Atlas SCT Test DAQ Release Notes

The following versions have been tagged in the CVS repository and may be retrieved as follows:

This PRODUCTION version is built to run on Windows platforms with version 3.02/07 of ROOT (WITHOUT experimental GDK driver).

Alternate project files are included such that users with access to the Microsoft Visual C++ compiler may rebuild the software to run with more recent versions of ROOT: please note that this is not recommended. If for some reason you must use SCTDAQ with a recent version of ROOT under Windows I would recommend that you avoid even the "optimised" gdk versions: when I tested version 3.10/02 I found it to be slower by at least a factor of 2!

SCTDAQ may also be built under Linux to run with ROOT version 3.02/07 or higher.

To use this version under Windows:

• 1) Install ROOT 3.02/07 (version WITHOUT experimental GDK driver)
• 2) Download and unzip sctdaq_3_42.zip
• 3) Choose the version of sctdaq\bin\stdll_*.dll appropriate to your system (CLOAC or NO_CLOAC) and rename it stdll.dll
• 4) Start ROOT from within the sctdaq directory (IMPORTANT!) and run ST as usual .x ST.cpp

The built versions will work for the following system configuration, with all VME boards set to the usual addresses:

• 1 CLOAC (CLOAC version only)
• 1 SLOG
• 1 MuSTARD
• 0-3 SCTLV (any flavour)
• 0-2 SCTHV
• c:\sctdaq
• d:\sctvar

Users with non-standard file system paths or VME configurations will need to edit the appropriate source files and recompile as usual. (For a system without a d: disk, a user with administrative priviliges may simply mount the c: disk as drive d.)

Additions/fixes/improvements:

• sctdaq/macros/AccumulateAnalysis.cpp, AccumulateTest.cpp
  • Simple test of the accumulate function of ABCD.

• sctdaq/macros/ConfirmationTest.cpp
  • Option to vary setting of SELECT for analogue part of the test (unused by default).

• sctdaq/macros/EndCapCold.cpp
  • Sequence to rerun the cold part of the endcap hybrid test (used at RAL).

• sctdaq/macros/EndCapHLTT.cpp
  • Modified time profile.

• sctdaq/macros/FullBypassTest.cpp
  • AERO mode now argument (with default=1).

• sctdaq/macros/NO.cpp
  Set FAIL if:
  • more than 15 NO_HI defects
  • more than 7 consecutive NO_HI defects
  • any chip has noise less than 400 ENC (to catch chips with gross calibration errors)

• sctdaq/macros/RedundancyTest.cpp
  • Use preburst to disable failing links. Hence if one hybrid/module fails the test, the others should still pass.

• sctdaq/macros/ResponseCurvePlot.cpp
  • When run online, add DB comment showing state of SELECT

• sctdaq/macros/StrobeDelay.cpp
  • Default SD fraction now 0.40

• sctdaq/macros/StrobeDelayPlot.cpp
  • Unnecessary dependency removed.

• sctdaq/macros/TLTT.cpp, TLTT.h
  • When testing ENDCAP hybrids, SELECT will toggle between consecutive confirmation tests.
  • Missing T sensors (-128C or less) are now excluded from calculations.
  • Set fail if a module is suspended during an LTT sequence.

• sctdaq/stdll/TST.cpp
  • Bugfix: NO test used to get stuck in semi-infinite loop if one module had been suspended, eg due to LV or T trip. This has now been corrected.

• sctdaq/macros/CharacteristaionTest.cpp ConfirmationTest.cpp ReceptionTest.cpp IVCurve.cpp
  Under Windows, sounds are now played at the end of these tests / test sequences.
  • sctdaq/bin/end.wav will play if the sequence ends normally.
  • sctdaq/bin/error.wav will play if the sequence is aborted.

• sctdaq/macros/ (General)
  • All standard plots now show hybrid/module temperature T0

• sctdaq/macros/ (General)
  • All standard plots now show hybrid/module temperature T0

Additions/fixes/improvements:

• sctdaq/macros/ (General)
  • All standard plots now show hybrid/module temperature T0

• sctdaq/macros/CharacterisationTest.cpp
  • Log start of test - see entry for TLogger
  • During CharacterisationReplay, use SW bad channel mask for NOPlot
  • During CharacterisationReplay, do not display plots for TimewalkPlot
  • During CharacterisationReplay, use correct arguments for StrobeDelayPlot

• sctdaq/macros/ConfirmationTest.cpp
  • Log start of test - see entry for TLogger
  • During ConfirmationReplay, use correct arguments for StrobeDelayPlot

• sctdaq/macros/FullBypassTest.cpp
  • New version to support use of AERO in mode 0 by means of adjustment of the VDCDAC threshold as Vdd is changed. Defaults are set for users with UNMODIFIED patch cards. Users of modified cards should edit this file to select an alternate value of the adjustment factor "Afactor".
  • Although the new option is the default, options are provided to use use AERO in modes 1 or 2 which use the SPY data. These may still be select if required, by editing the file.
  • Code drafted by PWP, completed by Scott Lindsay.

