[Sane-analysis] More on BigCal angles

[O. A. Rondon]

Hi,

To figure out the reason for the difference between the event angles
calculated with the Integral B.dl and the coordinates to angle
conversion fit, I've run the standalone tracking code, which should be
exact relative to the field map.

I updated the table and spreadsheet on the wiki. The results are shown
on the bottom four rows of the table. As before, th_lab and phi_lab are
the detected angles.
https://hallcweb.jlab.org/experiments/sane/wiki/index.php/Simple_Correction_for_the_Target_Field_Deflection#Comparison_between_Integral_B.dl_and_track_fitting

I see two issues:

- the event angles calculated by the fit don't agree with the thrown
angles. In particular, theta is off by a significant amount, about 1.3
degrees for 1 GeV particles.

- theta calculated by the fit is not independent of field angle, as it
must. The disagreement is worse for parallel field.

Things improve with increasing E', but even at 1.7 GeV the theta angle
for 180 deg. field is half a degree off.

The numbers from the magnet tracking code are in new tabs in the
spreadsheet.

It seems to me that the current fit cannot be used at least for E'<~ 1.3
GeV, and even at higher E' it introduces a needless systematic shift.

We need to decide whether we can proceed with the analysis using this
method of calculating the angles. The angles calculated with the  2012
and 2013 versions of the fit parameters agree with each other well, so
it does not seem a matter of just redoing the fits.

On the other hand, it might be that I'm not implementing the fit's
coordinate conversion correctly, so an independent check should be done.

Cheers,

Oscar


O. A. Rondon wrote:
> Hi,
> 
> I have updated my wiki section on the method to calculate the true event
> theta and phi, based on a simple fit to the target field's Integral B.dl.
> 
> https://hallcweb.jlab.org/experiments/sane/wiki/index.php/Simple_Correction_for_the_Target_Field_Deflection#Comparison_between_Integral_B.dl_and_track_fitting
> 
> The update is a new section, comparing the angles calculated by the
> Integral B.dl method and the angles calculated by the track fitting
> method currently being used in the analysis. The agreement is generally
> at the 10% or better for phi, and 5% for theta. It's not as great as I
> would have expected, but it could be due to the depth correction that
> shifts the apparent cluster coordinates, which also is more important
> away from the center, where agreement is quite good (better than 1%,
> except for phi at 180 deg, which is strange).
> 
> Feel free to play with the spreadsheet.
> 
> Cheers,
> 
> Oscar
> 
> 
>