[Halld-cal] Calorimetry work for PrimEx

Fri Sep 7 22:23:31 EDT 2018

Thanks Sasha -

 we all want all the GlueX programs to be successful -

Curtis

---------
Curtis A. Meyer MCS Associate Dean for Research
Phone:    (412) 268-2745 Professor of Physics
Cell:        (412) 260-6290 Carnegie Mellon University
Fax:         (412) 681-0648 Pittsburgh, PA 15213
cmeyer at cmu.edu<mailto:cmeyer at cmu.edu> https://urldefense.proofpoint.com/v2/url?u=http-3A__www.curtismeyer.com_&d=DwIFAg&c=lz9TcOasaINaaC3U7FbMev2lsutwpI4--09aP8Lu18s&r=VNruI5B3_Ie9AiOp0_GgpPscd-lfGLx1DNIPlMbQfwQ&m=2nPL76pU9_gjBXSKY86A7rw2gEgoX5PL2vsDCepHS0k&s=Nq5g9nqZ858kSiaA38STE8sfELAvzquxLwRsSmIOMjY&e= 

On Sep 7, 2018, at 21:58, Alexander Somov <somov at jlab.org<mailto:somov at jlab.org>> wrote:

Dear Curtis,

Though I've never seen you at the former L1 trigger meetings
(a joke), I agree, we'll need to set up a primex email list
and send out announcements for important discussions; so far,
we've been discussing a lot of technical things, related to
the CompCal construction.

Cheers,
      Alex

On Sat, 8 Sep 2018, Curtis Meyer wrote:

Dear Sasha

 the issues of the FCAL bases is well-known within GlueX (since fall of 2014), and is important input to any experimental program. Clearly this impacts PrimEx, but for us (broadly) to understand the issues, they need to be part f regular GlueX meetings. I try to participate in nearly all working group meetings on a regular basis and this is the first time that I have heard there are PrimEx meetings. We all want all of our GlueX programs to succeed, and I believe that we have built a structure than can make this happen, but we all need to work within this structure.In particular, meetings should have broad announcements to attract interested parties.

Curtis
---------
Curtis A. Meyer MCS Associate Dean for Research
Phone:    (412) 268-2745 Professor of Physics
Cell:        (412) 260-6290 Carnegie Mellon University
Fax:         (412) 681-0648 Pittsburgh, PA 15213
cmeyer at cmu.edu<mailto:cmeyer at cmu.edu><mailto:cmeyer at cmu.edu> https://urldefense.proofpoint.com/v2/url?u=http-3A__www.curtismeyer.com_&d=DwIFAg&c=lz9TcOasaINaaC3U7FbMev2lsutwpI4--09aP8Lu18s&r=zGSIfHWMfVP6ZGXTJMIIqw&m=qqOMTpBV9LaGrqib7RRW2uTPpO3BKx55AAVuCSL8VaA&s=5BbNGu_-Xjwc8ZqZ7Ny7sVYQILfEjfyqNOspZ5cLiGQ&e=

On Sep 7, 2018, at 21:07, Alexander Somov <somov at jlab.org<mailto:somov at jlab.org><mailto:somov at jlab.org>> wrote:

Dear Matt, All,

Thanks a lot for following this up.

As we have discussed, we need to understand whether the current
performance of the FCAL (with failing bases) and the level of
calibration is acceptable for precision measurements. We, indeed,
would like to participate in the FCAL calibration and related
activities if needed. Please provide us with a  list of specific
tasks (like generating MC sample), where we can contribute.

Regarding systematic uncertainties on the cross section measurement,
we are planning to organize a set of PrimEx related meetings after the
collaboration meeting, and discuss the detector performance and
measurements, which we'll need to do during the CCAL commissioning
run (currently, we have a weekly PrimEx meeting, where we mainly focus
on the construction of the CCAL, which will be completed in about
two weeks). I propose that we'll review the expected precision at
this stage. We need to get a broader audience involved in these
discussions.

Unfortunately the FCAL is not an ideal detector (as well as others),
so some MC corrections, which we'll need to apply in the cross
section measurement depend on it's performance. At the moment,
it's not clear to me to what extend we understand the FCAL performance;
it seems that a lot of work is still needed.

Cheers,
    Sasha

On Fri, 7 Sep 2018, Shepherd, Matthew wrote:

Dear Sasha, Liping, and Ilya,

Thanks for attending the calorimetry meeting yesterday.  It is important to move things along so we can be prepared for the PrimEx run.  We need members of the PrimEx team to work on specialized studies that are critical for these measurements.

