• Created by Unknown User (pelassa), last modified on Aug 06, 2008

!! CAUTION: WORK IN PROGRESS !!

This page is open : please feel free to add any comment, suggestion, question. But when you do that, please put it in a form like "Name - Date : Comment" so that we can keep track of everything, and please also send me an email (vero.pelassa@gmail.com) , where you briefly explain your change and the reason for it. Thanks!

The questions in violet are adressed to the Monitoring experts.

here we take the example of run 236084237 - first run processed.

All plots are accessible via 'expert' mode and filter 'Off' 

Some useful informations are on the duty scientist checklist, in part "Data Quality" : how to get the configuration (enabled triggers list, window trigger width, values of the thresholds, ...), some explainations about the plots.

In bold face : the name of the histogram or plot in the folders on the left
In italic : the real name of the histogram or plot as stored in the root files - yes there are root files with all these histograms and plots stored in them (smile)

What to look at first?

check the temperatures of each CAL tower from the telemetry trending page (see Duty Scientist Checklist)
(the external towers should be a bit cooler than the central ones, and show variation according to the orbit. all temperatures should be between 10 and 15 C)

for a particular run:

  • scan the first available directory between FastMon and Digi - as you can find many FastMon plots in Digi
  • In red bold face we tried to figure out some high priority plots - our guess...

Configuration of this run: 236084237

Moot Key

conSciOps_noCal :

  • FHE = 1500 MeV = threshold for CAL_HI trigger (conservative - nominal value is 1000 MeV)
  • FLE = 150 MeV = threshold for CAL_LO (conservative - nominal value is 100 MeV)
  • LAC = 4 MeV = CAL threshold (conservative - nominal value is 2 MeV)

in TRG_GEM file :

  • complement to window open mask: ROI, CAL_HI, CAL_LO should open no trigger window (in nominal config only CAL_LO and ROI can't open)
  • triggers enabled : (same as nominal)
    • 0 = External
    • 1 = ROI (+ no TKR)
    • 2 = Solicited
    • 3 = Periodic (4-range readout)
    • 4 = CNO + CAL_LO + TKR + ROI (4-range readout)
    • 5 = CNO
    • 6 = CAL_HI (+ no CNO)
    • 7 = TKR + no ROI (+ no CAL_HI + no CNO)
    • 8 = CAL_LO only => as it can't open the window, it's as if it were disabled!
    • 9 = CAL_LO + TKR + ROI (+ no CNO + no CAL_HI)
    • 10 = TKR + ROI (+ no CAL_HI + no CAL_LO + no CNO)

in filter_Associate file : filters enabled = GAMMA_normal, HIP_normal, DGN_gem (nominal config)

FastMon

CAL

Hit Map :
(FastMon_CalXHit_NHit_Counter_TowerCalLayer_TH2)
generally homogeneous, central towers being more hit than the external ones, ok.

one bin per log-end, ie 2 per crystal.
an empty bin means a channel - or even a crystal, if 2 empty bins together - is dead, a significant spike may show a pb of zero-suppression for this channel.

(add the map of positions of the towers?)

Hits per log end : (FastMon_CalLogEndRangeHitCounter_Range3_TH2_Tower , idem for 0, 1, 2)
one map per range: 0 = LEX8, 1 = LEX1, 2 = HEX8, 3 = HEX1
as above, these maps can indicate dead channels, dead crystals - empty bins, or noise pbs - spikes.

more hits in the central towers than in the external towers, ok.

Number of hits in each Cal log end for HEX1, HEX8 and LEX1 ranges :
central towers homogeneously hit.
patterns on the external towers are due to the X-Y orientation of the crystal logs : layer 0 parallel to X, layer 1 parallel to Y, etc. so that these towers appear more hit in the "inner sides" for side towers and in the "inner corners" for corner towers - all at the same level as the central towers, what looks good.

(add a sketch of a CAL tower with X-Y orientation of the crystals?)

Number of hits in each Cal log end for LEX8:
significant spikes appear in crystals:
(Tower, Layer, Crystal) = (0,0,6), (2,1,5), (2,7,7), (3,7,5), (7,1,5), (9,0,5), (9,4,5), (10,0,6)

it's consistent with the list of "hot"/noisy crystals pointed out by Sasha (see here)

Missed logs map : (FastMon_ZeroCalXHit_NHit_Counter_TowerCalLayer_TH2)
homogeneous, ok. and set at a high number, what looks normal, since an event lights on a small fraction of the CAL, in average.

Num. logs hit (LAT) : (FastMon_CalX_Total_NHit_TH1)
goes like few 10^5 x exp(-nb_of_crystals) + few 10^2
maximum is at 0
the (small) peak around 8-10 crystals is due to mips

Num. logs hit (tower) : (FastMon_CalX_NHit_TH1_Tower)
same as LAT level, but with a peak also at 96 crystals. might be an accumulation effect - events triggering one complete tower or more appear here all together. needs further investigation...

Num. towers hit : (FastMon_CalTower_NHit_TH1)
maximum at 0, and gently decreasing, what looks normal.
rmk Sasha: opssim1 shown that one hot log or low LACs make the mean move to higher values

GEM (almost same as Digi GEM)

Arrival Times

NB: 1 Tick = 50 ns 

CAL_HI, CAL_LO : (FastMon_CondArrCalHiTrigger_TH1 FastMon_CondArrCalLoTrigger_TH1)
most @ 31 = CAL_HI most of the time didn't open the window, ok
many @ 0 : cross-check with CAL_HI (resp. CAL_LO) OpensWindow (see next sections) suggests that when CAL opens the window it is not alone, in fact TKR may open and CAL simply trigger in the following 50ns - as many entries in CAL_HI (resp. CAL_LO) Opens Window TKR distribution as counts @0 in CAL_HI (resp. CAL_LO) arrival times distribution : 9000 (resp. 16000)

CNO : (FastMon_CondArrCNOTrigger_TH1)
peak @ 4

ROI : (FastMon_CondArrROITrigger_TH1)
peak @ 4, ok as ROI should not open the window here.

