Set ErrorNode for Dipole, result is different from MADX

[LIUXY96]

I use MADX and PyORBIT to track single particle through a drift, a dipole and another dirft, then set only fielderror to this dipole, the x coordinate is different from madx result, which based on same lattice and same error. Also, when I set only xsroterror to this dipole, and initial dE is zero, the final dE coordinate is nozero, but very small. I think this should be absolutely zero. Is there any bug for ErrorNode? Or I had something wrong?

[shishlo]

Dear LXY960427,
Can you send to me the MADX file and the PyORBIT script that you mentioned. I will try to take a look.

[LIUXY96]

Thanks for your help! I find that madx and pyorbit give differrent results, although dx, ds, dpsi, dtheta, dphi and field error are same. For example, I set that dipole dx error is 0.127mm, initial coordinate: x=0.001, y=0.001. Final coordinate from madx: (0.0004504628565, -0.0004948271389, 0.001, 0, -0.0008661142371, 0), final coordinate from pyorbit: (0.00050427978, -0.00050443287, 0.001, 0, -0.00077072941, 0). Y coordinate is same but x and z direction are quit different. Ds displacement error has same problem. I set that dipole xsrot error is -0.126mrad, initial coordinate: x=0.001, y=0.001. Final coordinate from madx: (0.0004370733215, -0.0006298123105, 0.001, 0, -0.0008661599623, 0), final coordinate from pyorbit: (0.00044315998, -0.00056678022, 0.001, 0, -0.00086599932, -9.9609841e-10). px coordinate is quit different, and dE changes from 0 to -9.9609841e-10, which I think can not happen in real accelerator. xyroterror and ysroterror have same problem with xsroterror. By the way, quadrupole xsroterror and ysroterror alse have this problem. Also, I get different results form same dipole field error (I set fracerr parameter in pyorbit). Codes ares in this two folders, you can check if I set something wrong. Looking forward to your reply!

[LIUXY96]

Thanks a lot! MADX and PyORBIT scripts are here: https://github.com/LXY960427/Balabala

[shishlo]

Thanks for explanations! I got the files and will take a look. I am not sure how long it will take. Best regards,Andrei

[chiragpatidar]

Dear all,
I simulated different types of errors for the given lattices in madx and pyorbit using the scripts. The difference in pyorbit with madx is mainly due to difference in rotation pivot and error point . Madx pivot point for rotation and error point for displacement are at the beginnign of element while for py-orbit they are at middle of element. With change in pivot and error point, pyorbit matches fairly well with madx except followings output varaibles.

1. z for dipole field error (dB/B), horozontal placement (dx) and longitudinal placement (ds) errors in dipole.
2. y and (x, z) for x-y and x-s rotatation errors in dipole, respectively. Except z, others can be corrected /minimised by changing a length varaible (replacing tan(phi/2) by sin(phi/2)) in the error model.
3. All difference with respect to madx are of the order of 1-2 mm for practical values of errors and they are independent of input co-ordinates so can be ignored for long and big bunches. I am attching a doc file of my results.

Thank you
comparative study of magnet errors -madx pyorbit short copy.docx