Unexpected change in particle momentum with skew quadrupole

[Galnander]

Dear Michael,

We are using Elegant for slow extraction simulations of SIS100 at FAIR. When introducing a skew quadrupole element to study coupling-induced resonant losses, we observe that vertically lost particles sometimes show a significantly modified total momentum, p, in the loss file, 1.43 instead of around 1.02. See the attached files.

Do you have any idea what may cause this effect?

The skew quadrupole in question is, line 153:
Z14KM1SQ: MULT, ORDER=1, KNL=0.03, TILT=0.7854

The same effect occurs also when I apply a TILT to one of the main quadrupoles (KQUAD).

The working point is intentionally chosen to excite the nonlinear Qx – 3Qy skew resonance. (Our resonant and chromatic sextupoles are also excited.) This could redistribute transverse actions, x-y, and lead to losses at aperture limitations in y, which is what we observe. But it should not give a change in the total momentum. We have not included any elements that may change the longitudinal momentum, (RF cavities, or so), to my knowledge.

Also, a suggestion, please add the pCentral parameter to the loss file as well.

I am using Elegant and Pelegant version 2025.3.0, under Windows, SVN revision: 34285

Best regards, and thanks in advance,
Björn

[michael_borland]

Bjorn,

That's a very old bug that went unnoticed for years, since elegant is rarely used for such low values of momentum. The error is that some loss mechanisms result in caching of gamma instead of p=beta*gamma. I'll fix this in the next release, along with adding pCentral to the parameters.

In the mean time, here's a quick fix

Code: Select all

sddsprocess ko_emi28_track_losses.lost "-redefine=column,p,1.2 p < ? sqr 1 - sqrt : $ "

In words, if p>1.2, then replace by sqrt(p^2-1), otherwise leave unchanged.

Thanks for reporting the bug.

--Michael

[Galnander]

Dear Michael,

Thanks for fast answer and follow up.

Best regards,
Björn

[Galnander]

Dear Michael,

Another thing. Could you please also add the 'ElementName' of the lost particles, in the loss file? Now only 's' is saved.

That would be convenient.

Best regards,
Björn