TKR : (FastMon_CondArrTKRTrigger_TH1)
almost all @ 0, ok as TKR has to trigger before anything else.

CNOOpensWindow

CAL_HI, CAL_LO : (FastMon_CondArrCalHi_CNOOpensWindow_TH1 FastMon_CondArrCalLo_CNOOpensWindow_TH1)
@31, ok, they should not open the window.

TKR,ROI : (FastMon_CondArrTKR_CNOOpensWindow_TH1 FastMon_CondArrROI_CNOOpensWindow_TH1)
all @0, ok, means that CNO doesn't open the window alone.

CAL_HI (resp. CAL_LO) OpensWindow

CAL_LO (resp. CAL_HI) : (FastMon_CondArrCalLo_CalHiOpensWindow_TH1 ...)
@31, ok as they are not supposed to open too often the window

CNO, ROI : (FastMon_CondArrCNO_CalHiOpensWindow_TH1 ...)
@4, ok as they should not open the window here.

TKR : (FastMon_CondArrTKR_CalHiOpensWindow_TH1)
all @0, ok, means that CAL_HI (resp. CAL_LO) doesn't open the window alone.

TKROpensWindow

CAL_HI (resp. CAL_LO) : (FastMon_CondArrCalHi_TKROpensWindow_TH1 ...)
almost all @31,some @0 what is consistent with CAL_HI (CAL_LO) arrival time distributions, and with CAL_HI (CAL_LO) OpensWindow.

CNO, ROI : (FastMon_CondArrCNO_TKROpensWindow_TH1 ...)
@4, ok as they should not open the window

Trigger Vectors

Keep an eye on the configuration file (enabled triggers, etc.) when you look at these plots.

CAL_HI map (tower) : (FastMon_CalHiTrigger_Tower_TH1)
central towers more hit than external ones in X direction. a small asymmetry in Y direction: +Y more hitted than -Y towers. probably due to the inclination of the satellite wrt zenith : to be checked.
Of course, if one GLAST CAL Analog Front-End (GCFE) is hot, one of the towers will be significantly out of family.

CAL_LO map (tower) : (FastMon_CalLoTrigger_Tower_TH1)
central towers more hit than external ones, ok.
Of course, if one GLAST CAL Analog Front-End (GCFE) is hot, one of the towers will be significantly out of family.

ROI map (tower) : (FastMon_AcdGemROI_Tower_TH1)
corner-towers more hit than side-towers, themselves more hit than central towers. it's due to the tiles on the sides of the ACD. asymmetry in Y direction: -Y > +Y, inverted wrt CAL_HI, ok.

TKR map (tower) : (FastMon_TkrTrigger_Tower_TH1)
same as CAL_HI, but the Y asymmetry is inverted, what seems normal.

Word

Warning, known feature:
Some of the following histograms are filled with a weight ~equal to the value of the word.
So that the number of entries does not correspond to what is indicated on the Y-axis. The Experts are looking into that already.
08-08-05, from Johan : Fixed now

Keep an eye on the configuration file (enabled triggers, etc.) when you look at these plots.

"reminder" of what the bits mean:
1 = ROI
2 = TKR
4 = CAL_LO
8 = CAL_HI
16 = CNO
32 = PERIODIC
64 = SOLICITED
128 = EXTERNAL

Look for the "Here" link in the label of any of the following plots to get the list of all combinations and corresponding values of the word.

CAL_HI : (FastMon_CondSummaryWord_WithCalHi_TH1)
when CAL_HI triggers : TKR always triggers, CAL_HI never triggers alone, ok. most of the time with all physical triggers: energetic events.
peak at 31 = CNO+CAL_HI+CAL_LO+TKR+ROI
peak at 15 =         CAL_HI+CAL_LO+TKR+ROI

CAL_LO : (FastMon_CondSummaryWord_WithCalLo_TH1)
when CAL_LO triggers : TKR always triggers

CNO : (FastMon_CondSummaryWord_WithCNO_TH1)
when CNO triggers : alone (allowed), or with TKR, or with CAL_HI or/and CAL_LO
peak at 31 = CNO+CAL_HI+CAL_LO+TKR+ROI

Condition Summary : (FastMon_CondSummaryWord_TH1)
zoom < 32 :
the highest peak is at 2: most of the events trigger only the TKR, so we expect this bin to be the highest
1 is at 0: ROI doen't trigger alone, ok.
4, 5, 8, 9, 12, 13 at 0: no CAL_LO nor CAL_HI without the TKR, ok as they should not open the window they can't trigger alone.
17 at 0: no CNO with just the ROI and no TKR, ok.
21, 25 at 0: no CNO with CAL_LO or CAL_HI and ROI without the TKR, ok.
29 at 0: no TKR, no trigger!
32: periodic trigger is on, ok.

Periodic : (FastMon_CondSummaryWord_WithPeriodic_TH1)
>32 <64: it is not much probable that the periodic trigger meets a real event, so there are very few counts.
Periodic trigger triggers mostly alone, as it should be.

ROI : (FastMon_CondSummaryWord_WithROI_TH1)
when ROI triggers : TKR always triggers
peak at 31 = CNO+CAL_HI+CAL_LO+TKR+ROI

Solicited : (FastMon_CondSummaryWord_WithSolicited_TH1)
Solicited trigger : very few entries, it is on, but rarely triggers.
Why 2 entries at 64 AND 66 ? - Solicited triggers with TKR?

