Subject: LIP analysis intercepting ERP one
From: Erik Katsavounidis (Erik.Katsavounidis@lngs.infn.it)
Date: Wed Jan 26 2000 - 11:31:14 EST
Ciao a tutti,
While I was looking at finer details of the LIP analysis I pursued
lately, I noticed a rather strange global ERP inefficiency at the
few % level that fluctuates *significantly* as a fuction of time.
This is mostly in SM2 but I have seen it in other SM as well.
I've checked all RUNs since RUN#16000 and have a summarizing
plot that shows the effect. This is rather news to me and I'm
bringing it to your attention as I think it is relevant to efficiencies
that you have been calculating. This is something that nominally should
be folded already to your efficiency calculations. I have been basically
hand scanning a fair number of events in order to form a solid picture
of what has been going on in the detector during this class of events.
A global ERP inefficiency has a direct result in my ERP-based
pre-analysis of LIP events. I have observed entire group of RUNs
where the average number of LIP candidate events was an order
higher respect to other RUNs (~10 vs 2), most of them coming
from SM2. In the final 13 months dataset of my LIP analysis, more
than 60% of the surving events come from certain period of the SM2
running. The events are fully analyzable through the PHRASE and or
the WFDs. The class of events I am talking about DOES have an ERP
muon trigger (in most of the cases) but the ERP boxes have nothing
to do with what e.g. LIP, CSPAM and STREAMER TUBES have recorded,
as in the example that follows:
>>> RUN: 17000 EVENT: 2959 LENGTH: 2258
>>> VAX TIME: 99- 1-19 18:20:34.540001 UT: 99- 1-19 17:19:24.076162
>> uVax 1 had 7 triggers!
>triggers > 2:ERP_MUON 4:BARI_MU 7:CIT_MONO 8:LIP 10:ST_MONOH 13:SPAM_MU
>eqp20 (ERP Muon) registered 3 hits in SM 2 :
2B11 ADC&TDC(1A/2A/1B/2B) 2398 2480 342 340 & 2012 1987 2028 1986 E/P 0 0
2C11 ADC&TDC(1A/2A/1B/2B) 1346 1420 214 211 & 2070 2132 2081 2143 E/P 0 0
2T12 ADC&TDC(1A/2A/1B/2B) 808 1060 202 210 & 2141 2260 2153 2274 E/P 0 0
Box -mTIME(sec,NoAt/At) PhotoElectrons(0/1) Energy(MeV, T/PE) Position(cm,T/PE)
2B11 334.18 333.85 1419.3 1378.3 68.7 68.7 -29.77 -16.92
2C11 348.72 348.74 670.9 799.8 38.2 38.2 99.64 93.35
2T12 364.42 364.68 582.8 1021.8 42.3 41.7 218.85 246.20
>eqp30(CSPM-uV):(LOW ) 2bL1 2bL0 2cL1 2cL0 2tH1 2tH0 -- THESE ARE THE SPAM HITS
>>Streamer tube tracking results for LIP hits: -- TRACK INTERCEPTS FOR LIP HITS
2T09 X/Y/Z: 2085.612 348.169 977.100 Plen: 24.186
2C07 X/Y/Z: 1935.338 659.670 538.000 Plen: 24.186
2B05 X/Y/Z: 1781.299 978.974 87.900 Plen: 24.186
2T10 no track intercept
2B08 no track intercept
The whole pathology seems to be related to general trigger processing time
issues of the ERP circuitry and/or subtle acquisition/syncronization problems.
To my surprise, this symptom is not uniform throughout the detector (SM2 seems
to suffer more from that) and more than that, it *varies* as a fuction of time.
In order to appreciate if global ERP rates affect this symptom, I have checked
for correlation between ERP trigger rate and inefficiencies-- this yielded
practically null result. The quantity that I checked (and show on the plot) is the
faction of CSPAM triggers that does not have an ERP hit in the corresponding patch
(= supercounter= sum of 8 adjacent counters). While this is typically at the
few per thousand (4-8%%) level, it climbs up for SM2 to 15-25 per thousand.
The effect is more obvious when you check for the fraction of missed (by the ERP)
of HIPT triggers as a function of time. HIPT discriminates at ~2V end-to-end
and any ERP inefficiency should certainly be *very* stable as a function of
time and throughout the detector and *very* close to zero (<0.001) Instead, during
problem runs, the fraction of missed HIP triggers increases to few percent
level, suggesting that the missed tracks are of higher energy rather than
of average energy.
I am going through similar plots for the rest of the detector and for the last
4 years of running. I'll point you to the final plots when they become ready.
BTW, the effect is not correlated to gain resets-- it happens also in between
gain reset period. Any LOOK-UP-TABLE effect might be hidden therein, although
I doubt it as it can hardly explain the lack of HIP hits. It'll take a bit more
of homework before rulling this out though.
Let me know if you have any comments, remarks or questions,
A presto,
--Erik
PS: check out http://wsgs02.lngs.infn.it:8000/macro/cspam_vs_erp.gif
for the SM2 summary plot -- "SPAM PATCH 2000 to 2999" label on the plots
implies that all SPAM supercounters of SM2 have been taken into account.
This archive was generated by hypermail 2a24 : Wed Jan 26 2000 - 11:31:23 EST
