta peak for 180 data, I noticed that the fit errors for the
eta mass (528+/-146) and width for the final calibration are huge
compared with the three previous ones. Something happened in the final
calibration. The fit errors for the 80 deg. data (550+/-0.1) look like
the errors for the first three 180 deg. plots. In any case, to go from
523 MeV to 548 MeV is a significant 5% change, which doesn't look easy
to achieve incrementally.

As we talked about at the meeting, it really would be a good idea to
check if the top-bottom split in A180 was present with the original
calibrations. I don't think we saw a plot like this before. This check
could rule out calibration as the reason and the focus could be shifted
to other causes.

Another thing that we talked about was the top-bottom kinematics
distributions, which seemed to be mislabeled. It would also be very good
to see how those plots really look like, to see if the split might at
least partly be due to kinematics (different W regions for top vs bottom
at same x and Q^2). Adding a plot of W vs x would help.

Cheers,

Oscar


Whitney R. Armstrong wrote:
> Hello Everyone,
> 
> I wanted to give a brief update regarding the Bigcal re-calibration and
> summarize the conclusions from our last meeting.
> 
> First, I showed on Wednesday that the new calibration *significantly*
> impacts the asymmetry results. The new calibrations improve the overall
> behavior bin to bin and are consistent with A_parallel measured by CLAS.
> 
> Secondly, when looking at the asymmetries for the RCS and Protino
> separately
> (http://quarks.temple.edu/~whit/SANE/analysis_main/results/combined-split_asym_para59_251.png)
> 
> it was noted that between x ~ 0.3 to 0.4 there is clear difference
> between the top and bottom. Possible explanations include the
> calibrations, noisy/bad blocks, or background contamination (through
> differing kinematics). The latter is nearly ruled out because it
> persists through the background subtraction (as seen here
> http://quarks.temple.edu/~whit/SANE/analysis_main/results/asymmetries/A1A2_59_noel_split_253.png).
> 
> However, it is possibly a combination of a noisy block causing bad
> causing the average kinematics to be skewed and thus calculating the
> wrong background correction. For reference here is the background
> corrections plot
> http://quarks.temple.edu/~whit/SANE/analysis_main/results/asymmetries/bg_corrections_para59_251.png.
> 
> 
> Or it is the calibration...
> In order to provide a check, I have been looking at the eta mass peak
> (548 MeV/c^2) in the two photon mass spectrum. After spending two days
> iterationg over just the two sections of bigcal to improve the
> calibrations, I produce the following fits.
> 
> The mass peak in the perpendicular calibration looks good (m_0 ~ 550): 
> http://quarks.temple.edu/~whit/SANE/analysis_main/results/pi0calibration/FitEtaMass_perp_517.png
> 
> 
> For the parallel calibration, the mass peak is shifted to lower values
> by about 20-30 MeV (m_0 ~ 520). 
> http://quarks.temple.edu/~whit/SANE/analysis_main/results/pi0calibration/FitEtaMass_para_517.png
> 
> 
> Previously I had reported that the perpendicular calibration seemed to
> need more work. This was true when looking at the overall shift of the
> pi0 mass peak, but, when looking at the latest changes in the
> calibration coefficients for the parallel they are quite large (bottom
> right plot
> http://quarks.temple.edu/~whit/SANE/analysis_main/results/pi0calibration/pi0CalibrationCompare_para_517_c1_0.png).
> 
> Because the changes are large and fluctuate around zero a shift in the
> pi0mass peak is not observed (just a narrowing is), but these changes
> are just as important as the perpendicular.
> 
> I will continue to try to improve the calibrations by focusing on
> smaller sections (instead of all of RCS or Protvino at a time). My hope
> is that with improvement the parallel calibration's eta mass peak will
> move towards 550. It
> appears to be doing so rather slowly. Look at these plots in order:
> http://quarks.temple.edu/~whit/SANE/analysis_main/results/pi0calibration/FitEtaMass_para_514.png
> 
> http://quarks.temple.edu/~whit/SANE/analysis_main/results/pi0calibration/FitEtaMass_para_515.png
