Elegant users forum

Dear Michael Borland

Thank for your reply.
I have two lattice file. One is bare lattice with 11 mm vertical gap. Other contain the undulator with 6mm vertical gap.
The second result get from second lattice with 20 undulator. I attached it.
I think that it is relate with twenty undulator. Because the effects don't comes from resonance line. But I not certain.
Sorry for bother you. Please check.

Best regards. Have a nice day.

From Hwang, Ji-Gwang

Ji-Gwang,

Can you also post the input files for systematic and random multipoles and any other input files you used?

--Michael

Dear Michael Borland

I'm post it. Please check.

Sincerely yours.

From Hwang, Ji-Gwang

Dear Michael Borland

I'm post it. Please check.

Sincerely yours.

From Hwang, Ji-Gwang

JI-Gwang,

The problem is with the WIGGLER elements. This is a matrix implementation that should not be used for ring tracking. You need to use UKICKMAP or CWIGGLER.

--Michael

Dear Michael Borland

Thans to your reply. I will try to check it.
Always thank you for your helpful reply.

Have a nice day. Best regards.

Hwang,Ji-Gwang

Hi,

In my opinion, it is very difficult to identify the right tune in every case, especially when there is coupling and lattice errors. The highest peak of spectrum does not mean it is the tune.

In the attachment, is the spectrum of x and y coordinate, 2048 turns, and the initial amplitude of x is ten times of y. The red is horizontal. The lattice I was using has various roll/misalignment errors. So the horizontal and vertical have identical frequency due to the coupling. But the power(height of the spectrum) are different.

Can you guess the true tune ? or the zero-amplitude twiss tune ? it is (.40, .33). the nux=.40 has no doubt. But please take a look at vertical tune around .33, the peak is so weak.

Of course we can apply different intelligence to identify the true tunes, but in general it is not easy to find a set of rules which applies to all cases. e.g. we can narrow down the tune range, but at high chromaticity it does not work well for the plots in x-dp/p plane.


Lingyun

Lingyun,

This is a very good point and a nice illustration of how difficult it is to determine the tunes when there is coupling.

Some years ago, I added tracking-based computation of tune shift with amplitude and tune footprint, as a way to perform optimization of nonlinear properties of rings. It failed for just the reason you describe. When the amplitude got large, it was hard to determine the tunes, but it is the large amplitudes that we are most interested in.

--Michael