Scattering analysis update.
Friday 18th June 14:00 to 17:00. We will use the regular mice zoom details:
[[computing-software:MICE_zoom_info|MICE Zoom Connection]] (login required)
"Copy of v0.2 of the paper" http://micewww.pp.rl.ac.uk/attachments/11471/MICE-MCS-LiH-paper-v0.2.pdf
"Copy of v0.3 of the paper" http://micewww.pp.rl.ac.uk/attachments/11472/MICE-MCS-LiH-paper-v0.3.pdf
- Update on the analysis (John N)
- Latest version of the paper (Paul)
18/06/2021 Referees meeting
John Nugent gave us a few updates on his status. He has collated his changes and
put things together into the MICE note.
John N commented on the chi2 calculation including correlated errors and explained how he calculated chi2. We agreed that the procedure is roughly correct, and that stat error is a reasonable thing to vary in the input bin (x_j) distribution when calculating the partial. There was a question about equation (2). We suggested that the error in each bin should be approximately the sqrt(n) in each bin. Chris proposed sending through the A matrix (or histo) - it should be nearly diagonal. John C proposed another routine would be to throw a gaussian error on the number of events in each bin and repeat the deconvolution. Look at the variation in the output.
We moved on to discuss the paper. John C commented that things were a lot cleaner than than they had be and the whole thing has been generally "shaken down".
Title and abstract have been modified. A lot of formulae and discussion on MCS have been removed. A couple of formulae were kept for self-consistency/use later on. ACTION: John Cobb commented that MICE should be in past tense.
Date was removed from schematic. Coordinate system was described. ACTION remove the sentence "A change in the number of radiation lengths... less than 0.3 %".
Detail of the angle definition maths was moved to the appendix. We discussed relevance of eq. 6 and felt it was useful. The description of theta_x and theta_y was slightly wrong due to the fact incident track is not axial. ACTION: correct the description.
ACTION: In table 1, John N said that the addition of the radiation lengths should be done as 1/x = 1/x_1+1/x_2+...
We discussed the relevance of the Table 1 (1 tracker, 2 trackers, 1 station should be listed?). We made no recommendation - status quo wins.
ACTION: indicate that 3 mm emittance is "nominal" in line 71.
ACTION: add a sentence indicating the angular divergence of the incident beam - it is a table in the MICE note.
ACTION: clarify the 172, 200, 240 MeV/c refer to bin momentum, not overall beam momentum. The beamline settings are merged and then binned in TOF.
We discussed the effect of merging beams having different momenta. We considered whether there are systematics. This was looked at in the early days and no effect was found.
ACTION: remove reference to MAUS conference paper.
Paul showed Table 2. ACTION fix title on 3rd column. Noted that the sentence on line 103 doesn't read well. Mariyan noted the electron peak. Paul pointed out that it is in the text. ACTION: John C noted that the electron peak is listed at 25.8 ns. ACTION: mention positron peak in the caption.
ACTION: John N - check what the fit for the electron peak shows us
ACTION: Fig.3 don't join the dots.
Table 3 - momentum resolution was added, to compare to the s.d. John C pointed out that the mean of the tof is not quite the same as the nominal momentum. Could either change the labelling; highlight labelling is nominal and measurement is different; or rerun the jobs. ACTION: Paul Soler sighed and groaned and we agreed to rerun the jobs.
ACTION: Ken points out branding on Fig. 3 and 4 needs to be added
ACTION: Agreed to remove Fig. 4 from the paper
We discussed fig. 4. We felt it was "dirty washing". Better to emphasise the systematic uncertainty on the alignment. ACTION: remove the figure. Remove correction factors from the wording.
We discussed fig. 5. There was concern that the different bins were confusing. ACTION: We agreed to highlight analysis bins in fig. 2, potentially with a different colour. However, the 172, 200, 240 bins are different. There was concern poor reader might be confused. ACTION: Paul S agreed to try to finesse things to make it clearer.
We looked at fig. 8. John N clarified that the distributions are for the samples not the beams. ACTION: Paul S to correct.
We wondered whether it was better to move table 3 and fig. 8 closer together. ACTION.
We discussed what plots to show for raw distributions MC vs data. ACTION: Looking at the plots, thought 3D scattering angle is not necessary. Include both absorber and no absorber. Remove fig. 10. Remove fig. 9 c and d. Discussed replacing with upstream x' and y' distribution. Again, include in the text comment saying RMS width of x' and y' (this is an action further up).
John Cobb proposed plotting theta_scatt squared. Linear core and tail. ACTION: Have a look at them.
ACTION: extra bin in the scattering plots (fig 14?) in the positive direction.
ACTION: need to check that the fig. 16 is really deconvolution. JN to advise PS. Agreed to leave it in for the next version and discuss next time.
ACTION: consider to remove table 6 or remove the chi2 from table 6 if we can't convince ourselves about the correlated errors.
Paul suggested that any further large changes should be agreed. He will sort that out. Then word smithing could be done later.
Agree: Paul S will produce a new version before Wednesday. Referees to read. Comments on wording/etc in writing. Then decide if/when we need to do a follow up meeting.
John Cobb suggested that it looks in pretty good shape.