• sctdaq/macros/IVCurve.cpp
  • bugfix from Dave Robinson
  • bugfix from Allen Haim
  • Add option to 500V in 10V steps, 10S wait, power off (req. by R Bates)
  • Extra options for CiS detectors - will appear only when Endcap devices are declared in the sysmap
  • Option to run the IV curve on a single module - useful in the event of a visit by SCTHV communication gremlins...

• sctdaq/macros/No.cpp
  • Make ST_THR the default (scanned range adjusted to accomodate trim)

• sctdaq/macros/NoLBL.cpp
  • Load NoPlotLBL.cpp instead of NoPLot.cpp (bugfix from Bob Ely)

• sctdaq/macros/NoPlot.cpp
  • Additions based on work by Peter Kodys:
    • Option to suppress data for known bad channels (based on the mask file). If this is used, a DB comment is generated to make it clear that any unmaskable channels have been excluded from the quoted occupancy figures.
    • TMarkers now correct when ST_TTHR option in use.
    • TMarkers y point truncated to LOW_RANGE when log scale in use.

• sctdaq/macros/parameters.h
  • #define DATE_HISTORY

• sctdaq/macros/ResponseCurvePlot.cpp
  • I have rationalised the way in which defective channels are masked during the 3PG test. The following channels are now masked:
    • DEAD
    • STUCK
    • LO_GAIN gain < (0.75*chip_mean_gain)
    • HI_GAIN gain > (1.25*chip_mean_gain)
    • LO_OFFSET offset < (chip_mean_offset-80)
    • HI_OFFSET offset < (chip_mean_offset+80)
    • V_NOISY innse > (1.25*chip_mean_noise)
    Note that V_NOISY is a subset of the list of NOISY channels which is given by a cut on innse at (1.15*chip_mean_noise). V_NOISY is supposed to match (replace) the cut at 1.25*output noise traditionally applied during the trim analysis. For each channel that is masked, a DB comment is generated. In addition to all this, the range of the hp_offset histograms are now tuned to match the new offset cuts - hence modules which are trimmed high will no longer fail the RC test.
  • Bad channels (low/high gain/offset, very high noise) are no longer considered when determining the response curve for each chip. This makes more sense as, in practice, these are the channels that will be masked off. (This feature is most useful during replay sessions.)
  • A new option, activated by setting cut_adjustment=1, lowers the partbonded cut by 150ENC during cold tests. This was requested by UK-B to reduce the probability of strips being incorrectly identified as unbonded during cold tests. By default, this option is switched off.
  • Bugfix: Exclude points having sigma>1.5* channel mean from the channel by channel RC fit. This gives sensible values for channel gain, noise etc. in the presence of the "8fC" effect.
  • Enhancement: Tally the number of channels of each link for which n (or more) points have thus been excluded. This gives an idea of the extent to which a module has been effected by the "8fC" effect.
  • Bugfix: Exclude points for which the mean sigma is more than 1.5 * the chip mean from the chip by chip RC fit. (This code was present in SCTDAQ 3.41, but was broken in the initial DEV release of SCTDAQ 3.42. Now it's fixed.)
  • Bugfix: Exclude points for which the mean VT50 is zero from the chip by chip RC fit. (Happens in extreme cases of the "8fC" effect.)

• sctdaq/macros/StrobeDelay.cpp
  • Turn on integrated fitting - fits a point to the SD peak for each channel

• sctdaq/macros/StrobeDelayPlot.cpp
  • Completely rewritten, useful plots, more sensible cuts - see file for details
  • First bugfix: to make the macro run relibably with old (unfitted) Strobe Delay data files. (The original code used to crash...)
  • Second bugfix: to make the code which identifies empty bins coming between the rising and falling edges of the chip SD projection histogram. (In the initial version I had made some incorrect assumptions about bin widths...)
  • I also included (recoded) some changes from Bob Ely. The macro identifies the "innermost" SD edges for each chip and plots them on the graph. Bob's original code used these inner edges to calculate the SD: although code to do this is there, it is commented out for now.

