[Olympus] SYMB Technical Paper
Frank Maas
maas at uni-mainz.de
Tue Nov 18 18:41:50 EST 2014
Dear Colton, Dmitry and Roberto,
thank you very much for this nice paper, congratulations to
you Colton, Dmitry, Roberto and Luigi for putting this together and discussing
this in detail. Somehow, only now, after some glass of a good italian red wine, I saw most of the things
which I could have seen earlier.….
I have a few questions/comments:
- Section 2, line 135: The „response“ time is not well defined,
how about "90% rise time“, that is more like a know technical term.
„response time" is to my mind something different including the
time the signal needs to go through crystal and PMT which is more like 30 - 50 ns.
Or how about "the whole
PbF2-PMT signal is contained within a 20ns time window“ (that is why we use a ~20ns
time window).
- Figure 4: The photo shows many things (lead brick etc.). I suggest
putting an arrow in the Photo to make it easier for a reader who does not know
what he sees to identify the collimator.
- Section 2.5, line 208: The overall dead time of the system is
smaller than 20ns due to the fact that a very fast FIFO buffer is used
to store the events before histogramming. After the FIFO,
the histogramming rate is up to 50 MHz.
The dead time is around 5ns.
- Fig. 8: The gain variation of a many stage PMT is (check Philips or Hamamatsu books
on Phototubes, text book) expected to change with some power of the
HV, i.e. (HV)**b, where b is larger than one. So it is a very surprising
observation that here it is linear.
In addition, this has been measured and used for
the phototubes we used in Olympus in the frame work of the A4 experiment. (See
Diploma and PhD thesis of Sebastian Baunack). We used this to calibrate 1022 phototubes online.
Last week, Luigi Dmitry and me, we discussed exactly this in our weekly jour fixe
and I have seen some other plots from other phototubes which show more clearly a nonlinear behavior.
Nonlinear behavior is the one which is expected.
Here the range ob HV variation is very limited (Delta U of about 35 V only). So it appears to be
linear most likely due to the small HV interval. But still this needs somehow to be addressed.
How about „… due to the small HV range used during calibration, we use here a linear fit…“
I’m not sure what your intended message of this plot is. If you want to show, that we have done a calibration,
you do not need to show a fit, and then I would rather show one of the other plots,
where one sees indications of the nonlinearity (which is text book expected). Anyway,
you can think about how to solve this, no need to follow my suggestions,
but any referee will be very surprised to see a PMT gain changing linearly with HV.
In addition, if I remember right, FIRST there was a optimisation of the HV by putting the
DESY test beam to the center of each crystal (Roberto and Dmitry please check whether I remember right,
I might be wrong here), then there was the additional HV-calibration by changing in a small interval
measured. So this is the reason why only a small HV interval was sufficient for later corrections of
transmittance losses. So I could see a reasoning:
a) DESY test beam: optimization of crystal HV, b) DESY test beam: additional HV calibration with small Delta U
c) due to small Delta U linear fit is sufficient.
- Fig. 8: The fluctuations of the points around the curve are much smaller than the
error bars. How did you determine the error bars? What is the chi^2 of the linear fit,
it should be extremely small (<< 1). In case you took the gaussian width sigma as an error,
the error is too big. it should be gaussian sigma / sqrt(N), N= events in the peak.
I’m not asking for tedious refitting or replotting, but there is an rather obvious
discrepancy (fluctuations around the fit much smaller than error bars) and that
needs to be resolved somehow. At least I think you could explain where the errors
come from and why you took those.
- Fig. 9: In chapter 2 you explained the DAQ for the SYMB involving an 8 bit ADC (max ADC value 255).
Here you show a continuous spectrum which goes up to ADC values of 2000. Of course
we used a different DAQ for the test beam at DESY with a single event read out.
I could not find this mentioned in the text and I think you should mention it.
- Fig. 9: There is a second bump/peak/"slight hill“ around channel 1600,
which is about twice the peak hight at around 800. You do not discuss it in the text.
It invites for referee questions. I would propose to add a sentence like
„The second smaller peak at ADC channel 1600 is due to pile up events."
