[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Thoughts on motivating tof for Review
Hi Elke,
See some comments below on the start time and dE/dx.
I was not trying to replace any discussion/analysis by Eugene on overall
pid. Just to think about how the tof would fall into that discussion.
On Mon, 25 Feb 2008, Elke-Caroline Aschenauer wrote:
> On Fri, 22 Feb 2008, Elton Smith wrote:
>
> Elton,
>
> did you discuss this on Friday?
We did not discuss this at the meeting. I was just trying to start an
outline/motivation for the tof presentation.
>
> Okay, I think before we can talk about PID we need to know the performance
> of the detectors. So I would think we need to make sure we know the timing
> performance of the BCal and the latest mails on this indicate we need a
> bit of work on this, but are already very close.
> For the forward tof we need the hardware performance, the uncertainty on
> the momentum and path length. If we have this we can look into the
> performance.
>
> I don't need a MC to know we basically have in the full momentum angle
> range a factor 1 to 10 for kaons to pions, so we will have a hard time to
> get kaon id. so all comes to pion to proton separation. Actually there
> will be also protons in the forward direction.
> If you talk about pid, than we might also consider some wider physics
> program, like we want to discuss in a week and than we need also electron
> id, which of course is not to difficult to get with our em-calorimeters,
> e/p gives normally a factor 100 suppression of hadrons in the lepton
> sample. Which might be okay, not great but ...
>
> As I was saying I asked Eugene to repeat his studies he did some time ago
> for the Cerenkov more focusing on the tof in the BCal and the forward.
> I think pythia gives a very nice possibility to do this, we could for
> example look to sigmas or lambda(1520) reconstruction if we want to make
> statements on kaons.
>
> For getting a good particle id you actually best measure parent
> distributions of your final detector with beam and than use the Bayes
> theorem to get probabilities. I showed this in my talk on the Hermes pid
> in the collab meeting last October.
>
> for the trigger and your timing questions. we should not to much focus on
> 10^7, because as I said in my other email the start counter will have a
> rough time at 10^8 and so will the tagger, so we should from the beginning
> have a start time which will work for all rates 10^7 and 10^8. So again
> what is wrong with the machine clock.
>
Werner has estimated the rates in the start counter and my understanding
is that the segmentation is design for operation up to 10^8 (based on
count rate alone). Obviously how effective it is in the trigger or to
determine the start time needs to be studied.
The determination of the start time at 10^7 is much easier and solved by
having a start counter. The machine clock is used in all scenarios to
determine a precise interaction time (modulo 2 ns), but one needs a
detector to select which beam bucket corresponds to the interaction. I do
not think a single solution will work for 10^8, so the argument to start
at low rates and develop algorithms as we increase the intensity makes
sense (empirically we can argue that a similar process has worked in Hall
B). For example, if only the barrel and forward tof are used, then events
where none of the charged particles reach the outer detectors will require
special attention, as well as events with detached vertices (which shifts
the time from the interaction downstream). Depending on how low in angle
coverage for the start counter, there can also be a loss of charged
particles down the beam aperture.
> For the trigger, I'm not sure how much we have to discuss in detail the
> trigger, because there we look for hits. I think the new material budget
> in the beam brings the rates down by a factor of 5 ( I have not all
> numbers here) so this will give us a certain hit rate in 100ns per paddle.
> We could make the inner paddles a factor of 2 smaller in width, this is
> not at all a cost problem. Actually one test would be to make the air in
> the beam pipe He and see what that helps in addition.
>
> I think in the barrel dE/dx has to be included in the picture, of course
> the threshold for protons coming out of the target is an issue, because it
> is close to what the range of protons is which can be detected by the CDC.
> This is something we have to look into very carefully, because the start
> counter contributes to this threshold for sure quite a bit.
This clearly needs to be checked quantitatively. However, the validity of
dE/dx does not depend on the momentum of the proton at the vertex, but
rather the momenum of the proton in the CDC (e.g. after it has lost energy
in the start counter). So the range of momenta where the dE/dx is valid is
simply shifted with the start counter.
As a side note, perhaps worth emphasizing, is that at high rates the
functions of the start counter will likely be taken over by more
sophisticated analysis of the whole detector. In this situation, one
should be able to remove the start counter from the detector and reduce
thresholds for detection of slow particles (i.e. protons). But when to do
this will depend on the experience gained on running the detector and
sophistication in both trigger algorithms and analysis.
>
> If the final pid includes a aerogel rich, which with the improvements in
> aerogel lello presented is the best thing to do, we will have actually
> pion id from 0.5 geV on, which is very nice, because having 2 pid
> detectors overlapping will increase your purity a lot. that is also true
> combining the rich and the em-calos for leptons.
>
> One other thing we have to make sure that the beam hole in the tof is
> matched to the one of the FDC.
>
> so this has become a long email, but as I was not present at the
> discussion, I thought I type some of my ideas into an email.
> I guess we will have a detailed discussion on Friday. I can be available
> for a phone meeting earlier if needed.
>
> cheers elke
>
>
> > Friends,
> >
> > I spent some time with Matt and Beni to go over the motivation/arguments
> > for tof in the detector, given that kaon id is very limited. I put these
> > out for comments/discussion especially in order to define what
> > calculations/work might need to be done in preparation for the review. We
> > came up with the following points
> >
> > - pion purity (how often are we fooled by other topologies?)
> > - positive proton id (mostly in the barrel)
> > - forward tof input to trigger
> > - timing information for analysis (start time for drift times, rejection
> > of accidentals coincidences with the tagger)
> > - part of future comprehensive pid detector package for detector (covering
> > pi/K id below 2 GeV).
> >
> > I would appreciate comments suggestions on what to emphasize and priority
> > for any studies that are required.
> >
> >
> > Questions:
> > - What is the granularity (segmetation) required due to rate
> > considerations?
> > - What is the spectra of all particle species (pi,K,p) over the detector
> > (e.g. what is the ratio of particle types in the Bcal and the TOF?)
> > - Study acceptance of charge particles (actual trajectories in B field)
> > for the 1deg beam hole and at the largest angles?
> >
> >
> >
> >
> >
> >
> > Elton Smith
> > Jefferson Lab MS 12H5
> > 12000 Jefferson Ave
> > Suite # 16
> > Newport News, VA 23606
> > elton@jlab.org
> > (757) 269-7625
> > (757) 269-6331 fax
> >
>
> ( `,_' )+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=
> ) `\ -
> / '. | +
> | `, Elke-Caroline Aschenauer =
> \,_ `-/ -
> ,&&&&&V Jefferson Lab +
> ,&&&&&&&&: HALL-D 12C / F381 121-A Atlantic Avenue =
> ,&&&&&&&&&&; Suite 8 Hampton, VA 23664 -
> | |&&&&&&&;\ 12000 Jefferson Ave +
> | | :_) _ Newport News, VA 23606 Tel.: 001-757-224-1216 =
> | | ;--' | Mail: elke@jlab.org Mobil: 001-757-256-5224 -
> '--' `-.--. | +
> \_ | |---' Tel.: 001-757-269-5352 =
> `-._\__/ Fax.: 001-757-269-6331 -
> +=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+
>