• sctdaq/macros/ST.cpp
  • Modified to save hits per link histograms if they have been filled.
  • At idle, use DCSAction instead of DCSQuery (such that temperature trip is applied)
  • Add VDCDAC scan to ABCD tests menu - but only when endcap devices have been declared in the sysmap.

• sctdaq/macros/Timewalk.cpp
  • Bugfix - previous version left invalid trigger configuration if module 0 not present
  • Add "online" option as argument to wrapper (supress plots during replay)

• sctdaq/macros/ToolBox.cpp
  • Add ReadMask to read in most recent "-1" mask data, (Note: this behaviour is different from that of the version published by Peter Kodys, which reads the mask file from the config directory. I felt that perusal of the results directory was more suited to offline use of the macro during the reanalysis of existing data.)

• sctdaq/macros/TLTT.cpp
  • Log start of LTT

• sctdaq/macros/TrimRangePlot.cpp
  • Changes to reduce the number of false positive TR_RANGE defects. See file for details.
  • Changes to the PASS/FAIL and PROBLEM cuts, better aligned to the use of 1 bad channel chips. PROBLEM is reported if any channels have been masked. FAIL is set if a module has more than 7 consecutive masked channels or more than a total of 15 masked channels. (In the original implementation, FAIL was set if a single channel could not be trimmed, other defects such as NOISY not being taken into consideration.)
  • Now that noisy channels are masked by the 3PG analysis, I have commented out the code that used to do that here. (Since the trim macro has no knowledge of the gain, it could confuse high gain channels with the noisy guys....)
  • Two stage calculation of mean and rms of offset and step. Outlyers now excluded from the calculation - makes defect identification more accurate. Also the rms and mean values now give a more accurate reflection of the performance of the trimmed channels, since the untrimmable channels are excluded from the calculation. (Code based on work by Carlos.)
  • Bugfix regarding handling of data from STUCKCELL modules (or, more accurately, datasets where the vt50 histogram contains unphysical entries with large negative numbers...)
  • Bugfix: protection against divide by zero error in calculate_mean_rms. (Xrms can be zero on the second iteration if a chip is dead!)
  • Bugfix: use local array for p1 on entry to calculate_mean_rms. This allows p1 to be set -ve to exclude a known untrimmable channel from the calculations for a single trim range without affecting the trim optimisation for trim range -1. (This bug, left untreated, could lead to channels being erroneously masked at the end of the test sequence.)

• sctdaq/stdll/Makedict.bat
  • add TLogger

• sctdaq/stdll/makefile
  • Add TLogger (thanks, Carlos!)

• sctdaq/stdll/stdll.dsp
  • add TLogger to project

• sctdaq/stdll/TModule.cpp
  • Low Icc and/or Idd values (indicative of ABCD chips which are not clocked and/or configured) are now shown in magenta

• sctdaq/stdll/TModule.cpp sctdaq/stdll/TModule.h
  • Show currents in black when a module has been suspended
  • Modified to book and fill hits per link histograms, after work at LBL by Robert Ely and Allen Haim. I have rewritten their code such that no changes were needed within TST.cpp, this means that the hits per link histogram will now be filled for all burst types which use software histogramming.
  • Changes to the implementation of the "hits_per_link" histograms. In the original implementation, a static array was used to compile the histogram, the ROOT histogram being filled upon completion of the scan. This implementation caused malloc problems under Linux. The relevant static arrays have been deleted and the ROOT histograms are now filled directly. (Thanks to Carlos for reporting this problem!)
  • Masked channels are now excluded from AUTOSTOP calculations. On the assumption that the 3PG test sets the mask bit for STUCKCELL channels, they will no longer break the AUTOSTOP calculations with the consequential increase in the time taken to test a hybrid/module.
  • Channels which are masked are now skipped by the TModule::FitScan method. Hence channels with unmaskable occupancy will no longer affect the mean (chip) Vt50 calculation.