If you mention the second, single event DAQ system used for these measurements,
it can explain, why we see pile up (unless there are really two electrons/positrons at the same time
which can be in test beam area 22).
- Fig. 9: The pedestal is not shown. I would feel much better if I have (for example as a referee)
to evaluate a spectrum, if I can see the measured zero. I know the pedestal peak is ver high,
one can suppress it or show logarithmic scale. anyway the minimum is to mention the
position of the pedestal (physical zero) in the caption. Or is this pedestal corrected?
If so, that too should be mentioned.
- Table 1: There is something wrong with the fit parameters and the error bars
of the column "right sector, e+“: the fit errors of the parameters a and b are somehow
very small (typo?). parameter c is zero (typo? or fit ran wild?), which can not be like that.
Shower fluctuations are purely due to geometry of the crystal matrix,
which should anyway be the same for electrons and positrons.
Same for electronics noise, since it is the same PMT, cables, ADCs etc.
for all channels. You could even imagine a common fit to two (in extremum to all four) curves in Fig. 11
where you have only one (common) parameter for the electronics noise a (this really
should be the same for all cases, at least for electrons an positrons it must be the same)
and even have only one (common) parameter for the shower fluctuations c.
- Table 1, Fig. 11 and Formula (1): There might be some double counting here:
you introduce parameter d, which is the energy/momentum spread of the beam.
In addition you have errors in „x“(energy)-direction in Fig 11. Did you use these error bars in the fit?
If yes: there is double error counting. If not: you should mention this. Also here, what is the
chi^2 of the fit? This would show (if around 1) whether we understand the energy-response of the detector.
- Figure 12: I would like to suggest to add an expression like „data taken with the OLYMPUS setup“ to make the contrast with
fig. 15 which is Simulations.
- Fig. 12: Furthermore, as you know, there is an deliberate offset in the ADC in order
to make more efficient use of the only 8 bit dynamic range of the ADC. This is not mentioned, but the spectrum
does not go to zero. This needs to be addressed somehow.
- Figure 12 a) and b) show two completely different spectra taken with the same type of
8 bit ADC and this is by intention.
b) has been taken deliberately so that one can see elastic scattering, which should not
be present in coincidence mode…
This is mentioned below the explanation of Fig. 13, so to my mind this comes too late.
Some explanatory words in the figure caption would help too.
- Please put into the caption of Fig. 12 also the measurement time for this spectrum so that
people can appreciate the high trigger/read-out rate capability of the system, which is
mandatory for measuring the luminosity under such small angles.
- Fig. 13: The explanation of Fig. 13 in the text (after lines 378) explains the Fig. 12, but the fact that
we see two lines in Fig. 13 is not explained here. One nice linear line is good,
two nice linear lines without explanation are bad. Is it positrons and electrons in the same plot?
if so you should say so (also in the caption). Or even put arrows in the figure to point out what is what.
- Fig. 15: figure caption says „SYMB digitization process“, but only histogram c) in fig. 15 adresses the digitization.
I would propose to replace „SYMB digitization“ by "SYMB Monte Carlo Simulation process“. It
would in addition make very clear that this is simulations as opposed to Fig 12. which is online (and offline) data.
- Fig. 14: No error bars at all or are they hidden under the symbols? If the latter, please say so in the figure caption.
How about errors in „x“-direction? Are the BPMs that accurate? If so please mention in the caption.
What I miss:
a) The photos are very nice and show a lot of details, one can other wise not grasp.
I like them very much.
But without some simple sketch (no complicated CAD-drowing but a simple sketch with rectangular boxes representing the
different items) showing the principal functionality of the SYMB,
like a beam, a target cell, a block on each side of the beam with the calorimeter
consisting of crystals, PMT etc. The photos are full of details and difficult to digest for somebody who does not know.
So I propose to add such a simple sketch-like figure, then the photos will be much more appreciated.
b) A plot where one sees the Moeller- and Bhabha/Annihilation-cross section together
with a short discussion would be extremely helpful. The physics processes on which the
SYMB relies is not very well explained. This includes a (short!) discussion of the
conincidence losses due to collimation etc. which have cost us many, many nerves in discussion.
