Hi Emil,

I fixed the single event display, but I first have to install Java 1.2 on
vesna, which I will do on Monday. If you have access to a Java 1.2
installation already, you can pull off PBSed.jar from
stefan@vesna:pbsed/PBSed.jar and run it with

java -jar PBSed.jar runxxx.mid

otherwise wait until monday. I also added a feature that when displaying a
DSC channel and hitting the "next event" button, it looks for the next event
which has a hit in this particular channel. This way you can nicely inspect
the DSC data. I need this to make the DSC fitting work (probably also next
week).

Best regards,

        Stefan.

Hi Emil,

Bryan installed now Java 1.2.2 on most machines. I put the single even
display to dirac under the pibeta account as PBSed.jar, so you should be
able to run it from there with a recent data file like

java -jar PBSed.jar runxxx.dat

I had a quick look at the three body events and concluded that our low level
threshold is too low to really trigger on three particles. Mostly you get
some dirt (crosstalk from cosmics, backsplash) to trigger the third arm. I
could also be that the trigger is not workign correctly, but I'm not
convinced. We first would have to raise the low level threshold to let's say
20 MeV to see "real" particles in three segments.

Best regards,

        Stefan

Dear Emil and Weidong,

I realized that there was a bug in the SED since we upgraded the analyzer to
5 clups and still had 4 clumps in SED. I fixed that and updated PBSed.jar on
stefan@vesna and pibeta@dirac.

I quickly looked at the logarithmic cluster summing and found indeed that it
is much better than Thomas' old approach. I wonder why he has not realized
that some years ago.

By looking at the theta/phi map I found that CsI #0 is missing, probably
because it's HV has been turned off in the last manual gainmatch yesterday.
Since it has not be noticed in the detector checks (channel 0 is always at
the bottom), you (WL?) should add a theta/phi map for the cluster centers in
the detector check, then you see immediately if channel 0 (=P0) is missing.

Best regards,

        Stefan

Hi Emil,

th PBSed won't work with old data, since it assumes all 336 DSC channels.
You can maybe fix that problem by filtering out the CDSC bank with the
analyzer. With the new data it works, I copied a part of a recent file to
dirac and tried it.

Making the program run with old data (including DSC) would take some time
which I don't have right now.

Regards,

        Stefan

Dear Weidong,

I have attached the three necessary files: First the new JAR-File with
the possibility for a fixed scale in y and lines. The second one is the
actual dsc.c which I'm using now. The third file is waveform.dat where
the average waveforms are stored. In addition I have stored these files
on pibeta in folder tmp. The data file I have used is run35613.mid. The
waveform.dat file is for runs at about this one okay. Later I had to
exchange a whole module. This one has different domino speed for the
beamcounter signals and needs a different waveform.dat. 

First of all bring the dsc.c into your analyzer. urn the analyzer
shortly and the change odbedit. The bank switch DFIT has to be 1. Then
run the analyzer with e.g. -o sed.mid. Start java with the new jar file
and load the sed.mid file. Now under options you can set the range in Y
- both values identically uses a dynamic scale in Y.

You know how to use the trigger mask and I guess you know, that if you
type in the channel number in the field 'channel' it shows only events
where this channels has a DSC hit.

Some additional things: there is a connector rotation between DSC and
ADC/TDC - I don't know how that could be, but it is 48 -> 63, 49 ->
62,...

>From my point of view the fitting on CsI looks good to me - I did not
find any mis-fitting. On the individual target channels it does not look
to nice, on the beamcounter aceptable and on the degrader nice.

Please have a look also on the fitted waveforms and please discuus it in
Virginia. Then we should talk about how we will continue with that.

About your questions: My actual version of dsc.c hast the 'new' fitting
inside; there arenow different fitting fuctions: in a first step it
looks only for peaks and fits peaks and in a second step it fits with an
offset parameter. Another change is the treatment of the first 8 bins
where we might have a problem with the GATE crosstalk.

Your second question was due to the possibility of additional trigger
filters. This is not foreseen yet, but you can use the ANA_SKIP function
in the analyzer and start it with -f when you create your mid-file. I
think, you know how to use this ANA_SKIP?!

I'll hope this helps you and you are successfully with the installation. 

In addition I will put a file sed2.mid in tmp on pibeta which I have
created with the actual dsc.c on my machine. The you should see the
fitted waveforms.

Good luck and best regards   HP