tuneshift with energy: tracking versus using twissparameters

[nadolski]

Dear All,

When I compute the tune shift with energy by tracking I am not able to reproduce the results obtained using the variation of the twiss parameter with energy

I have joined two scripts:

• offMomentumTwiss.tgz: Twiss parameters with energy (from Michael)
  runScan
  collateScan

  

  

• by tracking, I got very different curves. The half integer resonance is crossed around 5% (2.5% for Twiss)
  runTuneshiftwithenergy

  

I suspect something wrong in the script making the computation by tracking.

Thank you for your guidance.

Laurent.

[nadolski]

Somehow, I was not able to join the script in the previous post.

by tracking

tuneshiftbytracking.tgz

(581.83 KiB) Downloaded 332 times

using Michael script for the variation of the Twiss parameter with energy.

tuneshiftbytracking.tgz

(581.83 KiB) Downloaded 332 times

[michael_borland]

Laurent,

I suggest two changes for the tracking runs: first, use the sixth-order integrator by setting INTEGRATION_ORDER=6 on the KOCT, KSEXT, KQUAD, and CSBEND elements. Second, increase the number of slices by about a factor of 2.

--Michael

[nadolski]

Dear Michael,

Thank you for the suggestion.
I already did some studies playing with the number of slices and the integrator order.
There is no major effect. The converge is slightly improved but at the margin.
Frequency map are very similar.

I post for you information the tune shift by tracking with N_SLICES=10 and INTEGRATION_ORDER=6.

Question: the tracking is done around the closed orbit.
!track particles whether to center the beam transverse coordinates on the close orbit !before tracking
&track center_on_orbit=1 &end
Are you sure that the option is working properly?

Laurent.

[nadolski]

Finally I have got the point, looking directly at the phase space.
The RF cavity was defined twice.
The tracking was performed in 6D with a cavity instead of 4D.

[michael_borland]

Laurent,

Yes, that would do it. Sorry I missed it.

--Michael