CSRDRIFT after WIGGLER element

Moderators: cyao, michael_borland

nakamura
Posts: 13
Joined: 09 Sep 2011, 20:35

CSRDRIFT after WIGGLER element

Post by nakamura » 16 Aug 2020, 07:22

In the attached figure of the momentum spread along the beam path, there are two jumps pointed by red arrows.
Both jumps occur at L12 elements (CSRDRIFT elements with LSC on) just after two undulators made of WIGGLER
elements(UND0) and LSCDRIFT elements having the same effective length (LSCUND), which are defined below.

UND0: WIGGLER,L=0.024,K=1.421,POLES=2
LSCUND: LSCDRIFT,LEFFECTIVE=0.024,BINS=100

When I replaced one of the L12 elements with an LSCDRIFT element LL12, the sudden jump of the momentum spread
disappeared at LL12 and such a jump appeared at the CSRDRIFT element just after LL12 instead of LL12.
I cannot understand why these jumps of the momentum spread appear at CSRDRIFTs just after WIGGLER elements.
Please let me know the reason.
Thank you in advance.

Norio
Attachments
elegant_forum.png

michael_borland
Posts: 1927
Joined: 19 May 2008, 09:33
Location: Argonne National Laboratory
Contact:

Re: CSRDRIFT after WIGGLER element

Post by michael_borland » 07 Sep 2020, 11:42

Norio,

The reason for this is that CSRDRIFT tries to compensate for the missing effect of CSR in elements that don't explicitly model CSR. In drifts, most of the effect is on the longitudinal plane, so CSRDRIFT back-drifts to the end of the last CSRDRIFT or CSRCSBEND, then integrates forward. In this way, one can include CSR effects in straight beamlines even if they have other components besides CSRDRIFT.

--Michael

nakamura
Posts: 13
Joined: 09 Sep 2011, 20:35

Re: CSRDRIFT after WIGGLER element

Post by nakamura » 10 Sep 2020, 04:22

Michael,

Thank you very much for your kind answer.
However I did not well understand your explanation yet.
There is a CSRDRIFT element just before the 1st(or 2nd) undulator made of many WIGGLER elements in my lattice file.
So I think the 1st(or 2nd) jump of the momentum spread at the CSRDRIFT element just after the 1st(or 2nd) undulator in my plot
is caused by the integrated WIGGLER elements. Is it true ?
WIGGLER elements don't explicitly model CSR in your ELEGANT manual.
Do they implicitly model CSR and calculate the CSR effects in simulations ?
Thank you in advance for answering my new questions.

Norio

michael_borland
Posts: 1927
Joined: 19 May 2008, 09:33
Location: Argonne National Laboratory
Contact:

Re: CSRDRIFT after WIGGLER element

Post by michael_borland » 10 Sep 2020, 14:30

nakamura wrote:
10 Sep 2020, 04:22
There is a CSRDRIFT element just before the 1st(or 2nd) undulator made of many WIGGLER elements in my lattice file.
So I think the 1st(or 2nd) jump of the momentum spread at the CSRDRIFT element just after the 1st(or 2nd) undulator in my plot
is caused by the integrated WIGGLER elements. Is it true ?
WIGGLER elements don't explicitly model CSR in your ELEGANT manual.
Do they implicitly model CSR and calculate the CSR effects in simulations ?
Norio,

Yes, that's basically what happens. For example, suppose you have

Code: Select all

CSRCSBEND-CSRDRIFT1-WIGGLERs-CSRDRIFT2-CSRDRIFT3
CSR computations are performed for the CSRCSBEND and CSRDRIFT1. Inside the WIGGLERs, there are no CSR effects computed. However, when CSRDRIFT2 is seen, ELEGANT retroactively applies CSR over the intervening length from CSRDRIFT1. It appears as a jump in the CSR-induced energy spread at the end of the CSRDRIFT2, in this example. This will happen with any intervening elements that don't include CSR.

Assume that total length of the wigglers is LW. The way it works internally is that when CSRDRIFT2 is seen, ELEGANT performs a pure drift of length -LW, which takes beam from the beginning of CSRDRIFT2 to the exit of CSRDRIFT1. It then models CSRDRIFT1, but with the length increased by LW.

