Suggestion for RF devices

[JeanLucPons]

Hello,

Here is a proposal for pyAML RF management.
The main RFPlant device will provide global set_voltage() method that will either set up AT cavities or control system devices.

sr.live.get_rftransmitter("KLYS_TRA1").voltage.set(5e6)
sr.live.get_rf("RF").frequency.set(352372169.39)
sr.live.get_rf("RF").voltage.set(6.5e6)

An example showing a possible configuration for ESRF storage ring:

type: pyaml.rf.rf_plant
name: RF
masterclock:
      type: tango.pyaml.attribute
      attribute: sy/ms/1/Frequency
      unit: Hz
transmitters:
  - type: pyaml.rf.transmitter
    name: KLYS_TRA1
    cavities: [CAV_C05_01,CAV_C05_02,CAV_C05_03,CAV_C05_04,CAV_C05_05,CAV_C07_01,CAV_C07_02,CAV_C07_03,CAV_C07_04,CAV_C07_05]
    harmonic: 1
    distribution: 0.7692 # Default nbCavity/totalNbCavity
    voltage:
      type: tango.pyaml.attribute
      attribute: srrf/anodloop/tra1/Voltage
      unit: V
    phase:
      type: tango.pyaml.attribute
      attribute: srrf/phaseloop/tra1/Phase
      unit: deg
  - type: pyaml.rf.transmitter
    name: SSA_C25_1
    cavities: [CAV_C25_01]
    harmonic: 1
    distribution: 0.07692
    voltage:
      type: tango.pyaml.attribute
      attribute: srrf/ssa-driver/c25-1/RFAmplitude
      unit: V
    phase:
      type: tango.pyaml.attribute
      attribute: srrf/ssa-driver/c25-1/Phase
      unit: deg
  - type: pyaml.rf.transmitter
    name: SSA_C25_2
    cavities: [CAV_C25_02]
    harmonic: 1
    distribution: 0.07692
    voltage:
      type: tango.pyaml.attribute
      attribute: srrf/ssa-driver/c25-2/RFAmplitude
      unit: V
    phase:
      type: tango.pyaml.attribute
      attribute: srrf/ssa-driver/c25-2/Phase
      unit: deg
  - type: pyaml.rf.transmitter
    name: SSA_C25_3
    cavities: [CAV_C25_03]
    harmonic: 1
    distribution: 0.07692
    voltage:
      type: tango.pyaml.attribute
      attribute: srrf/ssa-driver/c25-3/RFAmplitude
      unit: V
    phase:
      type: tango.pyaml.attribute
      attribute: srrf/ssa-driver/c25-3/Phase
      unit: deg

[GamelinAl]

Could you explain what is the distribution parameter?
The name transmitter is strange to me, it is meant to refer to the power amplifiers that drive the rf right?
Other than this, it seems fine for a first proposal to implement. Later on, we will probably need to add some more complex design to deal with particular cases like passive harmonic cavities but that can wait.

[JeanLucPons]

distribution is just a factor that tell with ratio of the total voltage is generated by this amplifier. It can be changed to an array if a single transmitter power different cavities.
A RF transmitter is a system that power a set of RF cavities including driver/amplifier/phase sifter.

[TeresiaOlsson]

I also think transmitter is a strange name. I'm not sure that any of the physicists will understand it. Is it the same as the low level RF system?

I'm confused about the structure of the data models. Is it one model RF which inside it has other models for masterclock and a list of transmitters. And is there a Cavity device or not? Do you perhaps have the pydantic base models somewhere for it so I can see how they are nested together?

[JeanLucPons]

RF transmitter is a name we use at the ESRF but it can be changed.
Apparently this name is used in other places
Low level RF is rather the driver part.

class ConfigModel(ElementConfigModel):

    masterclock: DeviceAccess|None = None
    """Device to apply main RF frequency"""
    transmitters: list[RFTransmitter]|None = None
    """List of RF transmitters"""

class ConfigModel(ElementConfigModel):

    model_config = ConfigDict(arbitrary_types_allowed=True,extra="forbid")

    voltage: DeviceAccess|None = None
    """Device to apply cavity voltage"""
    phase: DeviceAccess|None = None
    """Device to apply cavity phase"""
    cavities: list[str]
    """List of cavity names connected to this transmitter"""
    harmonic: float = 1.0
    """Harmonic frequency ratio, 1.0 for main frequency"""
    distribution: float = 1.0
    """RF distribution (Part of the total RF voltage powered by this transmitter)"""

[gubaidulinvadim]

As I was mentioning in the pull-request, I think the Master Clock should be available as a separate pyAML device. Right now, I understand that this is a part of an RFplant. In some use cases, for example chromaticity measurement, only RF Master clock is necessary to do the measurement. Any other information is not necessary. Are there any reasons to have it only in the RFplant? Do I understand correctly that I can just create RFMasterClock (in the source code) device and attach the same RWFrequency scalars to it?

[JeanLucPons]

It has to be in the RF plant in order to set the RF frequencies on all powered cavities of the model.
For the control system, yes it is a standalone device. sr.live.get_rf("RF").frequency.set(352372169.39) just access the master clock.