> 
> http://quarks.temple.edu/~whit/SANE/analysis_main/results/pi0calibration/FitEtaMass_para_516.png
> 
> http://quarks.temple.edu/~whit/SANE/analysis_main/results/pi0calibration/FitEtaMass_para_517.png
> 
> 
> Thank you for your time, and as always, your comments and questions are
> appreciated.
> 
> Cheers,
> Whit
> 
> 
> On Wed, Jun 04, 2014 at 03:29:54PM -0400, O. A. Rondon wrote:
>> Hi Whit,
>>
>> Neat results. I agree that for perp the eta might be there, but I would
>> suggest plotting errors on the histos (both 80 and 180 data) to see if
>> at 80 deg it's not just statistics, like the big jump near 300 (MeV)
>> probably is.
>>
>> Could you also add an explanation of what is being plotted in each
>> panel? The meaning of the mass plot is clear (at least of the blue
>> histo), and I guess the mass vs xi, yi (i= 1, 2) probably are the mass
>> distributions vs the x, y coordinates of clusters 1 and 2 used to
>> reconstruct the eta, but I'm not sure about the red mass histo in the
>> mass 3 panel; the E2 panel's red and black histos; what D is in the
>> massVSD; and what the clustx vs y1, y2 are (the y1,y2 scales here don't
>> agree with the scales of the mass vs y1, y2 plots).
>>
>> Before devoting too much effort in finding the eta in the perp data, I
>> would suggest recalculating A180 with the new calibration, to see if it
>> makes a difference in the asymmetry.
>>
>> Cheers,
>>
>> Oscar
>>
>> Whitney R. Armstrong wrote:
>>> Hello Everyone,
>>>
>>> I wanted to share some results in my refinement of the BigCal
>>> calibration from
>>> pi0s.
>>>
>>> Below you will find plots showing the mass spectrum (above the pi0
>>> mass) from
>>> two photons (top central plots) for parallel and perpendicular
>>> configurations.
>>>
>>> In the mass plot you can see a bump near the eta mass (547 MeV).
>>>
>>> Parallel:
>>> http://quarks.temple.edu/~whit/SANE/analysis_main/results/pi0calibration/etamass2-2000_1.png
>>>
>>>
>>> Perp:
>>> http://quarks.temple.edu/~whit/SANE/analysis_main/results/pi0calibration/etamass2-2000_5.png
>>>
>>>
>>> Please note the following:
>>>
>>> 1) Only after recalibrating did the eta mass peak start to show up.
>>> This is
>>> because the mass reconstructino of the eta is most sensitive to the
>>> energy
>>> resolution and less sensitive to the position (due to the larger
>>> opening angle
>>> for the eta compared to the pi0).
>>>
>>> 2) The parallel calibration is better than the perpendicular. That
>>> is, the perp
>>> calibration can certainly be improved compared to the parallel, thus
>>> the eta
>>> mass is resolved better for the parallel.
>>>
>>> 3) The background is much higher than the pi0, therefore, it does not
>>> seem
>>> feasible to directly calibrate (like pi0) using the eta mass peak. 
>>> However,
>>> because the two eta clusters are separated bny large distances (>90
>>> cm), it
>>> provides a nice confirmation that the opposite sides (eg the top and
>>> bottom)
>>> are calibrated properly against each other.
>>>
>>> 4) All previous results were contaminated by the insidious behaviour
>>> of the old
>>> calibration. The top (RCS) energies were shifted higher while the bottom
>>> (Protvino) was shifted lower (thus producing a nicely centered
>>> pi0mass when
>>> summing over the whole detector). This likely exlains why Anusha
>>> consistently
>>> saw energy differences in the elastics. It also explains (the problem
>>> seen long
>>> ago) that the top appeared to have all the higher energy events
>>> compared to the
>>> bottom.  It would be interesting to look at both of these problems
>>> again.
>>>
>>> Please let me know if you have any suggestions on how to clean up the
>>> eta mass
>>> spectrum.
>>>
>>> As always, your comments and questions are greatly appreciated.
>>>
>>> Cheers,
>>> Whit
>>>
>>>
>>> _______________________________________________
>>> Sane-analysis mailing list
>>> Sane-analysis at jlab.org
>>> https://mailman.jlab.org/mailman/listinfo/sane-analysis
>>>
>>
>>
>>
>> Also,
>>
>> _______________________________________________
>> Sane-analysis mailing list
>> Sane-analysis at jlab.org
>> https://mailman.jlab.org/mailman/listinfo/sane-analysis
>