Transverse impedance discrepancy with measurement
Moderators: cyao, michael_borland
-
- Posts: 49
- Joined: 13 Oct 2016, 09:17
Transverse impedance discrepancy with measurement
Hi Michael,
We are facing the discrepancy between the measurements and simulations in the vertical plane while studying the head-tail instability for NSLS-II. We use the equilibrium bunch length for a particular beam current and then add 0.1mm transverse kick. Strangely, we also need a very high shunt resistance in the vertical plane to fit the experimental data.
We see the same behavior with feedback in the vertical plane.
I’ve attached the files for your reference. Can you please check what might be the reason?
Here is the damping rate comparsion b/w measurements and simulations. Thank you,
Aamna
We are facing the discrepancy between the measurements and simulations in the vertical plane while studying the head-tail instability for NSLS-II. We use the equilibrium bunch length for a particular beam current and then add 0.1mm transverse kick. Strangely, we also need a very high shunt resistance in the vertical plane to fit the experimental data.
We see the same behavior with feedback in the vertical plane.
I’ve attached the files for your reference. Can you please check what might be the reason?
Here is the damping rate comparsion b/w measurements and simulations. Thank you,
Aamna
- Attachments
-
- coupling_lattice.zip
- (5.91 MiB) Downloaded 319 times
-
- Posts: 1945
- Joined: 19 May 2008, 09:33
- Location: Argonne National Laboratory
- Contact:
Re: Transverse impedance discrepancy with measurement
Aamna,
One problem I see is that the lattice you use (lattice.lte in ILMatrixFromTracking) has default values for N_KICKS on the CSBEND, KQUAD, and KSEXT elements. As described in the manual, this isn't typically sufficient to reproduce the beam dynamics accurately. Typically I would use N_KICKS=20 for dipoles, 30 for quads, and 10 for sextupoles. Using too few kicks will result in inaccurate determination of the tunes, chromaticities, etc.
You seem to be modeling the resistive wall impedance with a broad-band resonantor, which I wouldn't think would be a good choice. You are better using a known functional form for the short-range resistive-wall impedance. For example, you could use ImpedanceWake2D (from CERN) to compute the resistive wall impedance.
--Michael
One problem I see is that the lattice you use (lattice.lte in ILMatrixFromTracking) has default values for N_KICKS on the CSBEND, KQUAD, and KSEXT elements. As described in the manual, this isn't typically sufficient to reproduce the beam dynamics accurately. Typically I would use N_KICKS=20 for dipoles, 30 for quads, and 10 for sextupoles. Using too few kicks will result in inaccurate determination of the tunes, chromaticities, etc.
You seem to be modeling the resistive wall impedance with a broad-band resonantor, which I wouldn't think would be a good choice. You are better using a known functional form for the short-range resistive-wall impedance. For example, you could use ImpedanceWake2D (from CERN) to compute the resistive wall impedance.
--Michael
-
- Posts: 49
- Joined: 13 Oct 2016, 09:17
Re: Transverse impedance discrepancy with measurement
Hi Michael,
Thank you for your response! I increased the n_kicks in the lattice file as you suggested, and I don’t see any major change.
Aditionally, I’m not able to understand why the initial vertical beam size is not equilibrium with or without impedance. I observe this behavior for a lattice with transverse coupling or without coupling. I suspect this might also be the reason for the difference.
Best regards,
Aamna
Thank you for your response! I increased the n_kicks in the lattice file as you suggested, and I don’t see any major change.
Aditionally, I’m not able to understand why the initial vertical beam size is not equilibrium with or without impedance. I observe this behavior for a lattice with transverse coupling or without coupling. I suspect this might also be the reason for the difference.
Best regards,
Aamna
-
- Posts: 1945
- Joined: 19 May 2008, 09:33
- Location: Argonne National Laboratory
- Contact:
Re: Transverse impedance discrepancy with measurement
Aamna,
The damping time is 20,000 turns, so you would need to track longer to see equilibrium. Here, for example, is 40,000 turn tracking with no coupling and charge=0. The vertical emittance is trending to zero as expected. Also, I noticed that the files you posted are not using the latest version of the script (ILMatrixFromFmap) to determine the ILMATRIX parameters from tracking. It's available in the elegant examples file, but I've attached it for your convenience.
--Michael
The damping time is 20,000 turns, so you would need to track longer to see equilibrium. Here, for example, is 40,000 turn tracking with no coupling and charge=0. The vertical emittance is trending to zero as expected. Also, I noticed that the files you posted are not using the latest version of the script (ILMatrixFromFmap) to determine the ILMATRIX parameters from tracking. It's available in the elegant examples file, but I've attached it for your convenience.
