Fitting dispersion data in linacs

[Max]

Hi Michael,
I have dispersion measurements (difference orbits at BPMs) in a linac. I would like to fit the eta_x and etap_x at the
start of the line (that contains earth field and correctors) in a way that gives the measured dispersion values at the BPM. One way would be:
-Fit the measured trajectory with dp/p=0 at the start of the line, using MALIN as the optimization parameter. Repeit with dp/p=1%.
The difference between the two fitted trajectories at the start of the line should give me the initial dispersion and dispersion prime values I am seeking.
If am correct, I ask: is there a more straightforward way ?
One more question: am I correct in assuming that the twiss_ouput instruction will give me the lattice functions with the
dipolare perturbations (earth field and correctors) included provided I set "output_at_each_step = 1" ?

Thanks, Max

[michael_borland]

Max,

The best way to do this is using fitting of the R16 matrix elements. This ensures that dispersion is being computed in terms of the initial energy deviation (note that &twiss_output gives the "local" dispersion, which isn't what you want). You can also use an EMATRIX element to specify and vary the initial dispersion.

I've attached a file giving an example of how to do this.

--Michael