Inconsistent Results in Orbit Correction Testing

Moderators: cyao, michael_borland

Post Reply
NuanYa_Huang
Posts: 2
Joined: 16 Mar 2023, 07:51

Inconsistent Results in Orbit Correction Testing

Post by NuanYa_Huang » 07 Nov 2023, 20:13

Dear all,

I am currently using the correct command to test the closed orbit correction configuration in ELEGANT 2022-2-0. However, I have encountered an inconsistency between the *.scor, *.orb, and *.bpm results. This inconsistency appears to be limited to the horizontal information, as other statistical results for the vertical orbit and correctors are consistent.

For instance, as shown in the attached slide, the *scor file indicates that the final corrected Pmax is 0.7 µm, but the raw data for the x orbit shows Xmax to be approximately 400 µm. I have also attached my ele and lte files for your reference.

If you have any insights into what might be causing this issue and how I can correct it, I would greatly appreciate your valuable comments.

Thank you.
Attachments
TPSII-HOA-BVB-2S4SH-Error_V5-RF-CX1CY1.new
(4.69 KiB) Downloaded 139 times
OrbitCorrection.ele
(2.69 KiB) Downloaded 137 times
ELEGANT2022-2-0_Question-1_toForum_20231108.pdf
(1.05 MiB) Downloaded 133 times

michael_borland
Posts: 1959
Joined: 19 May 2008, 09:33
Location: Argonne National Laboratory
Contact:

Re: Inconsistent Results in Orbit Correction Testing

Post by michael_borland » 19 Dec 2023, 14:42

This seems to be a bug that appears when n_xy_cycles=1. Set n_xy_cycles to 2 or higher in &correct and you should see consistency.

Thanks for reporting the problem.

--Michael

michael_borland
Posts: 1959
Joined: 19 May 2008, 09:33
Location: Argonne National Laboratory
Contact:

Re: Inconsistent Results in Orbit Correction Testing

Post by michael_borland » 19 Dec 2023, 18:18

After a little more looking, I've concluded that this isn't a bug, but rather a result of the presence of linear and nonlinear coupling, through tilted quadrupoles and sextupoles, respectively.

Elegant alternates between correcting in the x and y planes. With n_xy_cycles=1, it corrects the x plane, then the y plane. However, with coupling, the y-plane correction changes the horizontal orbit, which is reflected in the trajectory output but not in the statistics, which are computed for a the plane that was just corrected.

Generally, I tend to use n_xy_cycles fairly large (e.g., 10), with n_iterations fairly small (e.g., 5). This gives good convergence in both planes.

--Michael

Post Reply