Trigger Background Experience & Predictions Su Dong Sep/23/2003 Su Dong BaBar Background Workshop
Trigger Background Perspective The most significant issue is simply the impact of L1 rate growth on the DAQ dead time. The only systematic data for extrapolation is still the Feb/02 background runs. The Drift Chamber Trigger upgrade to reject tracking trigger background using Z information will be a major improvement. Its design robustness at high background conditions was evaluated. L3 now has considerable head room after moving to Linux and have straightforward path for upgrade with faster CPU and more nodes still. Sep/23/2003 Su Dong BaBar Background Workshop
BaBar Background Workshop L1 Rate Projection Feb/02 background run result suggest Possible model at good conditions: L1 Rate (Hz) = 125 (cosmic) + 50*ILER(A) + 190*IHER(A) + 180*I2HER(A) + 70*L/1033 + 100*ILER*IHER How well does this work ? How should we use such a model ? Sep/23/2003 Su Dong BaBar Background Workshop
BaBar Background Workshop The need for head room Assumed ground rules: Need to find strategy to keep dead time at <~3% even at bad conditions. Cannot operate on the DAQ rate limit. Always need head room. Sep/23/2003 Su Dong BaBar Background Workshop
Checking L1 Rate Projection All colliding beam runs longer than 10min and dead time <40%. Examine ratio of average L1 rate vs prediction. Now we know the background runs were really at best conditions ! There are significant tails in the L1 rate. Sep/23/2003 Su Dong BaBar Background Workshop
Checking L1 Rate Projection Derived background L1 rate compared to prediction: >2*backgr 24% >3*backgr 4.6% >4*backgr 2.2% Use prediction with caution ! Sep/23/2003 Su Dong BaBar Background Workshop
BaBar Background Workshop L1 Trigger Rate Growth vs Trigger Types Sep/23/2003 Su Dong BaBar Background Workshop
Improvements from DCT Upgrade (FDR Apr/03) Very simple usage: DCT triggers 1Z track (|z0|<10cm) Can reject ~53% of DCT background (A crude test with more elaborate usage: 1Z Pt>0.6GeV OR 2Z Pt>0.25GeV can reject ~67% of DCT background ) Sep/23/2003 Su Dong BaBar Background Workshop
DCT Upgrade Effect on L1 Rate CDR study Sep/01. Simple 1Z cut, assuming EMT unchanged. Can assume New DCZ allow Reduction of all Background triggers by 40%. Sep/23/2003 Su Dong BaBar Background Workshop
DCT Upgrade Robustness at High Background Open background D2 track triggers with multiple background overlays. Background rejection Capabilities largely Preserved up to ~5-6 times of Sep/01 Background. Physics efficiency at higher background behaves even better. Sep/23/2003 Su Dong BaBar Background Workshop
Real boards running at IR2 for 1/8 of f ! Status of DCT Upgrade Real boards running at IR2 for 1/8 of f ! Sep/23/2003 Su Dong BaBar Background Workshop
BaBar Background Workshop L1 Rate Projection Feb/02 background studies concluded: L1 (Hz) = 125 (cosmic) + 50*ILER + 180*IHER + 190*I2HER + 100* ILER* IHER + 70*L/1033 Prediction based on Seeman model (Feb/03): Lumi LER HER No-DCZ DCZ `2*bck’ `3*bck’ 4.8 1650 950 1043 760 1110 1460 7.5 2400 1250 1589 1163 1727 2290 15 3300 1600 2633 2000 2874 3750 25 4500 2000 4100 3160 4495 5830 Given the discrepancy even extrapolating into 2003, this projection clearly has to be read with caution ! Sep/23/2003 Su Dong BaBar Background Workshop
Trigger Background Perspective L1 Rate is worse than Feb/02 projection. We clearly need to be cautious in extrapolating rates based on a particular machine configuration and condition. We need to continue to preserve head room. L1 DCT upgrade is on its way. We still have some work to do to make the best use of this upgrade through optimized trigger configuration. There is on more possibility of L1 upgrade to veto Bhabhas in GLT, which only needs GLT firmware updates. However, benefit is only significant for luminosity well beyond 1034. Sep/23/2003 Su Dong BaBar Background Workshop