[Halld-cal] [EXTERNAL] Re: CPP/NPP BCAL calibration status
Igal Jaegle
ijaegle at jlab.org
Mon Jun 26 20:43:40 EDT 2023
All,
After a week's break, I am coming back to this task.
In short, the status is the following, the 3rd layer, cannot be calibrated with and without the trigger bit used (1 / FCAL+BCAL or 3 / BCAL alone).
Iteration 20 of period 3 (https://halldweb.jlab.org/primexd/bcal-cal/2022-05/3/) shows the results when using the trigger bits 1/3. And the individual cell gg inv. mass for each layers are aslo there *.py files.
It seems (for this last iteration) layers 1 and 2 did converge but the background level remains the same as before ...
Any new inputs?
tks ig.
________________________________
From: Igal Jaegle <ijaegle at jlab.org>
Sent: Thursday, June 15, 2023 7:33 PM
To: Sean Dobbs <sdobbs at jlab.org>
Cc: soy062 <soy062 at uregina.ca>; halld-cal at jlab.org <halld-cal at jlab.org>; Rory Miskimen <miskimen at physics.umass.edu>; Karthik Suresh <ksv656 at uregina.ca>; Benedikt Zihlmann <zihlmann at jlab.org>
Subject: Re: [Halld-cal] [EXTERNAL] Re: CPP/NPP BCAL calibration status
Thank you, Sean, I will let you know if it is improving things.
tks ig.
________________________________
From: Sean Dobbs <sdobbs at jlab.org>
Sent: Thursday, June 15, 2023 12:06 PM
To: Igal Jaegle <ijaegle at jlab.org>
Cc: soy062 <soy062 at uregina.ca>; halld-cal at jlab.org <halld-cal at jlab.org>; Rory Miskimen <miskimen at physics.umass.edu>; Karthik Suresh <ksv656 at uregina.ca>; Benedikt Zihlmann <zihlmann at jlab.org>
Subject: Re: [Halld-cal] [EXTERNAL] Re: CPP/NPP BCAL calibration status
Hi Igal,
As a follow-up, currently there is ~10% of the CPP run skimmed out
including the trigger data in:
/cache/halld/RunPeriod-2022-05/calib/ver02/BCAL_pi0
So you should be able to look and see if you are getting a better pi0
peak by selecting e.g. the BCAL-FCAL trigger, and how that affects the
total statistics of the measurement.
Cheers,
Sean
On Wed, Jun 14, 2023 at 10:31 AM Sean Dobbs <sdobbs at jlab.org> wrote:
>
> Igal,
>
> For what it's worth, I started skimming the data to include the
> trigger information. The whole data set will probably take a couple
> of weeks, but I'll let you know when there's enough data to get some
> feeling for how the signal/background might improve - this will
> probably be a couple of days.
>
> Cheers,
> Sean
>
> On Wed, Jun 14, 2023 at 3:08 AM Igal Jaegle via Halld-cal
> <halld-cal at jlab.org> wrote:
> >
> > All,
> >
> > I made a test for period 3, https://halldweb.jlab.org/primexd/bcal-cal/2022-05/3/pi0gainmatrix_11.pdf
> >
> > I started with period 1 table. As you can see between iteration 0 and 4, there is no changes. Then, I put the ADC gain for all L3 cells to the default value (3.5e-05 - after each iteration). And between iteration 5 to 11, the behavior starts to converge but not as well as when a pi0 signal is seen in L3. I will try to use for all L3 cells, the average gain instead of the default value for iteration 12 and above. Let me know if you have a better idea.
> >
> > tks ig.
