Impossible to retrieve Twiss parameters from JSON files if dispersion is not zero.

[jureyherrera]

Hello,

If the e-beam dispersion is not zero, there is a loss of information since we cannot retrieve the Twiss parameters from the Syned JSON files.

Please find below a PDF with more details about this issue, including a possible solution. Please let me know your thoughts on the matter.

Twiss_syned.pdf

Thank you in advance for your attention.

Cheers, J

[lucarebuffi]

This is a simple mistake, I think the whole thing just slipped. Would you meet me one-on-one and implement it?

[srio]

Thanks Luca for the coding. Relations are not trivial.

There is a problem. Run:

    a = ElectronBeam.initialize_as_pencil_beam(energy_in_GeV=2.0,current=0.5, energy_spread=0.00095)
    a.set_moments_all(1e-6,9e-6,4e-6,16e-6,36e-6,25e-6)
    print(a.get_moments_all())

It gives NaNs.

After testing and further discussions with Juan, I would propose

The stored values in _moment* do not have dispersion terms, which are stored separately in _dispersion*

Therefore:

• set_moments*() and set_sigmas*() arguments are moments/sigmas WITHOUT dispersion (as always)
• get_momets*() and get_sigmas() return values WITHOUT dispersion
• We should add methods get_moments_all_with_dispersion() and get_sigmas_all_with_dispersio() that return values corrected with dispersion:

(the beta subscript means stored values without dispersion)

• set_twiss*() stores _moment* calculated WITHOUT dispersion using alpha and beta that do not contain dispersion by definition:

Note that the emittance is the "True" emittance, calculated WITHOUT dispersion:

In the case that the user has the "experimental values" of moments and emittances, i.e. values that do consider dispersion, the dispersion term has to be deduced from the moments before using set_moments*() (obtain and enter values with subindex "beta" here)

, and from the emittance before using set_twist():

In conclusion:

• Values with _moment* set/get_moment() set/get_sigma() DO NOT CONTAIN dispersion. set_moment or set_sigma receive values WITHOUT dispersion. If the user has experimental data (with dispersion), it has to be corrected.

• New methods should be created to return moments_with_dispersion and sigmas_with_dispersion (alternatively, to be discussed, is to use get_moments() WITH dispersion and get_moments_without_disperion() WITHOUT dispersion)

To be discussed....

[lucarebuffi]

The NaN problem is a stupid mistake I already fixed.
It all sounds too complicated to me, at the point that the user will have trouble understanding the behavior.
If you open the syned widget and work with Twiss parameters, you expect that the stored data reflect that. But instead, the stored data decided to ignore the user input, and the moments calculated from the user input are different to what is expected.
It's a big no no for me. To be discussed.

[lucarebuffi]

I discussed with Juan a little bit. I have a possible solution to do both the visions. Will implement it asap.