Simulation Group

  Keep Out Zones a la Allena
  Katherine's Status Report
  Michael's Web Page
  Klaus's Status Report
  Burnham's Status Report
  Previous Reports

   GEANT3 Geometry Summary   OLD!

The following pages require a username and password:

  Qweak elog at JLab
  Qweak Simulation Forum
  Klaus' elog

*Note: From now on use the simulation mailing list - if you aren't on it and you want to be make sure you sign up!

You can now get the Qweak simulation from the subversion repository at JLab!

new gukine.F

New Inelastic Generator

Figure 1 - Rates for a lower x of 318 and 319 cm, with (cases 3,4) and without secondaries (cases 1,2), old generator.	Figure 2 - Rates for a lower x of 318 and 319 cm, with (cases 3,4) and without secondaries (cases 1,2), new generator.

Case Lower x (cm) Secondaries? 1 318 no 2 319 no 3 318 yes 4 319 yes

Katherine and I first realized there was something wrong with the inelastic generator when we tried to compare the inelastic percentages from simulations run on her machine compared to mine. Figures 1 and 2 show the comparisons for the cases in the table to the left, 1 for the old generator, and 2 for the new generator (still has the wrong EPMAX def'n). We checked our error calculations, made sure we were using the same euclid file, etc. We finally tracked it down to the fact that we were using different theta ranges. In the mean time, we had discovered that a factor of sin(&theta) had been added to the rate calculation that we all believed shouldn't be there. Unfortunately removing that factor didn't fix the problem.

To illustrate the issue, figure 4 shows the inelastic rate vs.decreasing generated upper theta. For comparison, a similar plot for the elastics is shown in Figure 3. Basically, the weighting in the old inelastic generator was being done incorrectly, so that the rate as a dependence on the upper theta that you generate in. As long as the generation phase volume is larger than the acceptance phase volume, the size of the volume shouldn't matter. The main effect that if it is too large, you waste time computing events that don't make it into your acceptance. Figures 5 and 6 show the results with the new generator, where we tried to generate "flat" in theta, phi and energy. Unfortunately there was an error in the energy generation and so Figure 5 is an intermediate, also incorrect generator. This will be explained more in figures 7-9.

Figure 3 - Elastic rates for decreasing generated upper theta.

Figure 4 - Inelastic rates for decreasing upper theta, old generator.	Figure 5 - Inelastic rates for decreasing upper theta, new generator, bad epmax.	Figure 6 - Inelastic rates for decreasing upper theta, new generator, new (good?) epmax.

Figure 7 - Plots of unweighted theta_o, phi_o, E'_o and vertx_in on the bar for old generator (fewer events).	Figure 8 - Plots of unweighted theta_o, phi_o, E'_o and vertx_in on the bar for new generator, bad epmax.	Figure 9 - Plots of unweighted theta_o, phi_o, E'_o and vertx_in on the bar for new generator, new (good?) epmax.

Figures 7-8 show the unweighted inelastic theta, phi and E' distributions at the scattering vertex, as well as the unweighted x distribution on the bar (the theta distributions are scaled by the generation range in all 3 cases). You can see that with the old generator, the theta and E' distributions are not flat; we just were not using the proper weight factors to calculate the rate. Figure 8 shows evidence of the incorrect EPMAX - it was incorrectly coded as a function of theta. So though the distribution of theta is flat and the energy is "flatter" it is still incorrect. Figure 8 shows the correct (EPMAX=EBEAM to be safe) and the unweighted theta, phi and E' distributions are all flat. In this case a lot of the events hitting the bar have a weight = 0, so that is why that plot looks so different in this case.

Figure 10 - Plots of weighted theta_o, phi_o, E'_o and vertx_in on the bar for old generator (fewer events).	Figure 11 - Plots of weighted theta_o, phi_o, E'_o and vertx_in on the bar for new generator, bad epmax.	Figure 12 - Plots of weighted theta_o, phi_o, E'_o and vertx_in on the bar for new generator, new (good?) epmax.

Figures 10-12 show the weighted distributions. In the old generator (figure 10), even when we thought we were properly weighting, changing the generate theta range changes the scale of the distributions. In the new generator (figure 12) changing the generated theta range doesn't affect the distributions.