TKR : (FastMon_CondSummaryWord_WithTKR_TH1)
when TKR triggers : mostly alone, ok.

Digi

CAL

CAL Hit map (tower,layer) : (Digi_CalXHit_NHit_Counter_TowerCalLayer_TH2)
central towers more hit than external ones, ok
look for spikes(=hot logs) and holes(=dead logs)

Missed logs map (tower, layer) : (Digi_ZeroCalXHit_NHit_Counter_TowerCalLayer_TH2)
Number of times there is no hit in a calorimeter layer : homogeneous, ok.
Complement of "CAL Hit map" plot.

In the two following plots the bin 0 is not filled. Could you fill it? These plots are supposed to help us check zero-suppression...
08-08-05, from Johan : 0 is removed in both Digi and FastMon plots-> Jira

Num. logs hit (LAT) : (Digi_CalX_Total_NHit_TH1)
Distribution of CAL cristal hits per events (all towers) : few 10^5 x exp(-Number) + few 10^2, mip peak @ 8-10
(If zero suppression is bad, the minimum occupancy won't be zero. If it's really bad, the peak will be somewhere above zero. Find out which tower is causing the problem, or if all towers are. So far the first bin is just not filled.)

Num. logs hit (tower) : (Digi_CalX_NHit_TH1_Tower)
(If zero suppression is bad, the minimum occupancy won't be zero. If it's really bad, the peak will be somewhere above zero. Find out which tower is causing the problem, or if all towers are. So far the first bin is just not filled.)
Distribution of CAL crystals hit per tower : same as LAT level plot, same for all towers, only "plateau" level may change from 150 to 300. ok
peak @ 96 : accumulation effect? backsplash from energetic events hitting 96 crystal for the tower, and also crystals in the other towers. further investigation needed...

Readout Range (cal crystal) : (Digi_CalXReadoutRange_TH2)
number of hits in a given channel and for a given range
in mean number of hits : range0=LEX8 > 1=LEX1 > 2=HEX8 and 3=HEX1.
central towers more hit than external ones, ok.
helps to find "hot ranges"?

GEM (almost the same as FastMon GEM)

only plots that aren't in FastMon appear in bold face

Arrival Times

CAL_HI : (Digi_CondArrCalHiTrigger_TH1)
CAL_LO : (Digi_CondArrCalLowTrigger_TH1)
CNO : (Digi_CondArrCNOTrigger_TH1)
ROI : (Digi_CondArrROITrigger_TH1)
TKR : (Digi_CondArrTKRTrigger_TH1)

CNOOpensWindow

CAL_HI : (Digi_CondArrCalHi_CNOOpensWindow_TH1)
CAL_LO : (Digi_CondArrCalLo_CNOOpensWindow_TH1)
ROI : (Digi_CondArrROI_CNOOpensWindow_TH1)
TKR : (Digi_CondArrTKR_CNOOpensWindow_TH1)

CalHiOpensWindow

CAL_LO : (Digi_CondArrCalLo_CalHiOpensWindow_TH1)
CNO : (Digi_CondArrCNO_CalHiOpensWindow_TH1)
ROI : (Digi_CondArrROI_CalHiOpensWindow_TH1)
TKR : (Digi_CondArrTKR_CalHiOpensWindow_TH1)

CalLoOpensWindow

CAL_HI : (Digi_CondArrCalHi_CalLoOpensWindow_TH1)
CNO : (Digi_CondArrCNO_CalLoOpensWindow_TH1)
ROI : (Digi_CondArrROI_CalLoOpensWindow_TH1)
TKR : (Digi_CondArrTKR_CalLoOpensWindow_TH1)

TkrOpensWindow

CAL_HI : (Digi_CondArrCalHi_TKROpensWindow_TH1)
CAL_LO : (Digi_CondArrCalLo_TKROpensWindow_TH1)
CNO : (Digi_CondArrCNO_TKROpensWindow_TH1)
ROI : (Digi_CondArrROI_TKROpensWindow_TH1)

Trigger Vectors

CAL_HI map (tower) : (Digi_CalHiTrigger_Tower_TH1)
CAL_LO map (tower) : (Digi_CalLoTrigger_Tower_TH1)

CNO map (GARC) : (Digi_AcdGemCNO_GARC_TH1)
discrepancies between different areas due to their various sizes.
asymmetry in Y direction: -Y > +Y, inverted wrt CAL_HI, ok.

ROI map (tower) : (Digi_AcdGemROI_Tower_TH1)
TKR trg map (tower) : (Digi_TkrTrigger_Tower_TH1)

Word

No weight-problems here (see FastMon/GEM/Word...)

CAL_HI : (Digi_CondSummaryWord_WithCalHi_TH1)

CAL_HI (Extended) : (Digi_CondSummaryWordWithCalHi_Extended_TH1)
<32: see not extended
@0 except for a few counts in bit 63: a few energetic events (31) coincident with peridoc trigger (32). very rare, as should be.

