photon (EM calorimeter) resolution in delphes

[Tathagata Ghosh]

Dear delphes experts,

I am trying to simulate pp-> hh -> bb~+di-photon using Madgraph.The version of Madgraph I use has Delphes_v3.3.0 as detector simulator. The default delphes card I used (attached as 'delphes_card_default.dat'), tells me that the EM calorimeter resolution should be ~0.8 GeV for a 60 GeV photon. So, naively assuming the uncertainties in angular vaiables are negligible and both photons coming from the Higgs decay has similar energies, the uncertainty in the di-photon invariant mass should be ~\sqrt{2}*0.8=1.1 GeV. However, the resulting di-photon invariant mass distribution that I obtained has a FWHM of 7 GeV. I figured that somebody else asked the same question in
https://cp3.irmp.ucl.ac.be/projects/delphes/ticket/1081.
You guys guessed at that time that the distribution is broad at the generation level itself. So, I checked my generation level distribution (attached as 'Inv_Mass_LHE_MG5.pdf') and the distribution is very narrow in accordance with the theoretical width calculated.

Next, I checked by setting the efficiencies of all objects in delphes to 1 and switching off SmearTowerCenter and suppressing the calorimeter resolution by a factor of 0.00001. These modifications did make the di-photon invariant mass distribution a bit narrower but it still has a FWHM of 5 GeV. Please see the attached plot 'diphoton_Inv_Mass_delphes_v3p3p0.pdf' for the comparison of the two runs of delphes. My modified delphes card is attached as 'delphes_card.dat'.

I am at a loss now why the di-photon mass coming out of delphes is so broad even when the smearing is switched off. I shall be highly obliged if you guys kindly look into the matter.

Thanks in advance,
Tathagata