GLAST Large Area Telescope: EM1/TKR Test Results Summary

Transcription

1 GLAST Large Area Telescope: EM1/TKR Test Results Summary Gamma-ray Large Area Space Telescope Hiro Tajima SU-SLAC I&T TKR Test Manager (representing the LAT EM/TKR Crew) Hiro Tajima, EM/TKR Test Status 1

2 Overview EM/TKR characterization Gain Noise Noise Occupancy Trigger timing Problems GTRC time out error GTRC failure Detector Biasing Cable mechanical interference Hiro Tajima, EM/TKR Test Status 2

3 Gain Gain measurement involves threshold DAC, calibration DAC and calibration capacitor in addition to the gain itself. Average gain: 101 mv/fc Stability: <0.1% change after 10 days, 0.3% change with new TEM-PS Gain dispersion within GTFE chip: 4 7% SPEC: < 10% (LAT-SS ) Keep track of dead GTFE channel. Gain vs Strip# ( Layer X1) Gain (mv/fc) Dead electronics channels Strip # Hiro Tajima, EM/TKR Test Status 3

4 Noise Average Noise: ~1500 electrons (ENC) Stability: 0.5% change after 10 fays. 0.6% change with new TEM-PS. Effective way to find open channels. Missing wire-bonds, broken pitch adapter traces. Perfect match with pre-ship review data: Nothing new. Fraction: 0.5% (SPEC: efficiency > 98%, LAT-SS ) 2.7% (includes mechanical problems with pitch adapter) Gain vs Strip# ( Layer Y3) Open channels Strip # Hiro Tajima, EM/TKR Test Status 4 Noise (electrons)

5 Noise Occupancy Hot channel is defined as occupancy > 10-4 (LAT-SS ) Fraction: <0.1% 6.7% (includes dead/noisy ladders due to bias problems.) Stability: 9/17: six hot channels. Five hot channels are subset of pre-ship review list. One new hot channel. 15 hot channels disappeared from pre-ship review list. 9/29: six hot channels. All hot channels are subset of pre-ship review list. 14 hot channels disappeared from pre-ship review list. Occupancy for most of disappeared hot channels are less than Hiro Tajima, EM/TKR Test Status 5

6 Timing Studies with Calibration Strobe Find optimum TACK (trigger acknowledge) delay for each strip. No significant propagation delay is observed. Large dispersion: ±0.5 µs SPEC: peaking time = 1.5 ± 0.5 µs (LAT-SS ) This is due to dispersion of fall time (not peaking time). After TKR/CAL integration (TEM is also changed), 0.16 µs shift of optimum TACK delay is observed. Further investigation. Significantly different fall time Strip # Hiro Tajima, EM/TKR Test Status 6 Optimum TACK delay (µs) 24/0 split

7 TOT Gain Dispersion Large dispersion observed for TOT gain. TOT Gain (µs/fc) TOT Gain: D(TOT)/ D(Injected Charge) Large dispersion observed Strip # TOT (µs) TOT (µs) Injected Charge (fc) Injected Charge (fc) Hiro Tajima, EM/TKR Test Status 7

8 More on TOT Gain Correlation with Gain measurement? Gain (mv/fc) TOT Gain (µs/fc) Strip # GTRC dependence? TOT Gain (µs/fc) TOT Gain (µs/fc) Strip # 24/0 split 0/24 split FE # FE # Hiro Tajima, EM/TKR Test Status 8

9 Trigger Jitter Studies with Cal Strobe Trigger jitter studies using calibration strobe. Trigger timing is defined as Tack delay at which trigger efficiency is 50% for a certain threshold and Cal DACs. Little meaning for absolute value. Trigger jitter is the dispersion of the trigger timing. This is not pure trigger jitter since it includes effects from threshold and gain dispersions. Trigger is measured to be less than ±0.25 µs. SPEC: trigger jitter < 0.25 µs (LAT-SS ) Trigger timing (µs) Strip # Hiro Tajima, EM/TKR Test Status 9

10 Trigger Timing Studies Relative Hit Multiplicity Hit multiplicity vs TACK delay TKR trigger CAL trigger EXT trigger TKR and CAL trigger timing difference is less than 0.2 µs. Further timing delay due to GASU might affect strip efficiency TACK delay (µs) Hiro Tajima, EM/TKR Test Status 10

11 Efficiency vs.. Tack delay 100.0% 99.5% Efficiency vs TACK delay Efficiency for TKR trigger is data capture efficiency. TKR trigger CAL trigger EXT trigger Efficiency 99.0% 98.5% TACK delay up to µs may be acceptable. 98.0% TACK delay (µs) Hiro Tajima, EM/TKR Test Status 11

12 Position Dependence of Efficiency (1) Efficiency vs. Track position TACK = 0 Efficiency Inefficient ladder? Track position (mm) Hiro Tajima, EM/TKR Test Status 12

13 Position Dependence of Efficiency (2) Efficiency vs. Track position TACK = 2.5 µs Efficiency Inefficient ladders? Track position (mm) Hiro Tajima, EM/TKR Test Status 13

14 Ladder Dependence of Efficiency 100% Efficiency vs TACK delay (Layer 1) 100% Efficiency vs TACK delay (Layer 2) 100% Efficiency vs TACK delay (Layer 3) 99% 99% 99% Efficiency 98% 97% X-L0 X-L1 X-L2 Y-L0 Y-L1 Y-L2 Efficiency 98% 97% X-L0 X-L1 X-L2 Y-L0 Y-L1 Y-L2 96% 96% 96% TACK delay (µs) TACK delay (µs) Efficiency 98% 97% X-L0 X-L1 X-L2 Y-L0 Y-L1 Y-L2 TACK delay (µs) Statistical error is % Hiro Tajima, EM/TKR Test Status 14

15 GTRC Timeout Problem with TOT Y0 seems to be more susceptible to the problem and unstable. Disabled Y2 completely. Use X1 single layer trigger to understand real TREQ/TACK rate. Threshold Trigger Rate (Hz) # of errors Time (sec) Error Rate x x x10-6 Error rate seems independent of trigger rate. Average TOT is about 9 µs for all cases. No timeout error from the GTRCs in X1 layer. TREQ and TACK is well synchronized. A lot of cable controller phasing errors if TACK>0 Could be a problem with further delay due to GASU and global trigger. Hiro Tajima, EM/TKR Test Status 15

16 Issues GTRC time out error due to a bug in TOT. Very rare with TOT disabled. (twice in 10 days.) Further test with higher track rate. GTRC/GTFE testing. One GTRC chip failed at 2.4V (SPEC: 2.5 ± 0.25 V) Broken wire bond for SSD bias at SLAC. No handling issue. Double or triple wire-bond for SSD bias pads. Note no new broken wire-bond for strips observed. Cable mechanical interference. Test scripts are not ready for flight models. Requires at least one month. Hiro Tajima, EM/TKR Test Status 16