Example configs for 8 GeV muon production #22
Conversation
Attached is a comparison of the distribution of the end z position for the hard brem at three different values for the muon conversion biasing factor. I chose to compare 1 order of magnitude above and below my nominal value of 3e4. The hard brem tail in the Ecal does not seem to vary much with an increase/decrease in the bias, so the chosen value of 3e4 seems like an ok choice.
|
exampleConfigs/8GeV_targetPN.py
Outdated
| from LDMX.SimCore import generators as gen | ||
|
|
||
| detector='ldmx-det-v14-8gev' | ||
| sim = target.photo_nuclear(detector, gen.single_8gev_e_upstream_tagger()) #thresh in target.photo_nuclear needs to be changed to 4850., currently configured for 4 GeV |
There was a problem hiding this comment.
@horohoo can you go ahead with this comment about updating the thresh to the 8 GeV setting please?
There was a problem hiding this comment.
And I'm not sure I understand why 4850 instead of 5000?
There was a problem hiding this comment.
4850 instead of 5000 came from a misunderstanding in how we thought the 8 GeV samples were biased. Since the trigger threshold at 8 GeV was <3150 MeV in the ECal instead of <3000 MeV, we wanted to be consistent with the 3150 value for the recoil electron and brem photon (8000 - 3150 = 4850) energy thresholds. I see that ECal PN uses 5000 as the bias threshold, so I will be consistent with how that biasing is configured when I update the target PN biasing.
There was a problem hiding this comment.
this brings up a topic that has been on my mind for... something like seven years 😄
we have historically been using the same (truth) biasing thresholds as the trigger threshold. the logic was that the trigger cut slashes everything below that cut anyway. before the thresholds were studied, they were assumed to scale with beam energy so 2500 became 5000. given the better conditions at 8 GeV, it turned out we could move this threshold little bit.
nowadays we have actual reconstruction applied in the systems used for the trigger decision and migration effects could be real. so i think it is high time to study what the relevant safety margin should be here to avoid strange edge effects from the choice of biasing thresholds.
I am uploading my example configuration files that I used to produce a sample of 8 GeV muon conversion events. I presented validation plots for these samples at two SWAN meetings:
3/11/24 SWAN meeting (target muon production)
2/26/24 SWAN meeting (ecal muon conversion)
Still on my to-do list, which I can add to this PR before approval, is to check if the biasing factor on the ecal muon conversion sample is correct. For this, I plan to generate small samples with biasing factors 10 times higher and lower than the value I originally used to see if the distribution is affected.