sPHENIX tracking discussion

[Anthony Frawley <afrawley AT fsu.edu>]

We had a long discussion at the tracking software meeting about how to pass the BCO timing information ( + offset) for each subsystem into the offline reconstruction. The initial proposal was to find a way to store this information in the trkrHitset object, which would require significant changes because of the large number of bits involved.

After some discussion, the issue is that our present scheme of storing the relative BCO number for the MVTX with respect to the level 1 trigger, for which we allocate 5 bits in the TrkrHitset key, will not work for fully streamed data from the MVTX, where there is no trigger reference. To time align fully streamed MVTX data, we need 52 bits to store the absolute BCO number for the MVTX signal. So, a change is needed to enable the use of fully streamed MVTX tracklet data for physics analysis.

Long story short, we agreed on the following:

1. We want to avoid having to substantially modify our current event-based Fun4All strategy. That would be a very big deal.
   
2. The global timing information for a triggered event is best stored in a trigger header word for the whole event, containing the BCO number for the event from the L1 trigger.  Then the TrkrHitset only needs to know the timing offset from the L1 trigger. The full BCO can be obtained by combining the absolute BCO from the trigger word and the relative BCO from the TrkrHitset.
   
3. For streamed MVTX data, the unpacker would make each strobe-fire into an "event". For that event, the trigger header word would contain the full BCO value for the strobe. This puts enough information in the data stream for aligning the MVTX hits in time for a tracklet analysis.
4. We thought that it would be best for the streamed MVTX data to go into separate files from the triggered data.
5. There will be a similar issue with the streamed digital current readout from the TPC, which could be handled in the same way as the streamed MVTX data.
   

As best we can tell, this would allow us to retain the current strategies while dealing with minimal disruption, which we would strongly prefer right now! The MVTX group will meet tomorrow at 1 pm, and will discuss a specific proposal for the information that will be needed.

We need to have a discussion with Chris and Martin about this, Joe has emailed them.

Best regards

Tony