CAL_LO : (Digi_CondSummaryWord_WithCalLo_TH1)

CAL_LO (Extended) : (Digi_CondSummaryWord_WithCalLo_Extended_TH1)
<32: see not extended
@0 except a few counts in bits 36, 39, 63: very few coincidences with periodic trigger, as should be. (probably the same as above: energetic events)

CNO : (Digi_CondSummaryWord_WithCNO_TH1)

CNO (Extended) : (Digi_CondSummaryWordWithCNO_Extended_TH1)
<32: see not extended
@0 except a few counts in bits 48, 51, 63: very few coincidences with periodic trigger, as should be. (probably the same as above: energetic events)

Condition summary : (Digi_CondSummaryWord_TH1)

Condition summary (Extended) : (Digi_CondSummaryWord_Extended_TH1)
<32: see not extended
very few counts between 32 and 63: coincidences with periodic trigger are rare, as should be.
a few counts in bits 64 and 66: solicited (+TKR) trigger

Periodic : (Digi_CondSummaryWord_WithPeriodic_TH1)

Periodic (Extended) : (Digi_CondSummaryWord_WithPeriodic_Extended_TH1)
shows the (very few) coincidences between periodic and physical triggers.

ROI : (Digi_CondSummaryWord_WithROI_TH1)

ROI (Extended) : (Digi_CondSummaryWord_WithROI_Extended_TH1)
<32: see not extended
@0 except a few counts in bits 33, 39,... 63: very few coincidences with periodic trigger, as should be. (probably the same as above: energetic events)

Solicited (Extended) : (Digi_CondSummaryWord_WithSolicited_Extended_TH1)
TKR : (Digi_CondSummaryWord_WithTKR_TH1)

TKR (Extended) : (Digi_CondSummaryWord_WithTKR_Extended_TH1)
<32: see not extended
@0 except a few counts in bits 34, 39,... 63: very few coincidences with periodic trigger, as should be. (probably the same as above: energetic events)
counts in bit 33 ?????
few counts in bit 66: solicited trigger

Trigger Engines : (Digi_TriggerEngine_TH1)
0, 1, 8 at 0: external trigger, ROI alone, CAL_LO alone. not allowed, ok.
7 highest: it's the TKR, ok.
4, 9, 10 high: contain the TKR

DigiLong

Baldini

The main difference between the plots present in this directory and those of the other DigiLong subdirectories is the computation method. Here real gaussian fits are performed on Gain Ratios and Pedestals.

08-08-06, about error bars : Luca changed his code so that it now displays the real error given by the fit (or 0 in some cases, like the plot of the number of DOF) instead of the square root of the central value - it's more readable and sensible, thank you!

CalGains

Ratio of 2 signals (big vs small diode, neg vs pos face, etc.) is computed for each hit in each crystal. Then the distribution for each crystal is fitted (gaussian, chi2 statistics).

legend for plots names:
R = ratio
P = big diode, positive face
p = small diode, pos face
M = big diode, neg face
m = small diode, neg face

  • Negative face, Big diode / Small diode :

calGainsAnalyzer_RMm_ChiSquare_TH1 :
total chi2 increases when the number of counts increases, so that tower-groups are recognizable: central towers still more hit than external ones, so that the total chi2 is higher.

calGainsAnalyzer_RMm_DOF_TH1 :
idem: DOF = nb of counts - 3 (assuming 3 parameters for the gaussian function)

calGainsAnalyzer_RMm_FitProb_TH1 :
goodness of fit.
probability for the chi2 function to have the value it has considering the corresponding nb of DOF.
either at 0 with small error bars (0.1) or near to 1 with huge error bars (between -0.2 and 2): are the fits bad?

calGainsAnalyzer_RMm_Mean_TH1 :
~8
calGainsAnalyzer_RMm_RMS_TH1 :
+/- 2 to 3

calGainsAnalyzer_RMm_ReducedChiSquare_TH1 :
goodness of fit.
in average ~1.2 +/- 1.4 ... what can one tell?

  • Big diode, Positive face / Neg face :

Is it possible not to take into account the "unphysical entries" that cause the peak at 1?
It is likely that it affects the fit very badly...
Johan - experts are working on it.

calGainsAnalyzer_RPM_ChiSquare_TH1, calGainsAnalyzer_RPM_DOF_TH1 :
tower pattern not much recognizable.
one can see the layer alternance...
spikes for a few crystals.

calGainsAnalyzer_RPM_FitProb_TH1 :
always 0 ...? probably because if the peak at 1 due to unphysical events (see Digilong/CalGains/AllChannels)

calGainsAnalyzer_RPM_Mean_TH1, calGainsAnalyzer_RPM_RMS_TH1 :
~ 1 + /- 0.2, what seems good... but what about the quality of the fit?

calGainsAnalyzer_RPM_ReducedChiSquare_TH1 :
goodness of fit.
~ 4 hum...
+ same spikes as in total chi2 histograms.
gaussian form not adapted? (chi2 stat should be ok considering the nb of DOF: ~ 100)

  • Pos face, Big diode / Small diode :

calGainsAnalyzer_RPp_ChiSquare_TH1, calGainsAnalyzer_RPp_DOF_TH1 :
fit. tower pattern recognizable. as above, the total chi2 and the nb of DOF increase when the nb of counts increases.

calGainsAnalyzer_RPp_FitProb_TH1 :
goodness of fit. here the chi2 prob is 0 +/- 0.2 or 1 +/- 1

calGainsAnalyzer_RPp_Mean_TH1, calGainsAnalyzer_RPp_RMS_TH1 :
~8 +/- ~ 2.5

calGainsAnalyzer_RPp_ReducedChiSquare_TH1 :
~1.2 +/- 1.2

CalPeds

It is not easy to fit with a gaussian function a histogram that has only 3 bins filled, what is the case of LEX1 and HEX1 peds: thus the value of the chi2 prob. Would it be possible to "mix" both methods? Gaussian fit for wide distributions (LEX8, HEX8), and the raw mean and rms for the narrow distributions?
08-08-05, from Johan : will be discussed in CAL meeting.