(Not to speak of the „symmetry factor“… :-)) One figure to be added to section 1.2
and maybe 10 lines of additional text would be sufficient: Expected event rate, target luminosity, symmetric
Moeller angle (90° in e-CM-system,…) a little bit of narrative to explain why and how we have chosen
Moeller/Bhabha scattering for this.
best regards
Frank
> Am 18.11.2014 um 22:09 schrieb Colton David O'Connor <colton at mit.edu>:
>
> My apologies if MIT's webmail system caused a problem with the attachment. If the paper was not attached to my first email, you can retrieve it here:
>
> web.mit.edu/colton/www/SYMB.pdf
>
> -Colton
> ________________________________________
> From: olympus-bounces at mit.edu [olympus-bounces at mit.edu] on behalf of Colton David O'Connor [colton at mit.edu]
> Sent: Tuesday, November 18, 2014 4:04 PM
> To: olympus at mit.edu
> Subject: [Olympus] SYMB Technical Paper
>
> Dear Collaborators,
>
> On behalf of Roberto, Dmitry, and the other authors, I present to you the paper we have prepared describing the OLYMPUS SYMB detector, intended for submission to NIM A. We have been through several drafts and revisions over the past few months. We now seek your approval for the contents of the paper to be made public.
>
> In particular, we are interested in comments about:
>
> (1) any data, plots, or descriptions in the paper that are not fit for release, such as broader physics results;
> (2) significant errors in our descriptions; and
> (3) important omissions, or large sections that ought to be omitted.
>
> Any comments you have on these or other topics will be welcomed, and we will not submit the paper until it is blessed by the collaboration, but any revisions actually implemented will be at the sole discretion of the authors.
>
> I suggest that anyone who would like to discuss the paper with us should join this Friday's luminosity meeting on SeeVogh if they are able (16:00 Hamburg time, 10:00 Boston time). If there is a need, we can plan for an additional meeting at another time, but my hope is that any comments will be offered promptly so that we can proceed toward submission sometime next week.
>
> If you wish to email comments, you may send them to me directly (colton at mit.edu), or you may broadcast them as a reply-all to this email if you wish for them to be visible to the whole collaboration, but if there are many comments then it will be good to discuss them over SeeVogh instead.
>
> Thanks,
>
> -Colton
> -------------- next part --------------
> A non-text attachment was scrubbed...
> Name: SYMB.pdf
> Type: application/unknown
> Size: 9520915 bytes
> Desc: SYMB.pdf
> Url : http://mailman.mit.edu/pipermail/olympus/attachments/20141118/bee5dc83/attachment.bin
> _______________________________________________
> Olympus mailing list
> Olympus at mit.edu
> http://mailman.mit.edu/mailman/listinfo/olympus
>
> _______________________________________________
> Olympus mailing list
> Olympus at mit.edu
> http://mailman.mit.edu/mailman/listinfo/olympus
--------------------------------------------------------------------------------------------
Prof. Dr. Frank Maas - GSI Darmstadt / JGU Mainz
Director, Helmholtz Institute Mainz
Hadronphysics II
eMail f.maas at gsi.de
Tel.: +49 6131 392 7447 (Mz)
Tel.: +49 6159 711601 (GSI)
Fax. +49 6131 392 7448 Mobile +49 174 3281502
GSI Helmholtzzentrum für Schwerionenforschung GmbH
Planckstraße 1
64291 Darmstadt
www.gsi.de
Gesellschaft mit beschränkter Haftung
Sitz der Gesellschaft: Darmstadt
Handelsregister: Amtsgericht Darmstadt, HRB 1528
Geschäftsführung: Professor Dr. Dr. h.c. mult. Horst Stöcker,
Ursula Weyrich
Vorsitzenderdes Aufsichtsrates: St Dr. Georg Schütte
Stellvertreter: Ministerialdirigent Dr. Rolf Bernhardt
More information about the Olympus
mailing list