--Michael

nakamura
Posts: 13
Joined: 09 Sep 2011, 20:35

Re: CSRDRIFT after WIGGLER element

Post by nakamura » 30 Sep 2020, 01:29

Michael,

I am sorry for my late reply.
Thank you very much for your kind explanation about the jumps of the momentum spread.
I understood that the jumps are caused by CSR effects from the CSRCSBEND elements upstream
of the WIGGLER elements because no CSR effects are not computed in the WIGGLER elements.
If so, I wonder why the jumps are so large. The distance from the last CSRCSBEND to the 1st WIGGLER
element is long (about 20 m) and most of the beam-line elements between them are CSRDRIFT elements,
where the CSR effects are computed and then reflected in the momentum spread.
I think the CSR effects induced from the CSRCSBEND elements are very weak at the positions of
the WIGGLER elements. I would like you to answer the question or comment it.
Thank you in advance.

Norio

michael_borland
Posts: 1927
Joined: 19 May 2008, 09:33
Location: Argonne National Laboratory
Contact:

Re: CSRDRIFT after WIGGLER element

Post by michael_borland » 06 Oct 2020, 14:07

Norio,

I can't comment specifically on the magnitude of the jumps in your simulation. It can be tested by replacing the CWIGGLER element temporarily with a CSRDRIFT element. The energy spread at the end of this element and downstream should be the same in both cases. Please try this and let me know if you see any discrepancies.

--Michael

nakamura
Posts: 13
Joined: 09 Sep 2011, 20:35

Re: CSRDRIFT after WIGGLER element

Post by nakamura » 20 Oct 2020, 07:13

Michael,

I replaced WIGGLER (or CWIGGLER) elements with CSRDRIFT elements.
The simulation results are shown In the attached figure of the momentum
spread along the beam path.
Fig. (a) is for the WIGGLER case and Fig. (b) is for the CSRDRIFT case.
In Fig. (a), there are two jumps pointed by red arrows at a CSRDRIFT element
just after the WIGGLER elements as I said in the previous mail, while in Fig. (b)
there are no jump. The two energy spreads are clearly different at the end of
the beamline. The WIGGLER elements seem to affect the momentum spread.
Please let me know the reason if you know. Thank you in advance.

Norio
Attachments
elegant_forum_nn_20201020.png

michael_borland
Posts: 1927
Joined: 19 May 2008, 09:33
Location: Argonne National Laboratory
Contact:

Re: CSRDRIFT after WIGGLER element

Post by michael_borland » 28 Oct 2020, 14:01

Norio,

Unless you have synchrotron radiation turned on in the CWIGGLER, this surprises me. Can you upload some input files so I can try it?

--Michael

nakamura
Posts: 13
Joined: 09 Sep 2011, 20:35

Re: CSRDRIFT after WIGGLER element

Post by nakamura » 28 Oct 2020, 21:31

Michael,

Thank you very much for your attention and reply.
I send the elegant input files for the WIGGLER and CSRDRIFT cases.
I am sorry these files are heavy in size and not well organized.
The betatron functions and QM K-values are different for the two cases
around the WIGGLER or CSRDRIFT elements because the optics
should be modified without the WIGGLER elements (for the CSRDRIFT case)
to suppress divergence of the betatron function.

Norio
Attachments
elegant_forum_20201029nn.zip
(8.44 MiB) Downloaded 234 times

michael_borland
Posts: 1927
Joined: 19 May 2008, 09:33
Location: Argonne National Laboratory
Contact:

Re: CSRDRIFT after WIGGLER element

Post by michael_borland » 09 Nov 2020, 10:56

Norio,

I think the problem is that you are including LSC, which depends on the beam size. When elegant gets to the end of the wiggler, it adds in all the "missing" CSR and LSC. Since the beam sizes are very different in the case with and without the wiggler, you get different longitudinal wakes from LSC.

When I disabled LSC, the two runs were nearly identical.

--Michael

Post Reply