> > ________________________________
> > From: soy062 <soy062 at uregina.ca>
> > Sent: Monday, June 12, 2023 1:38 PM
> > To: Igal Jaegle <ijaegle at jlab.org>
> > Cc: Karthik Suresh <ksv656 at uregina.ca>; Mark Macrae Dalton <dalton at jlab.org>; Rory Miskimen <miskimen at physics.umass.edu>; Zisis Papandreou <zisis at uregina.ca>
> > Subject: Re: [EXTERNAL] Re: CPP/NPP BCAL calibration status
> >
> > Hello Igal,
> >
> > I looked at the layer 3 plots, and indeed, they don't seem to be
> > converging. Even the fits in the 2y reconstructed mass look very weird
> > and don't seem to improve per iteration. This is probably very closely
> > related to the aforementioned issues with Layer 3. Usually I've never
> > seen this behaviour for the background and total fits, as things looked
> > rather clean for other experiments (see attached Be target from PrimEx).
> >
> > Also, looking at the coefficient of variation vs. iteration plot, there
> > is a big drop in the last iteration, which doesn't really reflect in the
> > other plots (it's essentially zero), so I'm not sure why that happened.
> > Currently, I'm not sure what advice to give for layer 3, since it seems
> > to be a complexity related to the CPP/NPP experiment conditions. Maybe
> > someone can give additional comments on this.
> >
> > Best regards,
> > Stjepan
> >
> > On 2023-06-10 10:47, Igal Jaegle wrote:
> > > Thank you, Karthik, for your input. Ok, the 12th iteration finished
> > > this morning, layer 1 and 2 did converge although their bands seems
> > > larger than what you are usually showing. The 3rd layer is all over
> > > the place because no pi0 resonance is seen, look at
> > > l3_gain_checking_1.py. What is your advice in this case?
> > >
> > > tks ig.
> > >
> > > -------------------------
> > >
> > > From: Karthik Suresh <ksv656 at uregina.ca>
> > > Sent: Friday, June 9, 2023 11:45 PM
> > > To: Igal Jaegle <ijaegle at jlab.org>
> > > Cc: soy062 <soy062 at uregina.ca>; Mark Macrae Dalton <dalton at jlab.org>;
> > > Rory Miskimen <miskimen at physics.umass.edu>; Zisis Papandreou
> > > <zisis at uregina.ca>
> > > Subject: Re: [EXTERNAL] Re: CPP/NPP BCAL calibration status
> > >
> > > Good evening Igal,
> > >
> > > The current procedure is such that the convergence could be slow (due
> > > the the iterative nature). Also, I doubt if the third layer can be
> > > calibrated within 1% precision with this method. But I think we could
> > > conclude that after a few more iterations.
> > >
> > > Thank you
> > > Karthik Suresh
> > >
> > > On 2023-06-09 21:26, Igal Jaegle wrote:
> > >> Thank you, Stjepan, for your answer and help. It seems very slowly
> > >> converging, see the latest report 8. The first values were not so
> > > bad,
> > >> then I did a mistake in the new gain calculation, correct it, then
> > > it
> > >> seems to slowly recover but very slowly. Could it be due to the 3rd
> > >> layer seeing no pi0s? What should be done to the 3rd layer, put
> > >> default values or the values of the run period precede it or succeed
> > >> it? What is your suggestion for the 3rd layer?
> > >>
> > >> tks ig.
> > >>
> > >> -------------------------
> > >>
> > >> From: soy062 <soy062 at uregina.ca>
> > >> Sent: Friday, June 9, 2023 10:56 PM
> > >> To: Igal Jaegle <ijaegle at jlab.org>
> > >> Cc: Mark Macrae Dalton <dalton at jlab.org>; Rory Miskimen
> > >> <miskimen at physics.umass.edu>; Zisis Papandreou <zisis at uregina.ca>;
> > >> Karthik Suresh <ksv656 at uregina.ca>
> > >> Subject: [EXTERNAL] Re: CPP/NPP BCAL calibration status
> > >>
> > >> Hi Igal,
> > >>
> > >> Sorry for the delayed response. The histograms in the "fit" ROOT
> > > file
> > >> are arranged in a cyclical way. For example, starting from 1 to 4
> > >> (channel number) would be layer 1, 5 to 8 would be layer 2, 9 to 12
> > >> would be layer 3, 13 to 16 would be layer 4 and then for 17 to 20 we
> > >> go
> > >> back to layer 1. The final one is 763.