Regarding Sasha's suggestion to establish a working group to discuss calorimeter calibration and other items in regard for preparation for the PrimeEx run, I believe the existing Calorimetry working group is the proper forum for this.  We have been meeting every two weeks for over ten years (all agendas and talks on the Wiki) to discuss construction, installation, calibration, monitoring, and reconstruction of the GlueX calorimeters, and have made status reports at every GlueX Collaboration meeting at least in the last decade.

If your measurement depends critically on the calorimeter, then this meeting and the halld-cal mailing list (cc'ed) is the right forum to have the discussion.

I wanted to raise a question regarding the plan for the measurement.  To make a cross section measurement you have:

sigma =  y / ( e n_t N_g )

y:  yield of events (statistically independent for Primakoff and Compton)
e:  acceptance * efficiency  (not common for Primakoff and Compton under any scenario)
n_t:  target density (common for Primakoff and Compton if the same target is used)
N_g:  number of photons on target (different for different run configurations, but likely to have correlated systematic errors for Primakoff and Compton)

Clearly sigma for Compton can be calculated.  You want to measure sigma for Primakoff.  I understand that for PrimeX everything on the RHS of the equation above was determined and one simply did the computation above for the cross section, and repeated for Compton as a cross check.

Can you please detail how you plan to constrain each of the items to achieve your desired level of precision in GlueX?  Clearly you can construct ratios to eliminate various systematics errors.  In the discussion yesterday there was talk of point-like targets to constrain acceptance measurements, but then when one uses that measurement at a different time with a different target some uncertainties may no longer cancel.  I went back and looked through Liping's ERR talks and couldn't find a clear answer.

Can you write the series of rations and measurements you plan to make and the anticipated precision on each?

As for the main action item:

I think we agreed the most pressing thing is to generate and analyze a pi0 MC sample as a route to calibrating the edge blocks.  I paste below an excerpt from an email message that I sent over two months ago to Ashot and Sasha.  It describes the status of edge calibration, which is what Sasha was asking about yesterday.  The status and needs are still exactly as described in this email.

Matt

Begin forwarded message:

From: Matthew Shepherd <mashephe at indiana.edu<mailto:mashephe at indiana.edu><mailto:mashephe at indiana.edu>>
Subject: Re: PrimEx issues / work day?
Date: June 28, 2018 at 12:10:21 PM EDT

What is needed:  someone to devise an algorithm for controlling the gain calibration at the edges of the detector.  My understanding is that PrimeEx is particularly concerned about the inner layers.  Here you must maintain control of the trigger efficiency for Compton vs. Primakoff events at the 1% level.  Just one misbehaving block will throw this off.

Our gain calibration procedure uses a sample of inclusive pi0 events.  We make pi0's matching a photon with most of its energy in one block with all other photons in the sample and then adjust the gain appropriately on the channel under study to put the pi0 peak at the right mass. Then iterate.

The problem with this technique is that it is unstable around the edges as one has leakage of the photon.  This leads to systematically higher gains in the edge blocks and a runaway condition when one iterates.  Therefore, we must fix gains around the edges so that our procedure is stable for most of FCAL, which is what we care about at leading order.  But you are likely going to be very sensitive to the inner edge.

We did use the LED monitoring system to try to "smooth" hardware gains by making HV adjustments.  There is a caveat here:  the LED system is not guaranteed, especially around edges, to generate a uniform quantity of light in every block.

We have discussed a better solution:  one needs to use MC to study leakage in blocks near the edge.  Then, when performing the gain adjustment on the block under study, one adjusts the gain to move the two photon peak to a modified value the pi0 mass (a function from the distance to the edge) to compensate for leakage.

*** This needs a person on it generate the MC and test the stability of the procedure on the edge blocks.  We can then examine if an HV adjustment to balance gains on the edges is possible.  Can PrimEx provide a person who can start working on this right away?  I think this may be the most useful exercise for PrimEx to do now to get familiar with the hardware and prepare for the run.

Matt, PLEASE, you probable still have an active person located at JLAB
and working on this part as an expert. We will need the name and contact
address to start working with him/her to possibly do this very important
stage of the experiment.

The FCAL expert on site is Colin Gleason (IU postdoc).  Colin will coordinate maintenance and we plan to service the detector at the beginning of August prior to fall run.  Colin can also help with iteration of gains as he has access to the skims and software to do it.

It would be useful for PrimEx to provide a person to do the MC study noted above at see if it provides a viable route to obtain gains and, consequently, do precise gain balancing of the edge layers.  (It has been on our to-do list but just hasn't risen to highest priority.)

_______________________________________________
Halld-cal mailing list
Halld-cal at jlab.org<mailto:Halld-cal at jlab.org><mailto:Halld-cal at jlab.org>
https://mailman.jlab.org/mailman/listinfo/halld-cal

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/halld-cal/attachments/20180908/d8145679/attachment-0001.html>