High-resolution dynamic aperture

[michael_borland]

The find_aperture command can be used for finding dynamic aperture. It uses various search methods to find the DA boundary.

Another method is just to perform tracking with a beam that fills a large rectangle in x-y space. By using the losses parameter of the run_setup command, one can get a record of the lost particles and cross-reference to the initial coordinates using sddsxref (making use of the particleID data in the bunch output file and losses file). This allows making a map of the lost particles color-coded by the turn on which the loss occurs.

Attached are some files that demonstrate how to do this. The script elegantRingAnalysis allows doing this on a cluster, which gives very rapid turn-around.

--Michael

[petrenko]

Thanks for the example!

Do you have plans to include higher order components (Kn, n>4) in CSBEND element?

[michael_borland]

Thanks for the example!

Do you have plans to include higher order components (Kn, n>4) in CSBEND element?

This is the first request I've had for this. I'll add it to my list of things to do. Out of curiousity, is there a particular motivation or application?

--Michael

[petrenko]

I thought that for dynamic aperture the strongest effect comes from the field errors in bends. So if you have these SYSTEMATIC_MULTIPOLES and RANDOM_MULTIPOLES up to an arbitrary order in quadrupoles why not have them in dipoles?

I'm quite new to dynamic aperture calculations. For electron machines probably the high-order resonances are not very important because of random kicks from synchrotron radiation. My current job is to see if the proposed magnet design is good enough for the NESR ion storage ring (this is a heavy-ion machine to be build in Darmstadt, Germany). I actually have a list of field errors up to 8-th order from the field simulation program -- just don't know what to do with the rest of them after N=4. Maybe it's not very important since it would be difficult to measure this kind of errors in dipoles.

Also I think it's easier to work then you have all the systematic and random field errors stored in sdds-files.

[michael_borland]

I thought that for dynamic aperture the strongest effect comes from the field errors in bends. So if you have these SYSTEMATIC_MULTIPOLES and RANDOM_MULTIPOLES up to an arbitrary order in quadrupoles why not have them in dipoles?

I'm quite new to dynamic aperture calculations. For electron machines probably the high-order resonances are not very important because of random kicks from synchrotron radiation. My current job is to see if the proposed magnet design is good enough for the NESR ion storage ring (this is a heavy-ion machine to be build in Darmstadt, Germany). I actually have a list of field errors up to 8-th order from the field simulation program -- just don't know what to do with the rest of them after N=4. Maybe it's not very important since it would be difficult to measure this kind of errors in dipoles.

Also I think it's easier to work then you have all the systematic and random field errors stored in sdds-files.

For electron rings for light sources (what I normally deal with), it is usually the sextupoles that dominate the dynamic aperture, so we tend to think less about higher order terms than others might.

Adding arbitrary higher order multipoles for dipoles is not straightforward as it is for other elements, because of the curvilinear coordinates. However, adding a few more orders is possible and is on my list as an optional feature.

Thanks for the suggestion.

--Michael

[wguo]

There is a way to add multipole errors for the dipoles--by using the MULT element. It can be used as a zero length kick element. Even though dipole magnets might be long , the multipoles are really generated by the fringe fields. From that aspect, a kick element might be ok.

The multipole field errors start to affect the dynamic aperture at very large transverse displacement. For dynamic aperture we usually track with 20~30 mm betatron oscillation amplitude. The off-momentum particles have bigger displacement at non-zero dispersion region. Therefore the dynamic aperture of the off-momentum particles are affected by the multipole errors first.

At NSLS-II we see some effect on the DA now that our peak dispersion is about 0.5 m.

[OZGUR]

Hi,

I provided a plot for DA and it gives a result as it can be seen in attached file.
I am just wondering how I can make it with more points and colors?

Thank you,

Özgür

[michael_borland]

Ozgur,

I think you want to zoom in on the DA boundary. If so, use the beam size controls on the &bunched_beam command, e.g., reduce the emittances specified by emit_x and emit_y.

--Michael

[OZGUR]

Hi Micheal,

I also increased the particle number in addition changing the emit_x and emit_y and it worked, thank you.

Best,

Özgür

[gwei]

Hi wguo,

Could you attach an example with multipole errors of the dipoles. Thank you.

Guohui

There is a way to add multipole errors for the dipoles--by using the MULT element. It can be used as a zero length kick element. Even though dipole magnets might be long , the multipoles are really generated by the fringe fields. From that aspect, a kick element might be ok.

The multipole field errors start to affect the dynamic aperture at very large transverse displacement. For dynamic aperture we usually track with 20~30 mm betatron oscillation amplitude. The off-momentum particles have bigger displacement at non-zero dispersion region. Therefore the dynamic aperture of the off-momentum particles are affected by the multipole errors first.

At NSLS-II we see some effect on the DA now that our peak dispersion is about 0.5 m.