FIX: Reading calibrated timing in MModuleLoaderMeasurementsROA

[fhagemann]

The ROA loader is currently assigning the values for Timing to m_TAC instead of m_Timing.
This is fixed now.

[fhagemann]

Also: what is m_Timing supposed to be? The calibrated TAC value in nanoseconds? Or the drift time since the creation of electron hole pairs (in nanoseconds)?

According to the comment in MStripHit, it should be a double denoting the //! Timing in ns

nuclearizer/include/MStripHit.h

Lines 189 to 190 in 07e64cd

	//! Timing in ns
	double m_Timing;

In the ReadOutAssemblyTiming in megalib, m_Timing is an unsigned int denoted as //! The timing:
https://github.com/zoglauer/megalib/blob/c5fa4af354fc36be4cd71511402a68f0d6d6952f/src/fretalon/base/inc/MReadOutDataTiming.h#L90-L91

Currently, this would write floating point numbers to the file, but expects integers when reading them, resulting in very funky behavior.

[ckierans]

Good catch!

m_Timing should be the TAC calibrated into ns. m_TAC was a variable Sean introduced awhile back. We used to only have the Timing keyword and it was overloaded to represent the ADC value and the ns conversion. m_Timing should be a floating point whereas m_TAC should be an int.

[fhagemann]

But changing the type of m_Timing this something that we need to tackle in megalib (on the develop-cosi branch)
then? If so, I can open an issue / draft a PR for that.

[zoglauer]

The MEGalib part is what we read from the roa file, the TAC value.
Thus is should be an integer

[fhagemann]

The MEGalib part is what we read from the roa file, the TAC value. Thus is should be an integer

But let's assume we write calibrated timing values to ROA, what would be the preferred reader then?
I would then need to use ReadOutDataTiming for the m_TAC and something else (that would then still need to be written?) for m_Timing.
I feel the naming is a bit inconsistent here..

[zoglauer]

Yes, Timing is from the old days. We didn't call it TAC in the balloon days. But why do we write ns timing to roa files? roa files are for the data from the raw measurements.

[fhagemann]

Well, the current ROA loader has to possibility to read "TAC" and "Timing" and attributes it to the m_TAC and m_Timing value of the strip hits (?)

No my question would be if this way of reading the data is actually correct or not?

nuclearizer/src/MModuleLoaderMeasurementsROA.cxx

Lines 188 to 233 in fc79014

	for (unsigned int r = 0; r < Event->GetNumberOfReadOuts(); ++r) {
	MReadOut RO = Event->GetReadOut(r);
	const MReadOutElementDoubleStrip* Strip =
	dynamic_cast<const MReadOutElementDoubleStrip*>(&(RO.GetReadOutElement()));
	const MReadOutDataADCValue* ADC =
	dynamic_cast<const MReadOutDataADCValue*>(RO.GetReadOutData().Get(MReadOutDataADCValue::m_TypeID));
	const MReadOutDataTiming* Timing =
	dynamic_cast<const MReadOutDataTiming*>(RO.GetReadOutData().Get(MReadOutDataTiming::m_TypeID));
	const MReadOutDataTAC* TAC =
	dynamic_cast<const MReadOutDataTAC*>(RO.GetReadOutData().Get(MReadOutDataTAC::m_TypeID));
	const MReadOutDataEnergy* Energy =
	dynamic_cast<const MReadOutDataEnergy*>(RO.GetReadOutData().Get(MReadOutDataEnergy::m_TypeID));
	const MReadOutDataOrigins* Origins =
	dynamic_cast<const MReadOutDataOrigins*>(RO.GetReadOutData().Get(MReadOutDataOrigins::m_TypeID));
	MStripHit* SH = new MStripHit();
	SH->SetDetectorID(Strip->GetDetectorID());
	SH->IsXStrip(Strip->IsLowVoltageStrip());
	SH->SetStripID(Strip->GetStripID());
	if (Timing != nullptr) {
	SH->SetTiming(Timing->GetTiming());
	}
	if (TAC != nullptr) {
	SH->SetTAC(TAC->GetTAC());
	}
	if (Energy != nullptr) {
	SH->SetEnergy(Energy->GetEnergy());
	}
	if (ADC != nullptr) {
	SH->SetADCUnits(ADC->GetADCValue());
	}
	if (Origins != nullptr) {
	SH->AddOrigins(Origins->GetOrigins());
	}
	Event->AddStripHit(SH);
	}
	Event->SetTimeUTC(Event->GetTime());
	return true;
	}

[fhagemann]

And there is also possibility to read/write ADC and Energy, the equivalent for the slow shaper output.

[fhagemann]

This is a screenshot of what I have:
adc = (Uncalibrated) energies in ADC units
tac = (Uncalibrated) TAC values in ADC units
energy = (Calibrated) energies in keV
timing = (Calibrated) TAC values in ns

Note that it's the same quantities, just differing in calibration/units.
Timing right now would NOT be drift time or something.
It's still the TAC value, just in nanoseconds.