Selecting between relative and absolute time tracking

[DavidSagan]

The switch also needs to select between using absolute time as opposed to the absolute time relative to the first turn reference time (see the Bmad documentation on bmad_com%absolute_time_ref_shift).

[DavidSagan]

Here is how the three time modes can be implemented given a selector switch. In the followinging, t_phase is the time used in the phase calculation:

  t_rf = phi0/omega + t_phase - z/(c*beta)
  phase = omega * t_rf

Let t_ref1 be the reference time computed for a given RF element on the first turn and t_refN be the (accumulated) reference time computed for the element at turn N. At turn N, t_phase will be:

  t_phase = 0                   # Relative time
  t_phase = t_refN - t_ref1     # Absolute time  with ref shift (bmad_com%absolute_time_ref_shift = T which is the default)
  t_phase = t_refN             # Absolute time, no ref shift (bmad_com%absolute_time_ref_shift = F)

[mattsignorelli]

I have some ideas... but first a question: Is there ever a case where you might want relative time tracking in some elements and absolute in others?

[DavidSagan]

With Bmad, no one has ever wanted mixed relative and absolute time tracking.

[mattsignorelli]

Correct me if I'm wrong, but wouldn't mixing the two be super useful for studies like resonant depolarization? The RF cavities all around the ring would stay relative time tracking, but the AC kicker would be absolute

[mattsignorelli]

I know having to set the phases for the rf cavities around the ring in FCC RDP studies has been a major pain point in Bmad

[DavidSagan]

Interesting. We should discuss.

[mattsignorelli]

I don't think handling it would be too hard. we can just have a flag in the RFParams, and then in the Time struct one could specify Time(absolute=false) for example. This would be super easy to implement. Mixing the two inside a single time function, e.g.

Time(absolute=false) + Time()

Is not as trivial, but i think we can find a solution. This also I imagine would be rare

[DavidSagan]

Thought: To get different types of time tracking for different elements, have a master switch that sets the default and have, for each RF element, a local switch that can override the default.

[eiad-hamwi]

Two possible routes:

1. Some Beamlines parameter group could contain absolute_time_tracking::Bool
2. The tracking method argument of track! could contain absolute_time_tracking::Bool