Page 1 of 1

Discrepancy between momentum aperture and Touschek

Posted: 14 Feb 2022, 00:42
by A.V.Bogomyagkov
Hello,

I calculated momentum aperture, ibs emittances with ibsEmittance and Touschek lifetime with touschekLifetime.

Then I plotted momentum aperture and square root of tmP and tmN, which look different from momentum aperture on attached figure. The difference is clearly visible while comparing the mmap and tlife files. The number of elements is identical.
Why is the difference? It looks like there is a some type of smoothing.

The second question is an attempt to understand the loss rate plot. The peak at s=588 m has minimum of H and minimum of dispersion comparing with immediate surrounding. What is the reason for the peak?

The beam energy is 1.5 GeV. I am using "elegant 2021.4.0 Nov 18 2021".
I attached the sorted mmap, tlife, and twi files.

Thank you,
Anton

Re: Discrepancy between momentum aperture and Touschek

Posted: 15 Feb 2022, 13:32
by michael_borland
Anton,

I can't reproduce your results. For this energy, sqrt(tmP) and -sqrt(tmN) are deltaPositive and deltaNegative. If I plot those, the agreement is essentially perfect.
comparison.png
The FN and FP columns are misleading, since they are not integrated vs s. If you do that, the spike doesn't look so significant.
integrals.png
For this plot, I normalized FN and FP to the circumference, then integrated vs s, so the units are 1/s.

--Michael

Re: Discrepancy between momentum aperture and Touschek

Posted: 16 Feb 2022, 22:27
by A.V.Bogomyagkov
Thank you, Michael,

It is puzzling. I read sdds files into python pandas and plot them. The difference is obvious.

I also converted sdds files into ascii (both files are in the new attachment)

Code: Select all

sddsconvert -ascii results/ct_ring.tlife results/ct_ting.tlife.ascii
sddsconvert -ascii results/ct_ring.mmap.sorted results/ct_ring.mmap.sorted.ascii
and compare mmap and tlife for element and the difference is visible

L2.MQ23 (s = 8.198064599999999e+01)
in mmap: deltaPositive = 5.000000000000000e-03
in tlife: tmP =5.424767157239484e-05 and the square root is sqrt(5.424767157239484e-05) = 0.0073653018656668

R1.MQ20 (s = 4.444997969999995e+02)
in mmap: deltaPositive = 1.4e-2
in tlife: tmP = 1.575137583744795e-04 and the square root is sqrt(1.575137583744795e-04) = 0.01255044853280071

For this energy
beta=math.sqrt(1 - ((me_mks * c_mks * c_mks) / (e_mks * 1.5e9))**2 ) = 0.9999999419733219
and doesn't change much.

Even if I use sdds tools to create new columns sqrt_tmP and sqrt_tmN I still get the same results

Code: Select all

sddsprocess results/ct_ring.tlife -pipe=out \
-define=col,sqrt_tmP,"tmP sqrt" \
-define=col,sqrt_tmN,"tmN sqrt" \
| sddsconvert -pipe=in results/ct_ring.tlife_new

!sddsconvert -ascii results/ct_ring.tlife_new results/ct_ting.tlife_new.ascii
I use SDDS_ToolKit-x64.msi version 5.1

Integration of FP and FN is a good comment. I will implement this.

Anton

Re: Discrepancy between momentum aperture and Touschek

Posted: 03 Jun 2022, 03:28
by Skamarokha
Michael,

I have reproduced the problem mentioned above by Anton with different lattice. Maybe the problem is in touschekLifetime program?

Here is version of the program I am using on Windows:

Code: Select all

 Program by A. Xiao, M. Borland.  (This is version 8, March 2017, M. Borland)