• sctdaq/stdll/TModuleDCS.cpp
  • bugfix re. suspended modules (need only suspend once!)
  • bugfix - SetLog method was missing
  • bugfix - show correct value of VDD_OUT
  • Make default values of T1, T2_minimum -128 to disable low T trip

• sctdaq/stdll/TST.cpp sctdaq/stdll/TST.h
  • Add LogEvent method (used to log events from interpreted tests)
  • Revise calls to ExecuteSCConfigs()
  • "burst_max_timeouts" is now declared as public instead of private. This means that the number of timeouts may be reduced if required to speed up a scan that will be prone to such failures, for example a scan of AERO thresholds (VDCDAC). "Burst_max_timeouts" has also been added to SaveState and RestoreState, so the original value may be recovered after a scan has completed.
  • Trivial, unimportant optimisation - use a known value for the number of active channels during AUTOSTOP calculations as part of the NO test. Probably faster than working it out again each time.
  • bugfix re. masked channels and AUTOSTOP (recalculate number of enabled channels at start of each scan)

• sctdaq/stlib/st_config.c
  • Additions to look for station specific default file (as requested by NIKHEF). Priority now: 1) Module specific file 2) Station specific default file 3) Generic default file

• sctdaq/stlib/stlib.c
  • Reorder startup sequence such that AERO is configured before LV power is applied (ensure clock or BPM signal present before power on)
  • Revise delays in st_execute_sc_configs: external delays are no longer needed

• sctdaq/macros/NoLBL.cpp
  • Version of NoPlot.cpp to include LBL's hits per link test. Will run DecodedVBurst if CLOAC present, DecodedBurst if CLOAC missing.

• sctdaq/macros/NoPlotLBL.cpp
  • LBL's version of NoPlot.cpp, unchanged from a version supplied by Allen Haim on 19th August 2003. (May be some memory leaks.)

• sctdaq/macros/ReceptionTest.cpp
  • Prototype Test Sequence for use at Macroassembly sites (based on requirements of Oxford)

• sctdaq/macros/STIV.cpp
  • Start ROOT with the argument "STIV.cpp" to run an IV curve without further intervention.

• sctdaq/macros/StrobeDelayPlot_old.cpp
  • old version

• sctdaq/macros/VDCDACPlot.cpp sctdaq/macros/VDCDACScan.cpp
  • Macros for setting up AERO thresholds. Contributed by Scott Lindsay. Minor revisions by PWP.

• sctdaq/sctlv/lv_watchdog.c lv_watchdog.dsp lv_watchdog.dsw
  • New "watchdog" application for users of SCTLV3. Designed to run in the background, this simple app will turn off the power if things get too hot. (This is in addition to the overtemperature trip code now implemented in TModuleDCS.cpp.)

• sctdaq/stdll/TLogger.cpp sctdaq/stdll/TLogger.h
  • New class implementing event logging to the text file strunXXX.log. A useful aid to the understanding of LV trips etc.

Additions/fixes/improvements:

• sctdaq/config/default_cold.det default_warm.det
  • For EndCap tests using AERO, example detector configuration files for use warm and cold. These may need to be adjusted to run reliably with other test systems. See EndCapHLTT for an example showing how one may run an automated test sequence at warm and cold temperatures.

• sctdaq/macros/AutoChar.cpp AutoConf.cpp
  • Simple examples of macros to run a test sequence, archive the data and trun off the power.

• sctdaq/macros/EndCapHLTT.cpp
  • Slightly revised timing
  • Option to use different default detector configurations (warm and cold)

• sctdaq/macros/FullBypassTest.cpp
  • Automatically select 62 configurations if Endcap devices have been declared in the sysmap
  • Option to use AERO mode 2 for this test only (off by default)
  • When use of AERO mode 2 is selected, the stream delay is changed to suit. (NB. The temporary value of stream delay is hardcoded in the macro and may require user adjustment to make it work.)

• sctdaq/macros/HardReset.cpp
• sctdaq/macros/LTPostRadTest.cpp
• sctdaq/macros/NoScurves.cpp
• sctdaq/macros/StrobeSpeed.cpp
• sctdaq/macros/T7IccVsFEScan.cpp
• sctdaq/macros/T7RadContTest.cpp
• sctdaq/macros/TLTT.h
• sctdaq/macros/ToolBox.cpp
• Increase size of filename arrays to 128 chars (from 64)

• sctdaq/macros/IVCurve.cpp
  • If mean T1 across all present modules is less than 10C, the current limits are reduced. By default, the "cold" warn and trip limits are set to hardcoded values of 1and 2 microA respectively. Code is also provided to scale the limits properly according to the measured temperature - however this is presently commented out. (The "warm" warn and trip limits are presently 5 and 10 microA respectively.)

