TRestRawMultiFEMINOSToSignalProcess: first two events are with the same id

[nkx111]

I found that the first two events from TRestRawMultiFEMINOSToSignalProcess are with same id. Don't know if is the file/daq problem or the process's problem.

I download this file: https://github.com/rest-for-physics/rawlib/blob/master/pipeline/data/R01208_Ar2Iso_Background14h_14Vetos_IccubFEC-000.aqs

and type directly in restRoot:

TRestRawMultiFEMINOSToSignalProcess pc;  
pc.OpenInputFiles({"R01208_Ar2Iso_Background14h_14Vetos_IccubFEC-000.aqs"});
TRestRun r; 
pc.SetRunInfo(&r);
pc.InitProcess()

pc.ProcessEvent(NULL)->GetTimeStamp().GetNanoSec() //293870925
pc.ProcessEvent(NULL)->GetTimeStamp().GetNanoSec() //293870925
pc.ProcessEvent(NULL)->GetTimeStamp().GetNanoSec() //375673055
pc.ProcessEvent(NULL)->GetTimeStamp().GetNanoSec() //317906618

And see the timestamp of the first two events are same. We can also try event id, and see they are same. However, the signal data seems to be different, meaning that the events are not actually same.

TRestRawMultiFEMINOSToSignalProcess pc;  
pc.OpenInputFiles({"R01208_Ar2Iso_Background14h_14Vetos_IccubFEC-000.aqs"});
TRestRun r; 
pc.SetRunInfo(&r);
pc.InitProcess();

((TRestRawSignalEvent*)pc.ProcessEvent(NULL))->GetSignal(0)->GetData(0) //249
((TRestRawSignalEvent*)pc.ProcessEvent(NULL))->GetSignal(0)->GetData(0) //261
((TRestRawSignalEvent*)pc.ProcessEvent(NULL))->GetSignal(0)->GetData(0) //269
((TRestRawSignalEvent*)pc.ProcessEvent(NULL))->GetSignal(0)->GetData(0) //249

[juanangp]

Looking at the code, the bug looks more or less clear to me:

rawlib/src/TRestRawMultiFEMINOSToSignalProcess.cxx

Lines 496 to 507 in 0838016

	if (fSignalEvent->GetID() == 0) {
	if (fLastEventId == 0) {
	fSignalEvent->SetID(tmp);
	fSignalEvent->SetTime(tStart + (2147483648 * r2 + 32768 * r1 + r0) * 2e-8);
	} else {
	fSignalEvent->SetID(fLastEventId);
	fSignalEvent->SetTime(fLastTimeStamp);
	}
	}
	fLastEventId = tmp;
	fLastTimeStamp = tStart + (2147483648 * r2 + 32768 * r1 + r0) * 2e-8;

So I am not surprised that the timestamp and the ID are the same for the first 2 events.

[nkx111]

It seems the first event is stampped by the header frame after it, and the other events are stampped by the header frame ahead. So actually the first two events are using same header. What is the correct place for the header in FEMINOS daq?

[juanangp]

Following the FEMINOS manual the Event Count and the Timestamp are written once per event, I guess that if you have several FEMINOS you should get one Event Count and Timestamp per FEMINOS.

As far as I understand, event boundaries are between PFX_START_OF_EVENT and PFX_END_OF_EVENT, this should work for a single FEMINOS set-up. However, in a multi FEMINOS set-up I think you should check for PFX_START_OF_BUILT_EVENT and PFX_END_OF_BUILT_EVENT because you should have one event count and timestamp per FEMINOS that indeed should match for every event.

Anyhow, perhaps would be better if the person that wrote the code comment on that.

[jgalan]

There are few comments above line 496. In principle, the code mentioned by @juanangp should only be accessed in case the event id of the event is =0. I guess this should not be the case, or it is? It is the DAQ event id starting by zero?

If that lines of code are at the origin of the problem, perhaps a solution would be to initialise ID to -1, and then use the condition if( fSignalEvent->GetID() == -1)?