Elegant users forum

Dear Michael,

Just to express my question more clearly. Since "csbend" is canonical, it seems to me "csbend will get more correct estimation in anyway, small and big rings, is it right?

When I did the DA study, my results with "sbend" is more close to AT results but DA is x direction is only about 40% with "csbend" compare to "sbend". Since my usage of csbend is with non-zero k1 , does it really affect ?

Thanks a lot!
Chinchi

Chinchi,

CSBEND = Canonical Sector BEND. It uses a canonical integration technique (Ruth integrator) to model the dipole, giving symplectic behavior and energy-dependence to all orders.

SBEN = Sector BEND modeled using transport matrices, up to second order. It is fine for transport lines or computing chromatic properties of rings (e.g., twiss_output).

--Michael

ccwang wrote:Dear Michael,

Just to express my question more clearly. Since "csbend" is canonical, it seems to me "csbend will get more correct estimation in anyway, small and big rings, is it right?

When I did the DA study, my results with "sbend" is more close to AT results but DA is x direction is only about 40% with "csbend" compare to "sbend". Since my usage of csbend is with non-zero k1 , does it really affect ?

Thanks a lot!
Chinchi

Chinchi,

Yes, CSBEND will give the most accurate results. However, for big rings it isn't necessary and just makes the tracking slower.

I'm not sure why SBEND would agree more with AT, since I don't know how AT models dipoles. I would assume that it uses a canonical integration technique. For other rings, elegant CSBEND results have shown good agreement with AT and TRACY.

--Michael

Dear Michael,

In my case, I use CSBEND with non-zero k1 and k2. I guess probably it is the reason.

BTW, is there any example of off momentum particle DA study? I would like to check the particle behavior with +-3% energy difference. Thanks a lot!

Best wishes,
Chinchi

Chinchi,

Please see the attached files for an example of off-momentum DA.

--Michael