Chromaticity

[twatanabe]

Hello Michael,

When an example of the dynamic aperture, PAR/dynamicAperture/track, is executed, tunes for each energy deviation are shown in addition to the dynamic aperture. I assume the chromaticities (linear and higher order) can be estimated by the tunes as a function of the energy, but it seems to me it is not consistent with the chromaticities estimated from the twiss parameters for the same lattice. For example, when I see the result of the dynamic aperture, PAR/dynamicAperture/track, it seems to me the linear chromaticity is something like +2.2, but as the other example (PAR/chromTracking/track) shows, the linear chromaticity based on the twiss parameter may be -0.63; even the sing seems to be not consistent.
I believe I am missing something, but I haven't found it. Could you please show me how I can properly estimate chromaticities from the dynamic aperture example?

Thank you in advance.

Takahiro

[michael_borland]

Takahiro,

There are several differences between those two runs, for example, the DA run includes chromaticity correction and large K2 and K3 values on the dipoles. That's why the chromaticities are different.

To show how to compare twiss calculations vs delta to tracking vs delta, I modified the chromTracking example and attach it here.

This shows good agreement, but I had to set K3=0 on the dipoles to get this. The reason is that elegant doesn't include K3 in the dipole matrix (which is only 2nd order), so the results would not agree with tracking.

--Michael

[twatanabe]

Hi Michael,

Thanks for the reply and the attached file. Yes, as you pointed out, I missed the fact that the example file includes the chromaticity correction etc. Your file worked well and gave the consistent result between tunes as a function of energy deviation, and the linear chromaticity calculated by twiss parameters.

Actually, however, I had found the above problem in my own calculation in which chromaticity correction, nor K2 and K3 values were not included. (I aseked the question based on the example files in order to make my question clearer, but maybe I failed.)

As far as I understand, all what I did was to modify the DA run example for my calculation, just by removing the chromaticity correction, replacing the lattice file to ours (no K2 and K3) and rewriting other parameters such as e-beam energy.
I noticed that the result (tune) was sensitive to "closed_orbit_iterations" and "iteration_fraction" when finding the closed orbit, so I optimized the parameters. I tried both KSEXT and SEXT, but the results were not so different with each other. I am now focusing on the linear chromaticity comparison so that I can forget about the higher order issue.

Still I somehow got inconsistent result as I mentioned. I am looking for what I am still missing in the modification of the DA run, and hope to make my question clearer... Sorry for the unclear question now.

[michael_borland]

Takahiro,

If you want to post your files (or send them to me at borland@aps.anl.gov) I can have a look.

--Michael

[twatanabe]

Michael,

Thanks, again. Attached is the files I am looking at.
Although I said I optimized "closed_orbit_interations" and "iteration_fraction", it seems like I did not; closed obits did not converge well even after 500000 iterations with the iteration_fraction of 0.25. I tried other options like fixed_length, but did not see improvement.

Thank you,

Takahiro