• sctdaq/macros/ResponseCurvePlot.cpp
  • Bugfix re. leaving (n-1) PS files open when testing n modules
  • Any CH_OVER channels listed as DB COMMENTS.
  • Main defect loops changed to be robust against CH_OVER channels
  • CH_INEFF channels were only listed as defects for RC test (ie not 3PG), except for the last channel of a link. Fixed.
  • CH_INEFF were always included in the defect tally. Now included for RC test only.

• sctdaq/macros/parameters.h
  • #define CH_OVER

• sctdaq/macros/TLTT.cpp, sctdaq/macros/TLTT.h
  • PS filename and corresponding web link now include host name
  • label all LTT output filenames with start_scan
  • close stop window if test ends due to alarm condition
  • fix s/n labels on postscript window to allow non-consecutive module numbers

• sctdaq/SCTLV/lvtest.c
  • Add new option for "Power Cycle Test" - switches card on and off 100 times, or until fails to start.

• sctdaq/SCTLV/sctlv_lib.c
  • Add propogation of rd/wt flag timeout error codes to sctlv_set_power and subroutines thereof
  • Add the corresponding error codes to sctlv_print_error

• sctdaq/stdll/TModuleDCS.cpp
  • Print method now selects which parameters should be shown according to the connected hardware (SCTLV2 or SCTLV3). In each case, all relevant parameters are shown (inc Vpin, Ipin in case of LV3).

• sctdaq/stdll/TModuleDCS.h
  • Add vcc_out and vdd_out
  • Increment class version number

• sctdaq/stdll/TModuleDefine.h
  • remove CH_OVER from BAD_TEST (such that CH_OVER channels are fitted)

• sctdaq/stdll/TST.cpp sctdaq/stdll/TST.h
  • Three new methods: GetNBarrels, GetNEndcaps, GetNUnkown

• sctdaq/stdll/stdll_cint.cpp sctdaq/stdll/stdll_cint.h
  • Rebuilt following changes

• sctdaq/stlib/sct_scans.h
  • Declare ST_MON_VCCOUT and ST_MON_VDD_OUT

• sctdaq/stlib/st_config.c sctdaq/stlib/st_config.h
  • Three new functions: st_n_barrels, st_n_endcaps, st_n_unknown. They return the corresponding number of devices of that type that have been declared in the sysmap

• sctdaq/stlib/stlib.c
  • Add support for vcc_out and vdd_out in st_retrieve variable
  • Trivial bugfix to st_execute_sc_configs regarding subsequent reloading of SLOG with selected trigger pattern
  • make HV ramp loop robust against "timeouts" (prevents the software getting stuck in a loop in the event of HV card failure)

Additions/fixes/improvements:

• sctdaq/macros/FullBypassAnalysis.cpp
  • Shorten comment "Failing Maximal tested Vdd" to "Fail max tested Vdd" (the original comment was too long for SCTDB 20 char limit)

• sctdaq/macros/IVCurve.cpp
  • Changes following identification of inconsistencies with regard to the implementation of SCTDB result files:
    • Record current as 99999999 if warning level exceeded at an earlier data point (and hence voltage frozen)
    • Record 99999999 instead of -1 if 150V or 350V not tested due to STOP or ABORT
    • Add DB comment if test aborted by user
  • Extra menu option (0 to 10V) for diagnostic tests

• sctdaq/macros/parameters.h
  • Raise version number to 3.39

• sctdaq/macros/ST.cpp
  • Call e->UseAero at system startup to enable configuration of AERO. (This will not interfere with other readout options as long as SCTLV is in use: the AERO configuration packet is followed by HARD RESET. Users of other power supply systems should comment this line out.)

• sctdaq/SLOG/slog_lib.c
  • Modify slog_init as follows: External commands are enabled if CLOAC present. External commands are disabled if CLOAC is missing. (Fixes hang of AERO due to burst of logic 1 at system startup when running without CLOAC due to transmission of "missing" external command input)

• sctdaq/stdll/TModule.cpp
  • Trivial bugfix re. minimum Icc or Idd current shown on scan canvas (relevant only where the minimum Icc or Idd during a scan exceeds 1000mA)

• sctdaq/stdll/TST.cpp sctdaq/stdll/TST.h
  • Add method TST::UseAero() to set do_aero flag to 1
  • Raise version number to 3.39

• sctdaq/stlib/stlib.c sctdaq/stlib/stlib.h
  • Modify st_startup to configure AERO if do_aero = 1 (default = 0)
  • Add function st_set_aero to modify value of do_aero flag

