1) We should understand these higher rates (factor of 2 ?)
2) run the *any_coinc.trigger*, even at 1 sccm, otherwise there is no data
for all L-R combinations and for all Cerenkovs. The above trigger file has
been updated to the last delay values after retiming (check settings against logbook)
If you have problems and use a more restricted trigger then you have to update the
delays for the L/R top/bot CFD. I supsect that data of last night had wrong delays,
but can still be analyzed
The run plan calls for varying the C threshold. That means the threshold value for
LEdgediscr and REdgediscr on the trigger panel. Only the first four inputs
(per panel) are used. Tavi also knows how to do this and before he leaves
he could check the CC timing with flasher or beam (try to intercept him).
The CC delays for the two sectors are set to 32 and 2 ns, this asymmetry is (partly)
explained by the fact that the R-sector cerenkov signals go through an additional FIFO
that adds delay. HOWEVER it is not do or die since the CC are not required in the trigger
and at this moment the ADC and TDC is all thatwe need.
Just to be clearer: we need some more analysis on 3 things:
and if you have some results it is nice to have plots to discuss it
_ L/R asymmetries (jason) uses show_protons.C
_ adc spectra field on/off (ben, others ?) Here we need a commis/phase2/show_adc.C
that uses init.C so that you plot the ON, OFF spectra right on top of
each other, and normalized. It's a side project.
_ cerenkov efficiency, after gain matching.
For the latter we prefer the any_coinc trigger. Tavi should take care of
updating the TOF_1300.sql hv settings. Reload/check the db of course.
CC analysis is from the eep ntuple, which as you should know is created by
ntuple.C If you do that in the phase2 directory (and you should) the
ntuple will have pedestal subtracted ADC and a proton ADC cut is easier.
Please talk with the previous shifts, or keep buggin me on the phone !!
There is a macro cc_eff.C (in phase 2), chi worked on it yesterday. I
am checking that out to make sure. The cuts have been discussed many times,
and a sample was given in the runplan yesterday.
If yu plot the variable ncl with a cut such that
1) proton on L/R sector (ADC >7000 and rear TOF hit and no cer_right_sector)
2) electron on R/L sector (forward TOF hit, if you wish ADC<7000 but that
won't help rejecting pions)
you can see how many events fired a cerenkov or not (ncl=-1) If events go
in the wrong CC there is just too much junk or the cerenkovs are too hot
and fire on noise. Note, w/ runs taken thursday you usually never had the
"wrong" cerenkov firing.
Elastic events with field off are correlated according to the kinematics
in commis/misc/ep_angles_nik.txt by nikolas. That, and the old geom file
will guide you in the selection.
-- ________________________________________________________________________________ Tancredi Botto, phone: +1-617-253-9204 mobile: +1-978-490-4124 research scientist MIT/Bates, 21 Manning Av Middleton MA, 01949 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^---------- Forwarded message ---------- Date: Fri, 1 Nov 2002 18:44:21 -0500 (EST) From: Tancredi Botto <tancredi@lns.mit.edu> To: "[Blast_Shifts]" <Blast_shifts@rocko.mit.edu> Subject: [BLAST_SHIFTS] runplan weekend 11/1-11/3
Good luck !
Runplan 11-01-02 - tb =================
Priority is on the Cerenkov. Tavi is in charge and will start internal Cerenkov gain matching. He has highest priority Friday night and from Saturday at noon. He will work in the D-tunnel with flasher ON and the scope. In principle you can negotiate beam on when he is in the d-tunnel (since he'll use the flasher as trigger)
We have a first wrap of mu-metal shield around the TOF pmts.
Sunday evening the beam goes back to CCR for development of snake on/field on tune.
Friday - Eve =============
Tavi will work on flasher. After that take two cosmic runs with field on and off. Monitor and log the TOF rates and analyze ADC spectra. Make conclusions about effectivness of added shielding.
After that we need the power supply OFF. Remind CCR that, so they can turn off the magnet cooling (IMPORTANT)
Save the phase1/cc_eff.C and phase1/tof_eff.C to cc_eff_paddles (tof_eff_paddles). We need an automatic cc_eff.C for beam runs, starting from the eep ntuple (or TBlcut)
You should repeat the analysis of last night: e.g. for L0 I used the following cut (from eep)
(ntr<14&&ntr>12&&ncr==-1)&&(ntl==1||ntl==2) && (atr>4000)
and varied the right sector (proton) ADC cut. Note enforce no cerenkov on right sector. However, we should definitely add a TDC cut to select the center portion of the TOF, Cerenkov
We should then extend this code for other cerenkovs
Saturday - Night =================
Start eep elastic data taking with field off. Accumulate statistic for in depth L/R asymmetry study.
Analysis should use show_proton.C in phase2 (pedestal subtracted) or show_ptof.C in phase1 (not subtracted). That macro has the option of selecting individual TDC-diff cuts, rather than only using an ADC cut. Note, Cerenkov cuts may introduce asymmetries
Saturday Day ===============
Keep running. Tavi will finish gain matching with scope.
After CC gain matching (has priority) we have starting HVs for a cerenkov efficiency study. *Shoud the separation between pedestal and peak be not too clear, scale all voltages in that box by a fixed amount*. Be aware (and tune against) DC offset from the CC fifo.
Saturday Evening - Sunday Evening =================================
Here the goal is to take runs at fixed CC voltage and as a function of threshold (currently at a minimum 25 mV). After gain matching of course.
It takes < 1 hr running to analyize the front 2 cerenkovs. It takes longer for the other two. So it may be more efficient to scan the thresholds of the front for fixed threhsold in the back (so that you can re-analize the same runs and have higher statistic in the back)
I would not expect the cerenkov to be really 100 % efficient. Part of the "efficiency" may come from noise CC-triggers, especially if the given box counts well above 100 KHz. Raising threshold may then lower the "efficiency" but give a more realistic reading. Increasing threshold will bring over the 1 photoelectron peak and so on.
If N is the photo-electron, and the threshold is belowe the 1 PE peak, the CC eff is ( 1-exp(-N) )
if the threshold is above the 1 PE threshold the CC eff is
(1 - (1+ N)* exp(-N) ) and so on
Sunday Evening the beam goes to CCR ============
I am reachable tomorrow at home, unreachable on sunday -tancredi
-- ________________________________________________________________________________ Tancredi Botto, phone: +1-617-253-9204 mobile: +1-978-490-4124 research scientist MIT/Bates, 21 Manning Av Middleton MA, 01949 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:28 EST