Collimators are not correct

[曾宪虎]

Dear BDSim expert,

I found the collimators (ecol and rcol) are not correct. If the collimator is very short, then the beam size is correct. But if the collimator is long or the collimation system consists of two collimator which has a long distance between two collimator(such as 1 m, then you will find that the beam size is smaller and the beam transmission is also decreased. I have compared a collimator based on ecol setting with a collimator based on ggmad g4 cons. I found the results based ggmad g4 cons are the same as other MC code. The results of ecol is smaller. I think this should be improved.

Looking forward to your early reply.

Thanks

Xianhu

‌

[Laurie Nevay]

Dear Xianhu,

Please note a G4Cons is not the same as an ecol nor rcol. ecol and rcol both have a square exterior, but G4Cons is a hollow cylinder. Please also note that the “xsize” and “ysize” parameters of the collimators are half-widths of the apertures and not full widths.

Both the ecol and rcol are used extensively and produce the desired geometry (correct size).

The effects you describe with beam size depend on not only the sizes of the ecol or rcol, but also the beam distribution, materials chosen and physics used. The ecol and rcol simply produce blocks of material. The only thing that can be wrong with them is their size.

If you change the length of the collimators you will surely change the beam size afterwards as the beam will have travelled a different distance, irrespective of the effect of the collimator. So I wouldn’t expect them to be the same size.

I don’t see what we can change in the code for this. Could you attach a simple example that shows what you see as wrong please.

Best,

Laurie

[Laurie Nevay]

One issue was found where the (invisible) container volume for the the cons solid in GGMAD geometry was twice as big as it should be, which caused overlaps in the geometry and therefore bad tracking. This has since been fixed with v1.4.

[Laurie Nevay]

I think we've fixed this issue now offline. The automatic container geometry for ggmad geometry has been corrected.