Additions/fixes/improvements:

• sctdaq/analysis/fC2mV.cpp stan_fitrc.cpp stan_util.cpp
  • new expression for "Grillo" function in y after Sandra Muijs

• sctdaq/analysis/mv2fC.cpp
  • added expression for "Grillo" function in x after Bill Murray

• sctdaq/macros/NoPlot.cpp
  • Cosmetic fixes
  • Increase size of h_name array to avoid bound overflow
  • Added expression for "Grillo" function in x after Bill Murray
  • Fixed expression for polynomial in x
  • Removed redundant code

• sctdaq/macros/ResponseCurvePlot.cpp
  • Replace INEFFICIENT defect by INEFF to match SCT DB
  • Trivial bugfix contributed by Dave Robinson
  • Mask HIGH OFFSET channels after 3PG tests. Just like "STUCK" channels, this type of defect drives the trimming towards the high end of the tested range. In turn this makes all channels susceptible to artefacts (such as increased noise) due to the saturation of the threshold DAC at high values, notably at low temperatures. In short, they cause more trouble that they are worth, so let's just mask them. This is probably the correct action to deal with "one bad channel" chips, but this decision will be reviewed in the near future.
  • New expression for "Grillo" function in y after Sandra Muijs.
  • New code to indetify and mask channels with VERY low gain (CH_VLO_GAIN). The cut is at 0.3* chip_mean_gain, so all channels masked by this cut would be essentially of no use in ATLAS. Such channels are still listed as LO_GAIN in the SCTDB but will be masked off at end of 3PG test to speed testing. DB comments are always generated when channels are masked by the 3PG anlaysis.
  • Also mask DEAD channels (as distinct from ones with very low gain)

• sctdaq/macros/ST.cpp
  • Fixes regarding st_dcs_log and st_dcs_display (update DCS info for all modules, not just those on display)

• sctdaq/macros/TLTT.cpp
  • call st_stop_stopper if exiting due to alarm
  • Print module names in the event of error conditions
  • Attempt to resynchronise comms before exiting due to current alarm (last ditch recovery from communication errors)
  • Revised to better match the needs of the endcap community. Alarm temperatures and currents are now supplied as arguments to LTT methods, to facilitate testing at a variety of different temperatures. I also took the opportunity to tidy up the code, removing redundant sections.
  • Adjust the length of the last big_loop such that the test duration will be correct.

• sctdaq/macros/TLTT.h
  • Change type of total_length from Int_t to Long_t.

• sctdaq/macros/CharacterisationTest.cpp
  • return int representing duration of test (in seconds)

• sctdaq/macros/HybridColdTest.cpp
  • Now uses TLTT::HybridLTT instead of the removed TLTT::HybridColdTest

• sctdaq/macros/parameters.h
  • New #define statement (CH_VLO_GAIN)

• sctdaq/macros/RedundancyTest.cpp
  • The test is now skipped if SELECT==1 for each device under test (consistent with readout of endcap hybrids/modules without opto)

• sctdaq/macros/DBUpload_Barrel_Hybrid.cpp
  • Contibuted by Dave Robinson

• sctdaq/macros/EndCapHLTT.cpp
  • "All-in-one" macro for warm and cold tests of endcap hybrids
  • Trim and mask files are copied across beofre the main LTT segment

• sctdaq/macros/Toolbox.cpp
  • New routines to facilitate the copying and/or deletion of trim and/or mask files to/from the config directory.

• sctdaq/macros/systest/ScurveFit.cpp
  • New expression for "Grillo" function in y after Sandra Muijs

• sctdaq/scthv/scthv_lib.c scthv_lib.h
  • add st_badvolts as sanity check on voltage setting
  • add st_badamps as sanity check on current measurement
  • Adjusted timeout settings and debug statements
  • Handle monitor timeouts internally

• sctdaq/stdll/functions.cpp
  • new expression for "Grillo" function in y after Sandra Muijs

• sctdaq/stdll/TModule.cpp
  • Cosmetic fixes

• sctdaq/stlib/stlib.c
  • Check for badvolts during HV ramping
  • scthv monitor timeouts are now handled internally by hv library
  • new expression for "Grillo" function in y after Sandra Muijs
  • added expression for "Grillo" function in x after Bill Murray
  • added expression for exp function in x
  • fixed expression for polynomial in x