> > >>
> > >> Looking at the plots (pi0gainmatrix.pdf), there seems to be
> > > something
> > >> missing in the pipeline. Basically for each iteration the pi0 mass
> > > is
> > >> not converging, which means that somewhere along the way the local
> > >> ccdb
> > >> gain constants are not getting updated after each iteration or the
> > >> local
> > >> ccdb is not getting copied to the right location. This can be seen
> > > for
> > >>
> > >> example in the sigma/mean vs iteration plot, it is constant, which
> > > it
> > >> should not be.
> > >>
> > >> Best regards,
> > >> Stjepan
> > >>
> > >> On 2023-06-09 16:45, Igal Jaegle wrote:
> > >>> Bump!
> > >>>
> > >>> -------------------------
> > >>>
> > >>> From: Igal Jaegle
> > >>> Sent: Thursday, June 8, 2023 9:20 PM
> > >>> To: Mark Macrae Dalton <dalton at jlab.org>; Rory Miskimen
> > >>> <miskimen at physics.umass.edu>; Zisis Papandreou <zisis at uregina.ca>;
> > >>> Karthik Suresh <ksv656 at uregina.ca>; Stjepan Orešić
> > >>> <soy062 at uregina.ca>
> > >>> Subject: CPP/NPP BCAL calibration status
> > >>>
> > >>> All,
> > >>>
> > >>> I am testing the upgrades of the BCAL gain calibration procedure.
> > >> The
> > >>> upgrades free the calibrator of all interventions between each
> > >>> iteration making the calibration fully automized. I made a wiki
> > >> which
> > >>> can be accessed from
> > >>> https://halldweb.jlab.org/wiki/index.php/Main_Page [1] go to,
> > > Barrel
> > >>> Carolimter, pi0 gain calibration, and CPP/NPP (2022-05) BCAL pi0
> > >>> calibration [2]
> > >>>
> > >>> The report and individual fit can be found here:
> > >>> https://halldweb.jlab.org/primexd/bcal-cal/2022-05/1/ [3]
> > >>>
> > >>> Stjepan/Karthik, how the layer ordering is working? 1 to 512 first
> > >>> layer, 513 to 1024 2nd layer, and 1025 to 1534 3rd layer? The 3rd
> > >> sees
> > >>> nothing is that an issue? How to solve it? Or leave with it?
> > >>>
> > >>> tks ig
> > >>>
> > >>>
> > >>>
> > >>> Links:
> > >>> ------
> > >>> [1] https://halldweb.jlab.org/wiki/index.php/Main_Page
> > >>> [2]
> > >>>
> > >>
> > > https://halldweb.jlab.org/wiki/index.php/CPP/NPP_(2022-05)_BCAL_pi0_calibration
> > > [1]
> > >> [1]
> > >>> [3] https://halldweb.jlab.org/primexd/bcal-cal/2022-05/1/
> > >>
> > >>
> > >> Links:
> > >> ------
> > >> [1]
> > >>
> > > https://halldweb.jlab.org/wiki/index.php/CPP/NPP_(2022-05)_BCAL_pi0_calibration
> > > [1]
> > >
> > > --
> > > Karthik Suresh
> > >
> > >
> > > Links:
> > > ------
> > > [1]
> > > https://halldweb.jlab.org/wiki/index.php/CPP/NPP_(2022-05)_BCAL_pi0_calibration
> > _______________________________________________
> > Halld-cal mailing list
> > Halld-cal at jlab.org
> > https://urldefense.com/v3/__https://mailman.jlab.org/mailman/listinfo/halld-cal__;!!PhOWcWs!wYZI7-yOAXDQlAlPpt0G3TedPB6zwn2MIOjbMWzG5i0OGwZK1kL-sMDmTYC8wJn7lESuYRotVEIyIlK4$
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/halld-cal/attachments/20230627/14861e76/attachment-0001.html>
More information about the Halld-cal
mailing list