Synchronous phase in Toutatis

[emre314]

Dear Romuald,

I am designing a 200Mhz RFQ in Toutatis whose cell structure is coming from PARMTEQ PARI.

I observed that there is a discrepancy between Parmteq and Toutatis results especially in longitudinal distribution even though I used the same input cell parameters (Parmteq=0.410 Mev-deg, Toutatis=0.210 Mev-deg). again transverse distribution --> Parmteq = 0.163 mm.mrad, Toutatis-->0.189 mm.mrad

I realized that although the synchronous phase in the input file is from -90 to -30, Toutatis generates a different trend of phase. Please check syncPhase pictures and rfq.dat file.

I would like to transfer geometry from Toutatis to Tracewin. I know that DatFlag comment is the key parameter. Interestingly, Synchronous phase trends change depending on DatFlag comment. Please check phase_with_datflag.png and phase_without_datflag.png. As a result, I could not transfer the data file to Tracewin properly.

Is there something that I made wrong or do you have any advice to me?

Thanks,

Emre

[RomualdDuperrier]

Dear Emre,

I am sorry for the delay in my response but I finally managed to analyse your problem. It seems that you are using a relatively old version of Toutatis. With this version, the code would either read the synchronous phase from your inp file or calculate it by analyzing the cell length. In principle the second method should give similar values but depending on how the cell lengths were set in inp, you could observe discrepancies. Finally, with this old version, it seems that a different behavior is triggered (first or second way to display the sync phase) if you enable or disable the datflag.

I am not capable of being more precise about the behavior of your version because I did not keep the source code for this old version. In general, Didier and I recommend to use the latest version because some bugs could have been fixed. By the way, the last version does not plot the sync phase anymore but just the phase shift between beam center and the sync phase read in inp file.

It is important to note one point: Toutatis is a t-code and does not use the sync phase value in inp at all for simulating the beam in the structure. Only geometrical parameters, voltage and input beam parameters should impact your simulation. When you ask for a dat file for TraceWin, it is the sync phase read in inp file that should be transferred to dat file.

I hope it helps

Romuald

[emre314]

Dear Romuald,

Thanks so much for your pieces of advice. I updated the new version of Toutatis and transferred the data file to Tracewin successfully. As a result, I solved my problem.
Now I have another issue about Toutatis. Is it a good idea I am not sure but I am comparing the long. and trans. emittance between Parmteq and Toutatis and I observe some discrepancies between them. As I mentioned before Toutatis cell structure and input distribution come from Parmteq Pari. I found that especially one long. emittance is almost double other one. (Parmteq -->0.410 MeV.deg, Toutatis -->0.23 MeV.deg). Again transverse emittance (Parmteq -->0.16 mm.mrad, Toutatis -->0.18 mm.mrad). Lastly transmission (Parmteq -->97.60% , Toutatis -->98.75%[accelerated])
I tried to apply a few filter method and I could not change the results. I attached the input file so that you may check it.

I need to say that I don't perform any optimization in Parmteq because the rfq is already designed in Parmteq and manufactured before.

I would be grateful if you have any idea or comments about it.

Bests,
Emre Cosgun

[RomualdDuperrier]

Dear Emre,

I ran rfqtest.inp with Toutatis 2.0 and got similar results. In details:

- the transverse emittance is a bit bigger with Toutatis compared to Parmetq. This seems to be consistent with the greater transmission with Toutatis. Toutatis considers that a particle is lost when it hits the electrodes when, if I remember well, Parmteq considers that a particle is lost when it gets out of a square of width = 2 x a with a the minimum aperture. The Parmteq criterion is less realistic and induces more losses, the transverse halo is then cleaned up artificially. The reason why Parmteq is limiting the particle motion to this square is because the analytical formulation used by Parmteq (8 terms) for the electric potential become wrong when you get out of this square.

- for the longitudinal emittance, I have to confess that I do not understand the importance of the difference, almost a factor 2. It seems excessive to me based on my knowledge of the two codes. I remember I investigated the impact of the Parmteq motion integrator that is not simplectic, but it resulted in only artificial longitudinal emittances by 10%. Could you share a longitudinal phase space portrait with me from Parmteq simulation? It might ring a bell. Also, have you tried playing the step size with both codes?

I hope it helps,

Romuald