W function start value is ignored for transfer lines in pyat

[oscarxblanco]

Dear all,
I want to see the effect of a transfer line on the chromatic amplitude and phase, however, when I try to set the start value of the W functions to zero it is ignored and calculated from the input twiss.

>>> tw_in_w_set_to_zero['W']
array([[0., 0.]])
>>> _,_,tw_out = my_transfer_line.linopt6(refpts=0,get_w=True,twiss_in=tw_in_w_set_to_zero[0])
>>> tw_out['W']
array([[12.91922806,  1.36202752]]) # I expected this to be zero

[swhite2401]

@oscarxblanco if I remember well W functions are derived values, dalpha, dbeta and ddispersion are the values in twiss_in used to derive it. So you should change these to modify the initial W function

[oscarxblanco]

@swhite2401 , it didn't work. I might be missing something else.

Moreover, even if I keep the initial values I see that there is a visible difference on the vertical W between pyat and MAD-X. Has this been bench marked ?

[swhite2401]

Ok I need to take a closer look then. I don't think this has been benchmarked for transfer lines, there could be something wrong in fact

[oscarxblanco]

@swhite2401 , it might be an issue with MAD-X not being able to reproduce the dynamics on the vertical plane.

For example, I know chromaticities are tipically different between pyat and MAD-X, but, pyat agrees well with the PTC libraries, and I think PTC is accurate. Unfortunately, I don't know how to calculate WX and WY in PTC but I could have a look.

[swhite2401]

@swhite2401 , it might be an issue with MAD-X not being able to reproduce the dynamics on the vertical plane.

For example, I know chromaticities are tipically different between pyat and MAD-X, but, pyat agrees well with the PTC libraries, and I think PTC is accurate. Unfortunately, I don't know how to calculate WX and WY in PTC but I could have a look.

Right this is possible... to be checked