Transcript ppt
Trigger Anomaly in BFEM? February 13, 2002 Tsunefumi Mizuno and Tune Kamae History (1) Discrepancy between BFEM data and G4 simulation persisted even after reasonably thorough validation G4. “Charged” counts in layers (Run55): Blue:BFEM data, Red: G4 simulation “Neutral” counts in layers (Run55): Blue:BFEM data, Red: G4 simulation History (2) With help of Tony and Heather, we retrieved the layer numbers participated in the 3-in-row trigger. The methods implemented are getXCapture() and getYCapture (). The layer count distribution (see below) indicates that layer 7, 14 and 15 didn’t participate in L1T. History (3) We then removed the 3 layers from the G4 simulation of 3-in-row trigger. Agreement between BFEM data and G4 simulation improved greatly for “charged” events (the flux was increased by approx. 20%) but not for “neutral” events. History (4) Another minor error was found by Dave Lauben: TrigXCapture and TrigYCapture seem to be swapped. The conclusion in the previous slides remain unchanged. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 --------------------------------------------------------------------------------------swapped o o o o o o o o o o o o o o o o o x x o o x x o x o Hit o o o o o o o x o o o o o o x x x o o x x o o x o x Trig not X7, but Y7(layer#15) not Y3, but X3(layer#6) not Y7, but X7(layer#14) Trigger efficiency of each layer (1) Further study showed apparent “inefficiency” in all layers. Blue histogram below shows counts in data. Red histogram is counts in fast capture. They disagree. Hit in the data Trigger fast capture Trigger efficiency of each layer (2) Select events where layers 0-5 are in the fast capture. Fast capture for Layer 23 seems to be only 50% efficient. Hit Trigger Possible Interpretation (1) Possibility 1) Gate width for “fast capture” (approx. 1us) is too short compared to the signal slew time of SSD, the propagation time through the readout chips in a layer, and variation in the trigger condition compliance. Possibility 2) Some layers are noisier than others and “data” may include accidental hits. Possiblity 3) Some layers are noisier than others and accidental hits prior to the event caused dead time to the fast signal that participated in the trigger. A study on fast capture efficiency Study was made if we see any events where no 3-in-row is found in fast capture. All events registered >=3 in-row. We can conclude that fast capture worked for the layers participated in the trigger. A study on frequency of accidental hits Events where a charged-particle passes through all layers were selected. The selection criteria are as follows 1) one or more ACD tiles show PHA above 0.2MIP (charged events) 2) Single track (number of tracks reconstructed is 2 -- x and y) 3) The track is straight (reconstructed value of chi^2 is below 0.1) 4) All 26 layers are hit. Eye-scanned 300 events by the Event Display and counted the number of hits that do not belong to the straight track. The percentage of is only 1-2%, well below the apparent inefficiency In fast capture (typically 10-20%).