GLAST LAT Project CERN Testbeam Work Shop Pisa
GLAST LAT Project CERN Testbeam Work. Shop, Pisa, 20 March 2006 CAL performance checks • A few key CAL features have barely been exercised in “real life”. • Persons most qualified to study these : Ø have visa issues (Sasha Chekhtman), Ø are busy getting the LAT out of SLAC Ø are preparing for environmental testing Ø or some combo of the above (Eric Grove), (Neil Johnson), (Mark and Byron and…). • Sasha & I propose to work together: I get & check the data sets Sasha needs, Sasha does the brain work. • Our reading of the Beam Test Plan (“BTP”) suggests that no additional runs are needed, provided some non-flight configurations are foreseen. • This talk: Sasha’s 5 favorite studies, and the data samples they require. David Smith (for the NRL gang) 1 CAL hardware tests
GLAST LAT Project CERN Testbeam Work. Shop, Pisa, 20 March 2006 Sasha’s 5 CERN studies 1. FHE thresholds –Could use the in-beam calibration runs (BTP § 10. 2, figure 1 left) but with diagnostics turned on and varying the trigger mask. 2. Range intercalibration – if four-range readout for § 10. 2 then okay. Else, pre -scaled separate trigger engine during SPS runs. 3. Test autorange mode (ULD performance) – any or all HE data sets. 4. Non-linearity correction – check charge injection result on scintillation data. Use SPS 90° data for good containment and resolution (BTP § 11. 2. 1, page 16). Avoid hitting diodes –> aim at Cs. I long edge. 5. Preamp reset – use trigger engine with 10 Hz periodic readout to study pedestal shifts after big energy deposits. David Smith (for the NRL gang) 2 figure 1 -left, BTP page. CAL 14. hardware tests
GLAST LAT Project CERN Testbeam Work. Shop, Pisa, 20 March 2006 Sasha’s Work. Shop. Six® Talk David Smith (for the NRL gang) 3 CAL hardware tests
GLAST LAT Project CERN Testbeam Work. Shop, Pisa, 20 March 2006 ( A remark ) • At GSI, the CAL EM had AFEEs 1 using GCFE 2 v 7 on two faces and v 9 on two other faces. CERN will be first high energy test of truly flight version (to say nothing of TEM and GASU). • The different settings discussed in this talk are all inside the GCFEs. 1) Analog Front End Electronics circuit board, one per face 2) Glast Calorimeter Front-end Electronics = asic containing the ADCs, DACs, LACs ULDs et cetera, one per crystal end David Smith (for the NRL gang) 4 CAL hardware tests
GLAST LAT Project CERN Testbeam Work. Shop, Pisa, 20 March 2006 FHE threshold study • First time we’ll see >1 Ge. V per crystal • Want ratio of energy histograms, with and without trigger requirement diagnostics must be ON). 3 different threshold settings is nice. . • In CAL, trigger info = sum of ends per layer. Standard “mu. Trg” method: run with even columns enabled, run again for odds. Takes three hours for muons, per threshold setting. • Issue: Lex 1 -> Hex 8 ULD threshold is near 1 Ge. V, so configure to force Hex 8 readout (simplifies analysis). • Sasha thinks this can be done during the calibration scans (readout unaffected by trigger masks), but the scripts would need some work. • Low statistics near shower tail? Try 12 runs, each with 1 column only enabled? David Smith (for the NRL gang) 5 (including Eduardo’s Work. Shop. Six study). (trigger CAL hardware tests
Range Inter. Calibration, Autorange Test GLAST LAT Project CERN Testbeam Work. Shop, Pisa, 20 March 2006 v Range intercalibration – at SLAC: muon calibrations of Lex 8, and charge injection to propagate to Lex 1, Hex 8, Hex 1. • at GSI: 4 -range readout (non-flight…) to propagate. • Since CERN = flight configuration, suggest a trigger engine with 4 -range readout, using CAL HI with pre-scale to keep deadtime low. A 600 Me. V ps g, simulated by Ph. Bruel, displayed with R. Giannitrapani’s “FRED”. 1. Autorange (ULD performance) 2. At GSI, sometimes (+end/-end)~8. 3. i. e. , recorded range used. 4. Need to test at high rate & large energy deposit. 5. Any HE data set in “autorange” mode (is flight config, so no problem). David Smith (for the NRL gang) 6 CAL hardware tests
GLAST LAT Project CERN Testbeam Work. Shop, Pisa, 20 March 2006 Non-linearity corrections • The ADCs and the charge injection (=ci) DACs are not linear. The calib. Gen script suites vary ci pulseheights to map the response. • But ci pulse shape scintillation pulse shape. Doubts persist, particularly for the Hex 8 range (1 to 8 Ge. V per crystal). • The HE SPS 90° run already planned provides the data set. • Simulations required to test our sensitivity to these “problems”. • Sasha speculates that crystal-to-crystal fluctuations will smooth out the effects, so that the overall non-linearity in e. g. Cal. Energy. Sum will be less than the ~5% single crystal non-linearity. David Smith (for the NRL gang) 7 CAL hardware tests
GLAST LAT Project CERN Testbeam Work. Shop, Pisa, 20 March 2006 Pre-amp reset • After big energy deposits, baseline restoration is slow (rc~100 u. S). • For E> ¼ Lex 1 range, a “reset” circuit kicks in. • Supposed to be always “on”. A script bug at GSI had it “off” except for one carbon run => only 0. 5 Ge. V/crystal => not really tested. • Sasha’s wish: a trigger engine providing 10 Hz periodic triggers with 4 -range readout, to look at baseline recovery (pedestal shifts) after large energy deposits. • Sasha further hopes that an external trigger scaler will count during event readout (during deadtime), so that time-since-last-big-event info will be available. • (He plots pedestals vs Gem. Delta. Event. Time). David Smith (for the NRL gang) 8 CAL hardware tests
GLAST LAT Project CERN Testbeam Work. Shop, Pisa, 20 March 2006 Conclusions • We have listed a few studies which will help get on-orbit CAL adc-to -Me. V conversions as-right-as-possible. • One would use trigger masks not foreseen outside of the non-beam mu. Trg runs, and have trigger diagnostics enabled. • Two would use Latte 4’s multi-trigger engine capability. • Ready to discuss which deviations from single engine, pure flightconfiguration the testbeam can support. David Smith (for the NRL gang) 9 CAL hardware tests
- Slides: 9