Page 1 of 1

Drop in longitudinal emittance in RFQ

Posted: Wed 25 Sep 2024 09:33
by jhustings
My simulation setup includes an ion source, a LEBT, and an RFQ. While running the simulation with Partran and Toutatis, I observed a discontinuity in the longitudinal emittance within the RFQ. Specifically, the emittance drops to a lower value midway through the RFQ and then recovers near the end. I have included a plot illustrating this behavior in the attachment. This phenomenon appears consistently in simulations involving 10,000 and 500,000 particles. However, when the simulation is run with only 1,000 particles, the discontinuity does not occur, and the emittance remains relatively low.

I am conducting these simulations using the Tracewin GUI on Windows, utilizing Toutatis. I included relevant files for the simulation with 10,000 particles.

Could you assist me in understanding the cause of this discrepancy?

Kind regards,
Jeroen

Re: Drop in longitudinal emittance in RFQ

Posted: Wed 25 Sep 2024 17:59
by Didier
Dear Jeroen,

As can be seen from your simulation below, a certain number of particles are detached in energy, making the longitudinal emittance larger and larger, until one of them is lost on the wall of the RFQ, abruptly resetting the emittance to a lower value. To overcome this problem in calculating emittances, it is necessary to exclude particles above a certain ΔW, for partan and for Toutatis (see picture below). This corrects the display problem.

On the other hand, nothing to do with this problem, I noticed that you were using the ‘Use aperture element (Envelope)’ option. This option was causing code stability problems in the current version, I've fixed this in the latest version on the site

Regards,

Didier
w.png
w.png (104.82 KiB) Viewed 85 times
options.png
options.png (69.93 KiB) Viewed 85 times

Re: Drop in longitudinal emittance in RFQ

Posted: Thu 26 Sep 2024 13:47
by jhustings
Dear Didier,

Thank you so much for clearing this out for us. This really helped us understand the problem and improve our simulations.

Cheers,

Jeroen