  • Chi square of the fits :

calPedsAnalyzer_CalXAdcPedChiSquare_HEX1_TH1, calPedsAnalyzer_CalXAdcPedChiSquare_HEX8_TH1, calPedsAnalyzer_CalXAdcPedChiSquare_LEX1_TH1, calPedsAnalyzer_CalXAdcPedChiSquare_LEX8_TH1 :
stays homogeneous from one tower to the other - but not inside a tower so that the pattern is recognizable.
chi2 average level: LEX8 > LEX1 > HEX8 > HEX1, what is in agreement with the number of hits: LEX8 > LEX1 > HEX8, HEX1 - see Hit maps per log end in FastMon (FastMon_CalLogEndRangeHitCounter_Range0_TH2_Tower ...)

  • DOF for each fit performed :

calPedsAnalyzer_CalXAdcPedDOF_HEX1_TH1, (}calPedsAnalyzer_CalXAdcPedDOF_HEX8_TH1, calPedsAnalyzer_CalXAdcPedDOF_LEX1_TH1, calPedsAnalyzer_CalXAdcPedDOF_LEX8_TH1 :
homogeneous from a tower to the other.
DOF nb: LEX8 > HEX8 > LEX1 > HEX1 - what does not seem to agree with the nb of hits in each range... ??

  • Chi2 probability :

Is it possible to plot it in log scale and range from, say, 10^-7 to 1?
08-08-05, from Johan : yes, I'll open a Jira

calPedsAnalyzer_CalXAdcPedFitProb_LEX8_TH1, calPedsAnalyzer_CalXAdcPedFitProb_HEX8_TH1 :
~0 +/- 0.4 or ~1 +/- 1.

calPedsAnalyzer_CalXAdcPedFitProb_LEX1_TH1, calPedsAnalyzer_CalXAdcPedFitProb_HEX1_TH1 :
often (always for LEX1) 0. it may be due to the small number of bins on which the fit is performed (3-4)

  • Mean value of the pedestal :

calPedsAnalyzer_CalXAdcPedMean_HEX1_TH1, calPedsAnalyzer_CalXAdcPedMean_LEX1_TH1 :
~ 220 to 230 adc counts, as should be.

calPedsAnalyzer_CalXAdcPedMean_HEX8_TH1, calPedsAnalyzer_CalXAdcPedMean_LEX8_TH1 :
~ 500 to 750, as should be.

  • RMS of the pedestal :

calPedsAnalyzer_CalXAdcPedRMS_HEX1_TH1, calPedsAnalyzer_CalXAdcPedRMS_LEX1_TH1 :
~0.6 +/- 0.8, what looks generally good. (the error bar is the square root of the mean value. - not any longer)

calPedsAnalyzer_CalXAdcPedRMS_HEX8_TH1, calPedsAnalyzer_CalXAdcPedRMS_LEX8_TH1 :
resp. ~ 3.5 and ~5.5 +/- 2.

  • Reduced Chi2 of each fit :

calPedsAnalyzer_CalXAdcPedReducedChiSquare_HEX1_TH1 :
mostly at 0 narrow peak...

calPedsAnalyzer_CalXAdcPedReducedChiSquare_HEX8_TH1 :
~ 1 +/- 0.2 : looks good.

calPedsAnalyzer_CalXAdcPedReducedChiSquare_LEX1_TH1 :
~ 16 (!!!!) narrow...

calPedsAnalyzer_CalXAdcPedReducedChiSquare_LEX8_TH1 :
~ 2.5 +/- 2.

CAL Pedestals

Here (and in the following section) the method is to cut the "feet" of each distribution (~10% at each end) and to take the raw mean and rms the distribution.

All Channels : (CalPed_CalXAdcPed_TH1_TowerCalLayerCalColumnFR)
NB: Range 0=LEX8, 1=LEX1 , 2=HEX8, 3=HEX1
Range 0 and 2 distributions are expected to be wider than ranges 1 and 2 due to a larger gain.
Range 0 is wider than range 2 due to larger diode.
As they are intrinsically related to the electronic, the mean values of pedestals sould not change from a run to another.
That is, in ADC counts: LEX1 and HEX1 both show a narrow peak centered around 200, and LEX8 and HEX8 both show a broader peak centered around 500.
Tip: in order to scan faster the (numerous) histograms, you can choose a single range at a time in the selection form (upper-left corner of the plot frame)

Deviation RMS : (CalPed_CalXPedDev_RMS_TH1)
RMS gives an idea of the average noise. Look for outliers: these are noisy channels.
(One can wonder if it is of interest to find by hand all these noisy channels and record them...)

Deviation mean : (CalPed_CalXPedDev_Mean_TH1)
This is the main plot (as we do not really care about absolute values of pedestals ...)
This plot shows that the mean values of pedestals for all channel does not deviates for more than 1 ADC counts.
Good !
NB: Is this plot really in auto-scale ? I am a bit surprised by this great agreement ...
08-08-05, from Johan : good remark, and good question! JIRA

HEX1 range RMS : (CalPed_CalXAdcPedRMS_HEX1_TH1)
Blue=value, black=error bars

HEX1 range mean : (CalPed_CalXAdcPedMean_HEX1_TH1)

HEX8 range RMS : (CalPed_CalXAdcPedRMS_HEX8_TH1)

HEX8 range mean : (CalPed_CalXAdcPedMean_HEX8_TH1)

LEX1 range RMS : (CalPed_CalXAdcPedRMS_LEX1_TH1)

LEX1 range mean : (CalPed_CalXAdcPedMean_LEX1_TH1)

LEX8 range RMS : (CalPed_CalXAdcPedRMS_LEX8_TH1)

LEX8 range mean : (CalPed_CalXAdcPedMean_LEX8_TH1)

CAL Ratios

the following ratios (P/M, P/p, M/m) measure global PDA bound stability:

CalPed_RMm_TH1_TowerCalLayerCalColumn :
same as  LEX1 ( - ) vs HEX8 ( - ) mean, but plot histograms for each crystals*.*
Ratio is expected to be ~6 (for signal events and pedestal substracted)
08-08-06, from Sasha : LEX1/HEX8 ratio was ~6 for HEgain=13, now with HEgain=15 (it's a bit misleading, but gain=15 means the lowest possible gain - in this case all four feedback capacitors are connected in parallel) LEX1/HEX8 ratio is ~7.5 .