• sctdaq/stlib/st_config.c st_config.h
  Additions to support use in "enter serial numbers at prompt mode"
  • serial number validity checking (length and "2022")
  • long and short endcap middle modules

Additions/fixes/improvements:

• sctdaq/macros/IVCurve.cpp
  • Modify "IV_LIMIT" state to be a SW trip state (HV will be set to 2V if 10microA current limit exceeded)
  • Also add diagnostic ramp, up to 50V in 5V steps

• sctdaq/macros/IVLTT.cpp
  • Generate dummy scan data (strun.root) file before starting test (needed to provide tags for DB output file)

• sctdaq/macros/NO.cpp
  • For users without CLOAC, use new burst type NSEBURST2 (17) to give faster NO measurements

• sctdaq/macros/parameters.h
  • #define CH_BADTEST (OR of all defined channel defects)
  • #define CH_MASKTEST (= low gain AND high noise)
  • Raise version number to 3.36

• sctdaq/macros/ResponseCurvePlot.cpp
  • Important bugfix re. DB defect listing for hybrids: DEAD or LOWGAIN/normal noise channels were not listed as defects by previous versions (module tests unaffected by this)
  • Mask channels having BOTH low gain AND high noise

• sctdaq/macros/ST.cpp
  • Adjustments for new burst type NSEBURST2 (see above)

• sctdaq/macros/TLTT.cpp TLTT.h
  • Re-read currents before exiting due to HV overcurrent alarm (provides some resilience against "one-off" bad readings)

• sctdaq/stdll/TModule.cpp
  • Fix autostop for scans of ST_TTHR (threshold wrt trim target)

• sctdaq/stdll/TST.cpp TST.h
  • New burst type NSEBURST2 (see above)

• sctdaq/scthv/scthv_lib.c
  • Handle "corrupt monitored volts" within scthv_monitor()
  • Include safety timings within scthv_repeat_request()
  • Replace delay loop with timing based on usleep
  • For windows, include implementation of usleep as static function

• sctdaq/scthv/scthv_lib.h
  • Adjust values used to identify "stuck" condition

• sctdaq/stlib/st_config.c
  • Trivial bugfix to st_write_config_xml re. months
  • Also now write trim target to xml config file

• sctdaq/stlib/stlib.c
  • Adjustments to accomodate revised SCTHV library, notably the removal of safety delays now moved to the lower level.

Additions/fixes/improvements:

• sctdaq/macros/ConfirmationTest.cpp
  • Added missing brackets

• sctdaq/macros/parameters.h
  • Raise version number to 3.34

• sctdaq/macros/ResponseCurvePlot.cpp
  • Three small bugfixes
  • List INEFFICIENT channels only for the ResponseCurveTest. (The 3PG test may be run untrimmed, in which case channels may have less than 100% efficiency due to large discriminator offsets. Such channels, if trimmable, do not have permanently reduced efficieny and hence should not be listed as defects.)

• sctdaq/macros/StrobeDelay.cpp
  • Comment out MAX_MODULES (clashes with parameters.h)
  • Change type of some variables from Int_t to Float_t to avoid type conversions

• sctdaq/macros/StrobeDelayPlot.cpp
  • Move declaration of TPad* inside if(StrobeDelay)

• sctdaq/macros/LTPostRadTest.cpp T7RadContTest.cpp systest/Accumbursts.cpp systest/idetmon.cpp
  • Use of long(gSystem->Now()) was not reliable in interpreted code. Now use e->Now() instead (returns type long).

• TLTT.cpp TLTT.h
  • Use of long(gSystem->Now()) was not reliable in interpreted code. Now use e->Now() instead (returns type long).
  • Added code such that the Burst Canvas DCS panel will be updated every 10 seconds during idle periods.

• sctdaq/stdll/TST.cpp TST.h stdll_cint.cpp stdll_cint.h
  • Added method Long_t TST::Now() as wrapper to gSystem->Now()

Additions/fixes/improvements:

• sctdaq/macros/FullBypassAnalysis.cpp
  • Changes with regard to PASS/FAIL status and defect identification:
    • set PASS if works at 4.0V
    • set PROBLEM if fails at any tested voltage
    • record DEFECT for any token link failing at any voltage
    • record COMMENT for each DEFECT EITHER giving the minimum Vdd needed for that link to work OR giving a message that the link fails at all tested Vdd

• sctdaq/macros/HybridColdTest.cpp sctdaq/macros/HybridLTT.cpp sctdaq/macros/IVLTT.cpp sctdaq/macros/ModuleLTT.cpp
  • As appropriate, the characterisation sequence is now called and the power is turned off from each of these macros, rather than TLTT.

