Elegant users forum

Hi, Michael,

I use elegant to do the frequency map analysis to a small ring with only linear elements (dipoles and quadrupoles only). But I saw a strange phenomenon which look like nonlinear effects. The tune change a lot VS the horizontal displacement.

The DA is also nonlinear like. It's pretty strange to me.

Then, I did the single particle tracking to observe the horizontal phase space and vertical phase space. Still the same, it looks that there are some nonlinear effect which I don't where come from.

I checked my code for many times but still can't explain this phenomenon. Could you please give me some suggestions on my code? Is there any mistakes in the using of the program?

Thanks a lot!

Regards,
Haisheng Xu

Haisheng,

There are nonlinearities that come about as a result of the smaller bending radius. The leading term is sextupole-like and scales as 1/rho^3 . You can turn these off by setting NONLINEAR=0 on the CSBEND elements, but it isn't physical. Small radius bends have very strong chromaticity effects, which will make chromaticity correction a struggle.

Having said that, your FMA didn't look bad (attached).

--Michael

Michael,
Thank you for reply!

Your result is better. Could you please post your file on line? I always observe the DA at the smallest \beta position. So I didn't see such large DA before.

Furthermore, could you explain a little about your model of the dipole ( and also the fringe field)? Or are there any references? I want to understand this in detail. As you know, the model of dipole magnet is different in small rings (because of the small bending radius).

Thanks a lot!

Haisheng

Haisheng,

My files are attached.

I've also attached a technical note that explains a little about how the dipoles are modeled. It doesn't describe the fringe field treatment, which at this point is pretty basic. You can either choose a purely linear edge focusing (sympletic) or a second-order matrix (non-symplectic). Some day I'd like to improve this.

--Michael

Michael,

Looks good! Thank you sooooo much!

Regards,
Haisheng

Hi, Michael,

I try to add "nonlinear=0" to the csbend. But the DA is still nonlinear like.

The lattice file is attached.

It seems that this flag doesn't work for me. Could you please have a look at my settings? Thanks!


Regards,
Haisheng

How was the beautiful picture plotted? what kind of sddsplot command for these 3 line data

gwei wrote:How was the beautiful picture plotted? what kind of sddsplot command for these 3 line data

I remember that there is no one-line command which can be used to plot the pseudo-color figure. Let me try to look for the script I used.

Haisheng

The plot command I use for FMA data is The -split option splits the data into subpages based on the values of the diffusionRate column. The -graph option sets up variation on the symbol "subtype" (i.e., color); it also chooses a square symbol and filling of the symbols; the "scale=2" parameter may need adjustment depending on how fine the FMA grid is. The "-order=spectral" option sets the order of symbol colors; you can also try "-order=rspectral".

--Michael

michael_borland wrote:The plot command I use for FMA data is

Code: Select all

sddsplot run.fma -column=x,y -split=column=diffusionRate -graph=sym,vary=subtype,scale=2,fill,type=2 -order=spectral 

The -split option splits the data into subpages based on the values of the diffusionRate column. The -graph option sets up variation on the symbol "subtype" (i.e., color); it also chooses a square symbol and filling of the symbols; the "scale=2" parameter may need adjustment depending on how fine the FMA grid is. The "-order=spectral" option sets the order of symbol colors; you can also try "-order=rspectral".

--Michael

Thank you very much!