Description of triggers for Y2003 run

The currently best way to select minbias over the entire dataset taken with the dAuMinBias (and dAuCombined) setup name is, from StEvent (and MuDst's)

StEvent:

      if (event->triggerIdCollection() &&
          event->triggerIdCollection()->nominal() )
	   if (event->triggerIdCollection()->nominal()->isTrigger(2001) 
	       ||
               event->triggerIdCollection()->nominal()->isTrigger(2003)) {
	       // Keep event
           }
      } 
MuDst:
      if (mu->event()) {
	if (mu->event()->triggerIdCollection().nominal().isTrigger(2001)
	    ||
	    mu->event()->triggerIdCollection().nominal().isTrigger(2003)) {
              // Keep event 
        }
      }

This should select a uniform sample of events across the various changes so far made to the trigger. Of course, it will be somewhat sensitive to time-varying backgrounds because of the basic definition of the trigger (cut basically on a valid signal in the East (Au side) ZDC only).

Centrality Selection

1. ZDC West code fragment (peripheral events)

        bool dAuAnalysis::zdcTagging(int zdcw){
         if(zdcw>17.) return true;
         else return false;
   }
    
   if(zdcTagging((int) zdc.adc(0))) .... 
          

   zdc.adc(0) is a method from the StZdcTriggerDetectorCollection
2. FTPC
   • First, was the FTPC read out in the event?

              StL0Trigger l0 = mu->event()->l0Trigger();
              if((l0.triggerActionWord()&0x8)!=0){
                 // Keep event
               }
             

     Also, see this page to get a list of good runs with the FTPC.
   • See the comments and code in StRoot/StEventUtilities/StuRefMult.hh . Note that, up until revision 1.5 (first in SL04e), this would only return a valid number for 2001||2003 trigger id's, hence making this code difficult to use for, say, high tower events. Since the MuDst's write out a reference multiplicity into the MuEvent for the FTPC's using this code, the same problem applies in the MuDst's.

This page gives a truly detailed list of triggers and definitions throughout the run. Please refer to this as the definitive resource.

The filling of the information was put in place on Feb. 20, 2003. Fast offline and the P03ia production have these in both the MuDst's and StEvent.

The trigger this year has gone to a different scheme of identifying the reasons for an event being taken by the trigger. This is in an attempt to resolve the overlap issues with scaledowns in the L0 trigger, which in hardware can only deal with unique, non-overlapping conditions.

The essential change is that the interface for a user is now based on "offline trigger id's", rather than the level 0 trigger word. This trigger id:

• Is unique across runs and trigger setup names
• Is built up on the fly from a combination of underlying L0 trigger words in the L1, L2, L3, and Event Builder hardware.
• Accounts for scaledowns and higher level trigger decisions, which are in principle different for the different underlying L0 trigger words, such that a selection on the trigger id should produce an unbiased distribution for the conditions under which it's defined.

This page (STAR->Group Documents->Daq->Trigger Accounting Database Details) gives all the details about the setup, and this page (STAR->Group Documents->Daq->Primordial RC/Trigger interface information page->Offline Bit Map) gives the current set of trigger id's defined. This page gives a truly detailed list of triggers and definitions throughout the run. Please refer to this as the definitive resource.

Currently we have 6 trigger id's defined in the dAuMinbias and dAuCombined trigger setup

Trigger Id	Name	Definition
2001	dAuMinbias-1	ZDC_E > 5 && 20 < ZDC_E tac < 225 && B&Y
2002	dAuMinbias-2	ZDC_E > 5 && ZDC_W > 5
2003	dAuMinbias-3	ZDC_E > 5 && 20 < ZDC_E tac < 225
2004	dAuMinbias-4	BBC Coincidence
2201	dAuHighPt-1	EMC High Tower 2.5 GeV threshold
2202	dAuHighPt-2	EMC High Tower 4.5 GeV threshold

2003 is our nominal minbias trigger. It is, however, susceptible to beam-gas backgrounds (since it only has requirements on one side of STAR), and so 2002, a ZDC coincidence, which should have low efficiency but high purity, was put in place to monitor the level of backgrounds. Unfortunately, as detailed below, this is of somewhat limited usefulness due to various problems.

The difference between dAuMinbias and dAuCombined trigger setup names is that the dAuCombined includes the high tower triggerId's, and uses prescales. TriggerId's (2001 or 2003) should still select the same minimumbias sample in this trigger setup.

Note that there have been some problems with the blue and yellow bits, and that the ZDC coincidence trigger (2002) does not require them. This led to a bias towards this trigger in some data, listed below. This was the reason for creating a new trigger id 2003 that did not require the blue and yellow bits. Assuming that the blue and yellow bit problems were uncorrelated with physics, the bias of this data should be recoverable by asking only for trigger id 2001, though this does lead to a smaller fraction of usable events in these files.

Runs	Active Trigger Id	Notes
< 4017040	2001	Every event satisfies 2001
4017040 <= run < 4024027	2001, 2002	Data on tape is biased towards 2002, or more peripheral events. Cut on trigger id needs to be made.
4024027 <= run < 4041023	2001,2002,2003	2001 was kept in as a cross-check of BY bit timing. Every event of dAuMinBias setup satisfies 2003.
4041024 <= run < present	2001,2002,2003,2004,2201,2202	Large set of mixed triggers. Cut on trigger id needs to be made to have any hope of analyzing the data.

There was also a problem in an extra killer bit requirement for the ZDCW that caused 2002 NOT to fire in a few sets of runs, leading to a fraction of 2002 to 2003 that was lower than it should have been. Hence for this set of runs the ratio of ZDC coincidences to ZDCE in the trigger id is not very meaningful.

To be complete, and to show how complicated the level0 trigger word definitions actually become under this scheme, here is a list of what have been the definitions up until we started to mix with prescales. One can still hack together the trigger id's from the L0 trigger words since there are no prescales or higher-order algorithms, but this is not sufficient information to reproduce the trigger id's past this date. Note that the correspondences are many-to-many.

Trigger word	Trigger Id	Name	Condition
Run < 4017040
0x1	2001	dAuMinBias-1	ZDCE && BY
4017040 <= run < 4024027
0x2	2001	dAuMinBias-1	(ZDCE && BY) && !ZDCW
0x3	2001	dAuMinBias-1	(ZDCE && BY) && ZDCW
0x1	2002	dAuMinBias-2	(ZDCE && !BY) && ZDCW
0x3	2002	dAuMinBias-2	(ZDCE && BY) && ZDCW
4024027 < = run < 4041023
0x3	2001	dAuMinBias-1	(ZDCE && BY) && !ZDCW
0x4	2001	dAuMinBias-1	(ZDCE && BY) && ZDCW
0x2	2002	dAuMinBias-2	(ZDCE && !BY) && ZDCW
0x4	2002	dAuMinBias-2	(ZDCE && BY) && ZDCW
0x1	2003	dAuMinBias-3	(ZDCE && !BY) && !ZDCW
0x2	2003	dAuMinBias-3	(ZDCE && !BY) && ZDCW
0x3	2003	dAuMinBias-3	(ZDCE && BY) && !ZDCW
0x4	2003	dAuMinBias-3	(ZDCE && BY) && ZDCW