--Michael
- Attachments
-
- coupling_lattice.zip
- (9.48 MiB) Downloaded 277 times
-
- Posts: 49
- Joined: 13 Oct 2016, 09:17
Re: Transverse impedance discrepancy with measurement
Thank you for your response, Michael.
I still don't see the equlibrium beam size for coupling lattice. The initial bunch size using the average beta function of 13.8m for emittance = 30pm is ~20 microns, but ILMATRIX gives betay = 3.366872m. How does elegant compute the initial vertical distribution from the vertical emittance in this case?
We want to compare the elegant results with internal code SPACE. We obtain the following results:
In SPACE, before the dipole kick at 10k turns, the bunch is in equilibrium.
After the kick, the bunch centroid envelope decays with a damping time equal to the radiation damping time (54ms), while the bunch size envelope (substracting the equilibrium value) decays with a damping time equal to half the radiation damping time (27ms).
Please correct me, if I'm interpreting/doing something wrong.
Best,
Aamna
I still don't see the equlibrium beam size for coupling lattice. The initial bunch size using the average beta function of 13.8m for emittance = 30pm is ~20 microns, but ILMATRIX gives betay = 3.366872m. How does elegant compute the initial vertical distribution from the vertical emittance in this case?
We want to compare the elegant results with internal code SPACE. We obtain the following results:
In SPACE, before the dipole kick at 10k turns, the bunch is in equilibrium.
After the kick, the bunch centroid envelope decays with a damping time equal to the radiation damping time (54ms), while the bunch size envelope (substracting the equilibrium value) decays with a damping time equal to half the radiation damping time (27ms).
Please correct me, if I'm interpreting/doing something wrong.
Best,
Aamna
-
- Posts: 1945
- Joined: 19 May 2008, 09:33
- Location: Argonne National Laboratory
- Contact:
Re: Transverse impedance discrepancy with measurement
Aamna,
The initial distribution is prepared using the lattice functions at the beginning of the lattice, since that is where the beam starts. Since you are looking at data from a WATCH element, elegant should provide data consistent with the lattice functions at the location of the WATCH element. The average beta function does not come into play at all, except insofar as you would need to weight your transverse impedance with it.
Since you didn't say how you are introducing coupling, it is hard for me to comment on that. I see the vertical emittance damp to zero in the uncoupled case.
--Michael
The initial distribution is prepared using the lattice functions at the beginning of the lattice, since that is where the beam starts. Since you are looking at data from a WATCH element, elegant should provide data consistent with the lattice functions at the location of the WATCH element. The average beta function does not come into play at all, except insofar as you would need to weight your transverse impedance with it.
Since you didn't say how you are introducing coupling, it is hard for me to comment on that. I see the vertical emittance damp to zero in the uncoupled case.
--Michael
-
- Posts: 49
- Joined: 13 Oct 2016, 09:17
Re: Transverse impedance discrepancy with measurement
Hi Michael,
For coupling, random errors were added to skew quadrupoles with names beginning from “sq” in the .lte file (originally set to zero) to adjust the vertical emittance to 30 pm.
For uncoupled lattice, vertical emittance is set to ~8pm.
Thanks,
Aamna
For coupling, random errors were added to skew quadrupoles with names beginning from “sq” in the .lte file (originally set to zero) to adjust the vertical emittance to 30 pm.
For uncoupled lattice, vertical emittance is set to ~8pm.
Thanks,
Aamna
- Attachments
-
- lattice.lte
- (142.13 KiB) Downloaded 263 times
-
- Posts: 49
- Joined: 13 Oct 2016, 09:17
Re: Transverse impedance discrepancy with measurement
Hi Michael,
Just a reminder, can you please check the attached file to comment on the vertical emittance damping with coupling?
Thanks,
Aamna
Just a reminder, can you please check the attached file to comment on the vertical emittance damping with coupling?
Thanks,
Aamna
-
- Posts: 1945
- Joined: 19 May 2008, 09:33
- Location: Argonne National Laboratory
- Contact:
Re: Transverse impedance discrepancy with measurement
Aamna,
You need to use element-by-element synchrotron radiation for tracking if you except consistency with the 6D beam moments calculation. You can't use SREFFECTS in this case. Also, we aware that the horizontal and vertical emittances computed by the WATCH element are the projected emittances. They won't agree with the emittance of mode 2 reported by moments_output except in the case that the beam has no x-y tilt.
--Michael
You need to use element-by-element synchrotron radiation for tracking if you except consistency with the 6D beam moments calculation. You can't use SREFFECTS in this case. Also, we aware that the horizontal and vertical emittances computed by the WATCH element are the projected emittances. They won't agree with the emittance of mode 2 reported by moments_output except in the case that the beam has no x-y tilt.
--Michael
- Attachments
-
- example.zip
- (18.36 KiB) Downloaded 124 times