EGamma setup: x=1 in SF, TChannelWeight bad momenta

[Frank Siegert]

In the default EGamma example setup the integration leads to the following two kinds of errors:

SF::CalculateWeight : x out of bounds 1 at 17038.3, xrange = 1e-06 ... 0.999999

TChannelWeight: bad momenta!!!! -0.991994 - 0.473135 (0.566092)
1: (35.2969,0,0,35.2969)
2: (35.2969,0,0,-35.2969)
3: (7.91879,0.0321595,6.52771,4.48277)
4: (62.675,-0.0321595,-6.52771,-4.48277)

The SF one goes away if I disable the electron PDF by PDF_SET_2=None, that might explain why it hasn't appeared before (we didn't use to active electron PDFs by default).

[Frank Siegert]

It seems like the combination of ISR handler and beam handler in this setup leads to the problems. Looking at Phase_Space_Handler::Differential(process,mode) then the momenta still look fine before the p_isrhandler->BoostInLab happens:

  401:p_lab[0]=(89.1145,0,0,89.1145)
  402:p_lab[1]=(89.1145,0,0,-89.1145)

After that they also still seem ok:

  404:p_lab[0]=(38.8405,0,0,38.8405)
  405:p_lab[1]=(204.461,0,0,-204.461)

Then comes the p_beamhandler->BoostInLab which makes them:

  407:p_lab[0]=(25.9257,0,0,25.9257)
  408:p_lab[1]=(306.313,0,0,-306.313)

(which is clearly wrong given that both beam energies are at 250 GeV).

[Stefan Hoeche]

Fixed with r17109.

[Stefan Hoeche]

Milestone rel-1.3.0 deleted