Centroid Shift without any magnet alignment error

[ranamonikaa]

Hi all. I am simulating a transfer line. I expect zero centroid shift as there's no magnet errors in the input file. But, I see centroid shift when I give non-zero emittance with n_particles_per_bunch >1 in the bunched beam command, however if I set the emittance to zero or don't mention it at all or set n_particles_per_bunch =1 with non-zero emittance, then I see zero centroid shift. Can someone please tell where am I going wrong. I have attached ele file for reference.

[michael_borland]

Without a lattice file it isn't possible to say what the problem might be.

--Michael

[ranamonikaa]

I have attached the lattice file below.

[michael_borland]

The reason for this is that your beam line contains dipole magnets with small bending radii. There are sextupole-like terms present in such magnets due to the small radius of curvature of the coordinate system. This results in distortion of the phase space, which shows up as a small centroid error.

You can demonstrate this to yourself by setting ORDER=1 on the RBEND and SBEND elements.

To understand the theory, see equation (27) in https://cds.cern.ch/record/283218/files/SLAC-75.pdf
Setting n=0 and beta=0 gives your case (magnets with no explicit gradient or sextupole).

Simulating such magnets is challenging. I suggest switching from RBEND and SBEND to CSBEND and CCBEND, which have more accurate models.

--Michael

[ranamonikaa]

Thank you Michael.