How does MNCP calculate an F6 Tally?

[MadGander]

TL;DR Summary

How does MNCP calculate an F6 Tally?

I'm wondering how exactly MCNP calculates an f6 tally? I'm trying to compare a theoretical result with an MCNP f6 tally (MeV/g). I have an initial energy spectrum and a thin layer of lead that attenuates the x-rays. Using the attenuation coefficient at each energy (bin width of 0.5 kev from 0-100 kev) and intensity at each energy, I produced an attenuated energy spectrum.

How do I compare this attenuated spectrum to an f6 tally? My thought process is to find the average energy within the spectrum and then divide by the mass of the f6 tally cell. However, this is a couple orders of magnitude off the MCNP ouput so I'm guessing it's wrong.

Another thing to note is that my average energy in the attenuated spectrum is higher than the average energy in an initial spectrum. The attenuator is a 0.08 cm thick lead sample, so there's no way the final average energy should be higher than the initial, right?

 

[Alex A]

Your simulation might need to account for scattered X-rays and electrons that don't make it to the detector. Orders of magnitude error makes me suspect a math mistake though.

Total Initial energy = Total Final energy + Total F6 + Total scattered (that does not hit the detector). Multiplying each energy bin by the midpoint of it's energy and summing the total should produce simple values that add up correctly. Individual bins don't have to obey conservation of energy but the total should.

I assume the F6 tally sums the energy loss in the cell and that value already has units of energy per cell volume (since twice the cell would tally twice the amount), that multiplied by the density would result in an answer that is in MeV/g.

All tallies are per source particle of course. Either multiply up the F6/scattered values or divide down the other two. If you have a largely monoenergetic source and you want it to look like a book value dividing down might make the most sense. Do each step one at a time and keep track of your units.

I'm wondering if an F8 tally would be better and I'm finding it hard to think clearly about this. If you are able to share your results and/or your MCNP input/output and/or your working out, that might give you the best chance of someone spotting the problem.