CalPed_RPM_TH1_TowerCalLayerCalColumn :
same as LEX1 ( + ) vs LEX1 ( - ) mean, but plot histograms for each crystals.
Distribution is wide due to longitudinal position of the event.
Peak at 1 seems not physical.
Due to ratio of default values setting ? it may happen that if a channel is read empty it is set to the same value as the other face of the crystal, or if both aren't read properly they're both set to 99. in both cases the ratio is 1.
Is it possible not to take these unphysical entries into account?

CalPed_RPp_TH1_TowerCalLayerCalColumn :
Same as first plot (LEX1 ( - ) vs HEX8 ( - ) mean) but for ( + ) face.

LEX1 ( + ) vs HEX8 ( + ) RMS : (CalPed_RPp_RMS_TH1)

LEX1 ( + ) vs HEX8 ( + ) mean : (CalPed_RPp_Mean_TH1)
Used for intercalibration (from low energy to hign energy).
Ratio is expected to be ~6 (anyway, it's almost flat and around 7.5).
08-08-06, from Sasha : LEX1/HEX8 ratio was ~6 for HEgain=13, now with HEgain=15 (it's a bit misleading, but gain=15 means the lowest possible gain - in this case all four feedback capacitors are connected in parallel) LEX1/HEX8 ratio is ~7.5 .

LEX1 ( + ) vs LEX1 ( - ) RMS : (CalPed_RPM_Mean_TH1)

LEX8 ( + ) vs LEX8 ( - ) mean : (CalPed_RPM_RMS_TH1)
Ratio is expected to be ~1 as diodes have the same dimension.

LEX1 ( - ) vs HEX8 ( - ) RMS : (CalPed_RMm_RMS_TH1)

LEX1 ( - ) vs HEX8 ( - ) mean : (CalPed_RMm_Mean_TH1)
Used for intercalibration (from low energy to high energy).
Ratio is expected to be ~6 (anyway, it's almost constant around 7.5)
08-08-06, from Sasha : LEX1/HEX8 ratio was ~6 for HEgain=13, now with HEgain=15 (it's a bit misleading, but gain=15 means the lowest possible gain - in this case all four feedback capacitors are connected in parallel) LEX1/HEX8 ratio is ~7.5 .

CAL Readout

 4-ranges readout (all events) : (CalPed_Range4AllEvents_TH1)
Why do we have more events in 4-range readout  (~5000 events) than for the periodic trigger (3000 events) ? 
Because engine 4 (heavy ions filter, used for calibration) is also in 4 range readout ...

 4-ranges readout (periodic) : (CalPed_Range4Periodic_TH1)
Check that periodic trigger occurs in 4-range readout. there should be no entry in "1-range".

Zero Suppress : (CalPed_ZeroSuppressForPeriodicAndRange4_TH1)
"Distribution of Zero Suppress bit. ZERO means evt is NOT zero-suppressed, ONE means it is zero-suppressed. Only periodic triggers with 4 range readout are being considered."
from the previous plot we only consider the "subgroup" that is in 4-range readout, and check that all these events were not zero-suppressed. these should be all the periodic triggers.

Merit

CAL

CAL Energy Corr : (Merit_CalEnergyCorr_Log10_TH1)

CAL Energy Raw : (Merit_CalEnergyRaw_Log10_TH1)

CAL Radiation : (Length Merit_CalCsIRLn_TH1)

CAL Track Angle : (Merit_CalTrackAngle_TH1)

CAL Transverse RMS : (Merit_CalTransRms_TH1)

CAL Transverse RMS (zoom) : (Merit_CalTransRms_Zoom_TH1)

CAL TwrEdge-Cntr Dist : (Merit_CalTwrEdgeCntr_TH1)

Recon

CAL

some of the following plots may give you rough idea of main thresholds - LAC, FLE, FHE - at CAL level or at crystal level. it might be quite long depending on the plot you inspect and is not as accurate as the dedicated methods used to measure LAC, FLE, FHE.

Distribution E. CAL_HI : (Recon_SuspCalHi_Highest0_EnergyDistribution_TH1)
gives a rough idea of the FHE at CAL level
here there is some "noise" below a rather clear threshold around 1.5 to 1.8 GeV, for a FHE set at 1.5GeV

Distribution E. CAL_LO : (Recon_SuspCalLo_Highest0_EnergyDistribution_TH1)
gives a rough idea of the FLE at CAL level
here there is some "noise" below a rather clear threshold around 180 to 200 MeV, for a FLE set at 150MeV

Distribution Energy (tow, lay) : (Recon_ReconEnergy_TH1_TowerCalLayer)
RECOGNIZED AS USELESS
distribution of measured energy per event in each tower/layer
Not sure it's usefull to have these histograms by layer, as it's hard to put alarms on it ( orbital fluctuation) and the shifters will never look at all of them.

