Fork me on GitHub

Opened 10 years ago

Last modified 9 years ago

#350 new How to

FCC detector card

Reported by: Sergei Chekanov Owned by:
Priority: minor Milestone:
Component: Delphes code Version: Delphes 3
Keywords: Cc:

Description

Hello,


I have a question about FCC detector card. It looks like, for the standard configuration of ECAL and HCAL, the energy fraction of pions and baryons is 100% for HCAL and 0% for ECAL. What is correct way to set 60% energy fractions for ECAL and 40% for HCAL. Can I set:

add EnergyFraction {0} {0.6}

for ECAL and

add EnergyFraction {0} {0.4}

for HCAL? This will be applied for all particles, and then I set 100% for electrons and photons for ECAL?

Change History (7)

comment:1 by Michele Selvaggi, 10 years ago

Hi,

yes the fraction of charged pions and baryons (actually everything that is not explicitly set below) is set so that 100% of the energy is deposited in HCAL, but you can change that according to your preferences.
Be aware that some generators do not decay pi0 so we also force pi0 to deposit into ECAL as a safety measure.

Michele

comment:2 by Sergei Chekanov, 10 years ago

Hi, Michele

Thanks, this is very useful. One reason why I've asked this question is that I see a very small effect on jet mass when I change HCAL segmentation, even when the ratio is 1 for pions/hadrons for HCAL. I'm using "delphes_card_FCC_basic.tcl". Is it possible that this card uses "eflow" for jets, so it does not really matter what I do for HCAL - charged particles are taken from "tracks". If this is true, is it possible to use only-CAL cell to build jets? Do you have such an example?

best wishes, Sergei

comment:3 by Michele Selvaggi, 10 years ago

Hi Sergei,

yes it is perfectly possible to use Calo-based jets. You can simply change the input collection to the FastJetFinder module:

############
# Jet finder
############
	
module FastJetFinder FastJetFinder {

set InputArray TowerMerger/towers
…
}

Please be aware, that for the FCC example there are two Calorimeters, producing two separate tower collections (that get merged before entering jet reconstruction), as opposed to CMS and ATLAS where ECAL and HCAL have the same segmentation and produce only one single collection.

You might also want to play with the ATLAS card, which produces calojets, and compare against FCC.

Cheers,
Michele


comment:4 by Sergei Chekanov, 10 years ago

Hello, Michele

Thanks a lot. Yes, I see how to do this. So far, all plots make sense, but there is one which looks strange. When I use the "tower" inputs for antiKT jets R=0.5 option and reconstruct jet masses for top-jet, the tower-jets option makes rather ugly jet mass. See https://atlaswww.hep.anl.gov/asc/wikidoc/lib/exe/fetch.php?media=vlhc:mass_towers.png. When I change the input to eflow, the jet mass ~2 narrower and it peaks where it should be (170 GeV). I do not think ATLAS has such bad jet mass (of course, we use topoclusters but, even for tower jets, this should not be so bad).

I also have applied 18% jet energy scale correction for tower-jet, by looking at the resolution plots. This shifts jet mass peak even further to the right, while fixing the jet pT.

Does it look strange? Can be an indication of some problem for the tower jets?

best, Sergei

comment:5 by Michele Selvaggi, 10 years ago

The top mass being shifted towards higher values using calo towers, I suspect this is due to the jet radius being too large (with respect to the large boost) and catching too much random radiation in the cone. This should go away with some pruning/trimming procedure. One way to be sure about that is to check if the shift is still present at gen-level plots. Actually I believe you also see the bias on e-flow jets, it simply appears to be more mild due to tracking efficiency loss.

Michele

comment:6 by laurapcard, 9 years ago

blockedby: 456

comment:7 by trac, 9 years ago

blockedby: 456

(In #456) Ticket #456 deleted

Note: See TracTickets for help on using tickets.