Add z coordinate to stat file

Proposal from @winklehner_d and @ext-bershanska_a:

Add the z coordinate of the reference particle to the stat file (seems like only s is given, which probably stems from the mismatch between OPAL-T and OPAL-cycl coordinate systems).

assigned to @snuverink_j

discussed with @gsell to add this also to 2.0. I will try to do this week.

I had a look at the code, and the stat file has the columns 18-23 reserved for the reference particle (see also Manual section 1.7.1 of the OPAL-T section However these are not filled for the ParallelCyclotronTracker. I can add this.

However, there are also the columns 43-48 for the R0 particle, isn't this also the reference particle? And although it mentions s at the header, it seems to me that z is actually saved ( beam->R[0](2) ).

@winklehner_d : so would column 45 ("R0 Particle position in s") be the right one for you? Or am I missing something?

As far as I understood, R0 is supposed to be the reference particle, yes. Filling the other columns (18-23) would be a nice touch, though. If nothing else then for legacy purposes maybe?

I wasn't aware that the cyclotron tracker just dumped z in the s column, but imo this is a bad idea. Accelerator people use s as the path along the design trajectory. This shouldn't double as the z coordinate (though they might be similar or identical in many systems). There should be a clear distinction. s is a continuously increasing parameter, z can oscillate. If anything, we should calculate s correctly in the cyclotron tracker, save it as s and save z in a new column. We could use 18-23 for the reference particle in lab coordinates and 43-48 for the reference particle in a local coordinate system? That would mean x=y=0 and px=py=0 basically and ps=total momentum, s = path along the design trajectory. @adelmann, @frey_m what do you think?

Thanks. For the cyclotron tracker the path length of the reference particle is actually in column 2 ("Average Longitudinal Position").

I think from looking at the code the TTracker also saves z in column 45. So shall I simply change the header?

Btw, note that columns 43-48 are only filled when OPAL is run on a single node since on multi nodes it is not clear anymore which particle is the reference particle.

In issue #227 the task "Reference particle(s) class, for both flavours" is mentioned. Not sure what is exactly meant, but it seems to me that if one wants reference particle information it is cleanest to track a separate reference bunch like is done in the TTracker.

@snuverink_j It is planned to do a similar implementation of the reference particle for the CyclotronTracker as done in the TTracker.

Should then R0 be eliminated? i.e. in single node (when only R0 is dumped) the same values should be dumped in columns 18-23 (appropriately labeled as z), and with multiple nodes proper reference particle is dumped in the right columns. At least, stat output file should be explained in documentation to eliminate the confusion about multiple reference particles.

One other thought: in OPAL-cyc we have also the xxx-trackOrbit.dat file. Here 2 reference particles in global (floor) coordinates are dumped. This information is available in single and multi core mode. To me this is the information w.r.t. the global coordinates. Column 2 is s.

I think we should add the xxx-trackOrbit.dat information to the stat if that solves the problem of @winklehner_d and @ext-bershanska_a

Discussed with @adelmann: we propose to add the reference particle information like it is calculated for the trackOrbit file to the columns 18-23. We'll keep the R0 columns for now for backwards compatibility, but these should probably be removed in the future.

mentioned in commit 8c97a04e

Implemented as suggested, I hope this is a workable solution for now (please reopen if not). This should be fixed properly when the reference particle implementation is done.

closed