• sctdaq/macros/IVCurve.cpp
  • Trivial bugfix

• sctdaq/macros/No.cpp
  • New version to adjust the scanned range of each module to take into account trim differnces by means of the target used during trimming. (The line to run the new version of the test is there, but it is presently commented out.)

• sctdaq/macros/NoPlot.cpp
  • TAxis* axis was declared twice.
  • if scanned range minimum is less than -10, new algorithm is assumed in which case markers are adjusted to add on the trim target (NB. graphs will still be labelled with negative numbers)

• sctdaq/macros/parameters.h
  • Raise version number to 3.33
  • Add MAX_BADSTRIPS and MAX_CONSECUTIVE_BADSTRIPS

• sctdaq/macros/ResponseCurvePlot.cpp
  Set FAIL if:
  • any chip DEAD, STUCK, INEFFICIENT, LOW GAIN or HIGH GAIN
  • more than MAX_BADSTRIPS (15) bad strips
  • more than MAX_CONSECUTIVE_BADSTRIPS (8) consecutive bad strips
  In each case, an appropriate DB comment is added to the results file. The bad channel identification and counting code has also been generally tidied up.

• sctdaq/macros/systest/sleep.cpp
  • rewritten to use TSystem::Sleep() instead of TStopwatch::GetRealTime() (ROOT 3.03 compatibility fix)

• sctdaq/macros/LTPostRadTest.cpp T7RadContTest.cpp sctdaq/macros/systest/AccumBursts.cpp idetmon.cpp
  • rewritten to use TSystem::Now() instead of TStopwatch::GetRealTime() (ROOT 3.03 compatibility fix)

• sctdaq/macros/TLTT.cpp
  • Set abort flag in event of alarm
  • do not run characterisation test at end of sequence
  • do not turn off power at end of sequence
  • rewritten to use TSystem::Now() instead of TStopwatch::GetRealTime() (ROOT 3.03 compatibility fix)

• sctdaq/macros/TrimRangePlot.cpp
  • set trim range value to SCTDAQ if run online
  • record trim target in trim file

• sctdaq/stdll/TModule.cpp
  • if all bins have the exact same occupancy, the channel is marked as STUCK (in addition to STUCK if all bins >98%)

• sctdaq/stlib/abc.h sctdaq/stlib/sct_scans.h
  Two new variables:
  • ST_TARGET - record of trim target
  • ST_TTHR - threshold in mV set wrt trim target

• sctdaq/stlib/st_config.c
  • Changes to initialise trim target to zero
  • load trim target from trim file
  • print trim range and target to screen in st_show_trims

• sctdaq/stlib/stlib.c
  • handle ST_TARGET and ST_THR

• sctdaq/stdll/TST.cpp
  • Add description of ST_TTHR and default scan

This PRODUCTION version is built to work with version 3.02/07 of ROOT (WITHOUT experimental GDK driver). It may also be built under Linux to run with ROOT version 3.02/07.

Additions/fixes/improvements:

• sctdaq/macros/CharacterisationTest.cpp
  • Do replay in BATCH mode - faster, no plots to screen, avoids memory leaks in TWin32Canvas

• sctdaq/macros/ConfirmationTest.cpp
  • Do replay in BATCH mode - faster, no plots to screen, avoids memory leaks in TWin32Canvas
  • Trivial bugfix re. Confirmation replay

• sctdaq/macros/NoPlot.cpp
  • If used online, use BATCH mode - avoids memory leaks in TWin32Canvas. Plots go by so quickly, they cannot be read in any case.
  • Make reference to gPad immediately after creation of TCanvas - improves stability when used offline

• sctdaq/macros/NoScurves.cpp
  • Close canvas at end
  • Close strun at end
  • Memory leak fixes

• sctdaq/macros/IVCurve.cpp
  • Display updated DCS information on the burst canvas at each point of the test

• sctdaq/macros/NoPlot.cpp
  • Make reference to gPad immediately after creation of TCanvas - improves stability when used offline

• sctdaq/macros/parameters.h
  • Raise version number to 3.31

• sctdaq/macros/TimewalkPlot.cpp
  • If used online, use BATCH mode - avoids memory leaks in TWin32Canvas. Plots go by so quickly, they cannot be read in any case.
  • Fix TText memory leak