bunch shape and CSRCSBEND

[libov]

Dear Michael,

I'm working on splitting a bunch longitudinally into two sub-bunches, by using a mask in a dispersive section.

This means that the bunch entering the last dipole of the dispersive section is actually two transversely-separated bunches; after the dipole dispersion is closed, the bunch is rotated and now these are two longitudinally-separated bunches.

Question: do you expect problems with CSRCSBEND element if there is such a discontinuity? And/or there are large offset particles (halo)?
What I see is some strange behaviour: often all particles are lost close to the beginning of the dipole.. And if I just use SBEND, everything is fine.

I will try now to track these two bunches through CSRCSBEND independently.

Thanks in advance!

Cheers,
Slava

[michael_borland]

Slava,

I can imagine some problems may occur if there are longitudinally isolated particles. These would drive noise in the CSR calculations. If this is what is happening, probably the best approach is to increase the number of particles used in order to better sample the distribution in the under-populated areas. Also, you could try using a CLEAN element to clean up the halo particles, since you probably don't care about them.

--Michael