Log10 Measured CAL Energy : (Recon_Log10ReconEnergy_HighEnergy_TH1)
distribution of total measured energy in the CAL per event
steep increase around 4 MeV - LAC
more statistics between 10 and 200 MeV than elsewhere.
decrease around 1 GeV.
two bumps at 10 and 20 GeV: related to geomagnetic cutoff.

Meas. Cal Ener. (tow,lay,col) : (Recon_ReconEnergy_VeryLowEnergy_TH1_TowerCalLayerCalColumn)
low energy tail (<5 MeV) of the reconstructed energy distribution in every crystal (...).
gives a rough idea of the LAC at crystal level
all begin around 4 MeV: it's where you read the LAC.

Meas. E. NegFace (tow,lay,col) : (Recon_Lac_Thresholds_FaceNeg_TH1_TowerCalLayerCalColumn)
as many histos as crystals in the CAL (...) showing the low-energy tail (<5 MeV) of energies reconstructed in each crystal from the Neg face: it can give a rough idea of the LAC for the negative face of each crystal - but it is quite long
all begin around 4 MeV, with only few noise below for a few crystals.

Meas. Energy Map (tower,layer) : (Recon_ReconEnergy_TowerCalLayer_TH2)
measured energy in each tower, and each layer.
more energy in the central towers, which is normal - geometry.
also more energy in the central layers, because of energetic events, that do not deposit all their energy at the "entry" of the CAL. but few - the most energetic ones - reach the bottom of the CAL.
it may happen that there is more energy in the first layers, if only CAL_LO and not CAL_HI is enabled - FLE calib runs for example.

Measured CAL Energy : (Recon_ReconEnergy_HighEnergy_TH1)
two peaks around 10 and 15-20 GeV: unique bump in the Cosmic Ray spectrum, modulated by the various cuts of the Gamma Filter - e.g. Pass_Thru lets everything enter above 20 GeV
statistic decreases as energy increases.
see Berrie's study of CalEnergyRaw peaks for decomposition by species.

Measured CAL Energy (low E) : (Recon_ReconEnergy_LowEnergy_TH1)
RECOGNIZED AS USELESS
if MIP filter is enabled: mip peak at 100 MeV - what is not the case here. a shoulder appears.
shoulder around 200 MeV
(bump of albedo photons?)
Diagnostic : may be the bump of the albedo photons ? not sure how much this plot is useful

Measured Cal E. (5 MeV trunc.) : (Recon_ReconEnergy_VeryLowEnergy_TH1)
gives a rough idea of the LAC at CAL level
here starts at 2 MeV, reaches a maximum at 4 MeV: it's where you read the LAC.

Recon_Lac_Thresholds_FacePos_TH1_TowerCalLayerCalColumn :
as many histos as crystals in the CAL (...) showing the low-energy tail (<5 MeV) of energies reconstructed in each crystal from the Pos face: it can give a rough idea of the LAC at crystal level - but it is quite long
all begin around 4 MeV, with only few noise below for a few crystals.
08-08-05, from Johan : this is a very precise measure of LAC thresholds
+++ NEW : SEE ALARMDIST/LACTHRESHOLD :
The distribution of LAC thresholds as measured by the alarm algorithm.
In addition to this, an alarm is put on each single plot of
Recon_Lac_Thresholds_FacePos_TH1_TowerCalLayerCalColumn so that you do not actually need
to have a look at all of them, but only look for alarms !

About the following 6 2D histograms:
for a given channel (= crystal face), we consider events firing this channel and triggering CAL_LO or CAL_HI. Here is plotted the highest energy (among events firing this channel) for which the tower containing this channel has requested CAL_LO or CAL_HI.
We try to find crystaltal faces that request CALLO or HI more often than their neighbors. There might be systematic differences from acceptance geometry.
We also look for outliers.

Suspect Cal-Hi Trigger (1) : (Recon_SuspCalHi_Highest0_TH2)
highest energy per crystal for events triggering CAL_HI
another rough idea of the FHE per crystal
here we sort of lack at statistics, but we can see a threshold around 1.8GeV

Suspect Cal-Hi Trigger (2) : (Recon_SuspCalHi_Highest1_TH2)
RECOGNIZED AS USELESS
second highest energy per crystal for events triggering CAL_HI
supposed to show counts at lower energy than Susp_CalHi_Highest(1), what is the case here
Useless if done without the diagnostic information. To be removed.

Suspect Cal-Hi Trigger (3) : (Recon_SuspCalHi_Highest2_TH2)
RECOGNIZED AS USELESS
third highest energy per crystal for events triggering CAL_HI
supposed to show counts at lower energy than Susp_CalHi_Highest(2), what is the case here
Useless if done without the diagnostic information. To be removed.

Suspect Cal-Lo Trigger (1) : (Recon_SuspCalLo_Highest0_TH2)
highest energy per crystal for events triggering CAL_LO
another rough idea of the FLE per crystal
here we sort of lack at statistics, but we can see a threshold around 180MeV

Suspect Cal-Lo Trigger (2) : (Recon_SuspCalLo_Highest1_TH2)
RECOGNIZED AS USELESS
second highest energy per crystal for events triggering CAL_LO
supposed to show counts at lower energy than Susp_CalLo_Highest(1), what is the case here
Useless if done without the diagnostic information. To be removed.

Suspect Cal-Lo Trigger (3) : (Recon_SuspCalLo_Highest2_TH2)
RECOGNIZED AS USELESS
third highest energy per crystal for events triggering CAL_LO
supposed to show counts at lower energy than Susp_CalLo_Highest(2), what is the case here
Useless if done without the diagnostic information. To be removed.

