Conclusions from trigger meeting Oct 31, 1997
Some major conclusions and recommendations from a LHC-B trigger meeting, held on Oct 30/31, 1997 at CERN ====================================================
1.) The L0 pileup veto, using two rear Silicon planes, proposed by the NIKHEF group is desirable for most triggers at higher luminosity. Since this involves a redesign of the vertex tank, the tracking group is asked for the favour to do so.
Since the electronics is relatively cheap, and the vertex tank has to be prepared beforehand anyhow, we do not consider this as a staging option.
2.) The electron trigger simulation shows, that the (mu1) pad chamber can in fact be located at 25 cm in front of the calorimeter. The foreseen mu1 pad size (ECAL block in y, ECAL block/2 in x) is fine.
From the trigger point of view the arrangement RICH-2, honeycomb, mu1-padchamber, preshower, ECAL is possible and reasonable.
3.) The electron trigger simulation shows also, that a spatial resolution of the ECAL cluster in the trigger algorithm is essential. This has still to be implemented into the hardware.
4.) An alternative muon trigger design study has been started by Marseilles. A note on it is expected by end of November. An alternative calorimeter trigger implementation has been studied by Marco Bruschi. It is available as a LHC-B note.
5.) For the vertex trigger two implementation options will be documented for the TP, one with dedicated hardware for the track finding part and commercial CPUs for the rest of the algorithm, and one with a full commercial processor solution. Both options need a non trivial data distribution system.
For reasons of flexibility of the applications the full processor system is now being considered as the baseline solution.
6.) Tracking Trigger performance has been studied for the version of tracking L0 candidates upstream through the magnet till the first tracking station. The performance numbers given for the TP are however taken from the output after station number 6.
7.) L2 and L3 algorithm studies start to show some results. The present L2 philosophy combines tracking and vertex information. More ideas need to be studied to come to anticipated reduction of a factor of about 8.
Identifieng decay chains in L3 has started for first channels, the goal to reduce the rate 1:200 for a given channel seems to be in reach.
Although we can not expect to have a final system for the TP, the minimal goal is to present a CPU estimation for at least one option for both L2 and L3.
--------------------------------------------------------------- Prof. Ulrich Straumann Univ. Heidelberg, Physikalisches Institut, Philosophenweg 12, D-69120 Heidelberg, office 103 Tel 0049 6221 54 9325
DESY -FH1- office 1B, 351, Tel 0049 40 8998 3856
CERN: office 14-1-003 Tel 0041 22 767 3542