basically, there are two stages.
Recon
1) ADC->DAC
2) DAC->MEV
digi is the opposite of course, but is done in peices. first the dac 'scale' values for the individual McIntegratingHit energies are summed together, once they're all summed, then we convert dac->adc.
here are the constants from CalXtalResponse/xml/idealCalib.xml file.
<item name = "LARGE" value = "0.36"> Large diode MevPerDac </item> <item name = "SMALL" value = "2.0"> Small diode MevPerDac </item> <item name = "ADC_PER_DAC" value = "11.2, 1.3, 11.5, 1.3"> ADC/DAC
for each ADC range </item>
so for LEX8 recon would go something like this:
dac = adc/11.2 // 11.2 is the LEX8 val for adc_per_dac mev = dac*0.36 // using large diode MevPerDac.
so final ratios are:
LEX8 ADC/Mev = 0.0321 LEX1 = 0.27 HEX8 = 0.17 HEX1 = 1.5
all ranges are currently maxing out at 3500 ADC (including pedestals) pedestals are about 500 in X8 and 200 in X1 this makes the max measurable energy in HEX1 (3500-200)/1.5=2.2gev per xtal. i don't know if that's a problem or not, like i said before these numbers are averages from muonCalib which uses different gain settings from flight.
incidentally, all these values can be changed instantly by editing the idealCalib.xml file (so you can see the effect of changing gains, peds, etc...).