Zero Rec. Ener. (tower, layer) : (Recon_ZeroReconEnergy_TowerCalLayer_TH2)
number of times that the reconstructed energy is zero in a layer/tower.
homogeneous, what looks good.
high number of entries, is it normal?
should be normalized to the number of times that a layer/tower is fired - coming soon according to the label.

Trending

Digi

CAL

CalX Hit (tow) : (Digi_Trend_Rate_CalXHit_NTowerHit_Tower)

CalX Hit (tow,lay) : (Digi_Trend_Rate_CalXHit_Hit_TowerCalLayer)
RECOGNIZED AS USELESS
Is there any meaning to "noisy layer" from the point of view of the electronic ?
the plot at the log level should suffice if there is a problem

CalX Hit (tow,lay,col) : (Digi_Trend_Rate_CalXHit_Hit_TowerCalLayerCalColumn)

CalX Hit=0 (tow,lay) : (Digi_Trend_Counter_CalXHit_ZEROHit_TowerCalLayer)
RECOGNIZED AS USELESS
If this counter means something then trending at the log level should be enough, why should we integrate by layer ?

CalX Hit=0 (tow,lay,col) : (Digi_Trend_Counter_CalXHit_ZEROHit_TowerCalLayerCalColumn)

Mean CAL Crystal Hits (LAT) : (Digi_Trend_Mean_Cal_NHit)

Mean CalX Hit (tow,lay) : (Digi_Trend_Mean_CalXHit_MHit_TowerCalLayer)
RECOGNIZED AS USELESS
Is there any meaning to "noisy layer" from the point of view of the electronic ?
the plot at the log level should suffice if there is a problem

Norm CalX Hit (tow,lay,col) : (Digi_Trend_OutF_Norm_CalXHit_Hit_TowerCalLayerCalColumn)

Norm CalX Hit=0 (tow,lay) : (Digi_Trend_OutF_CalXHit_ZEROHit_Fraction_TowerCalLayer)
RECOGNIZED AS USELESS
Is there any meaning to "noisy layer" from the point of view of the electronic ?

Norm CalX Hit=0 (tow,lay,col) : (Digi_Trend_OutF_CalXHit_ZEROHit_Frac_TowerCalLayerCalColumn)

Rate CAL Crystal Hit (LAT) : (Digi_Trend_Rate_CalXHit_Hit)

GEM

Acd Gem CNO Trigger (GARC) : (Digi_Trend_Rate_AcdGemCNO_GARC)

Acd Gem ROI (tower) : (Digi_Trend_Rate_AcdGemROI_Tower)

GEM Deadzone : (Digi_Trend_Rate_Deadzone)

GEM Discarded (LAT) : (Digi_Trend_Rate_Discarded)

GEM Prescaled (LAT) : (Digi_Trend_Rate_Prescaled)

GEM Rate (LAT) : (Digi_Trend_CounterDiffRate_GemRate)

Live Time Fraction : (Digi_Trend_DoubleDiffRate_LiveTimeFraction)

Rate CalHi Triggers (tower) : (Digi_Trend_Rate_CalHiTrigger_Tower)

Rate CalLo Trigger (tower) : (Digi_Trend_Rate_CalLoTrigger_Tower)

Rate TKR Trigger (tower) : (Digi_Trend_Rate_TkrTrigger_Tower)

TrgEngines Rate : (Digi_Trend_Rate_TriggerEngine)

DigiLong

CAL Pedestals

Pedestal Deviations : (CalPed_Trend_Mean_CalXPedDev_TowerCalLayerCalColumnFR)

Pedestals : (CalPed_Trend_Mean_CalXAdc_TowerCalLayerCalColumnFR)
RECOGNIZED AS USELESS
useless, only the deviation from the nominal value is interesting

CAL Ratios

LEX1(plus) vs HEX8(plus) mean : (CalPed_Trend_Mean_RPp_TowerCalLayerCalColumn)

LEX1(plus) vs LEX1(minus) mean : (CalPed_Trend_Mean_RPM_TowerCalLayerCalColumn)

LEX1(minus) vs HEX8(minus) mean : (CalPed_Trend_Mean_RMm_TowerCalLayerCalColumn)

FastMon

CAL

CAL Tower count : (FastMon_Trend_Mean_FastMon_CalTowerCount)

CAL log end range occupancy : (FastMon_Trend_Mean_FM_CalLogRHit_TowerCalLayerCalColumnFR)

Merit

CAL

Mean Energy Corr : (Merit_Trend_Mean_CalEnergyCorr)

Mean Energy Raw : (Merit_Trend_Mean_CalEnergyRaw)

Mean Radiation Length : (Merit_Trend_Mean_CalCsIRLn)

Mean Track Angle : (Merit_Trend_Mean_CalTrackAngle)

Mean Transverse RMS : (Merit_Trend_Mean_CalTransRms)

Mean TwrEdge-Cntr Dist : (Merit_Trend_Mean_CalTwrEdgeCntr)

Recon

CAL

Mean Cal E. (tow,lay,col) : (Recon_Trend_Mean_ReconEnergy_TowerCalLayerCalColumn)

Mean Cal E. (tow,lay,col,xfac) : (Recon_Trend_Mean_ReconEnergy_TowerCalLayerCalColumnF)

Mean Cal Energy (LAT) : (Recon_Trend_Mean_ReconEnergy)

Mean Cal Energy (tower) : (Recon_Trend_Mean_ReconEnergy_Tower)

Mean Cal Energy (tower, layer) : (Recon_Trend_Mean_ReconEnergy_TowerCalLayer)

Zero Cal Ener,frac. (tow,lay) : (Recon_Trend_OutF_ReconZeroEnergy_Fraction_TowerCalLayer)

  • No labels