RIKENFE: Provide calibrated Magnet temperature PVs

[davidkeymer]

As a support technician, I would like the RIKEN magnet temperature PVs to be calibrated values, so that they can be interpreted more easily. Currently, the values are raw voltages and have to be converted after extracting them from log files and/or the instrument database.

The provided calibration calculation steps are as follows (edited for clarity):

• The tap value is a signed 12-bit integer measured by the PLC ADC which has a 10V reference. This needs to be converted back to a voltage: V_adc = (PV^* / ((2¹²)-1))*10
• The PLC measures the signal conditioned voltage and so needs to be divided by a gain to get the actual voltage at the magnet terminals (gains are: RQ1 = 2, RB1 = 1.25, RQ2 = 5): V_tap = V_adc / gain
• From the magnet voltage, the equivalent resistance can be computed by dividing the tap voltage by magnet current: R_tap = RT2 = V_tap / I_magnet
• Then from this resistance, the temperature can be derived: RT2 = R23(1+0.004041(T2-23)) where RT2 is the resistance calculated and R23 is the initial resistance value, seen in linked excel sheet.

^* - PV currently provided by SCHNEIDER IOC (raw tap voltage from PLC)

Acceptance Criterion

• PVs provide actual temperature values.

Notes

• Suggest retaining the existing PVs for the raw voltages, and creating corresponding PVs for the calibrated values
• Raw voltage values are currently read by the main RIKEN PLC and in turn by the SCHNEIDER IOC running on the RIKEN control machine.
• The calculations could be added to the record loaded by the SCHNEIDER IOC, but it may be preferable to create another IOC, or use INSTETC for load-balancing and to prevent coupling.
• Historic data needs to be kept for at least a cycle so that potential trends in magnet temperatures can be identified. (may need to be covered by another ticket)

[zsoltkebel]

PRs

• https://github.com/ISISComputingGroup/EPICS/pull/339
• Ticket7975 add ioc rknmntr EPICS-Rknmntr#1
• Ticket7975 add ioc rknmntr EPICS-ioc#815
• Ticket7975 add ioc rknmntr ibex_gui#1614
• Record host name separately in channel access EPICS-IOC_Test_Framework#613 (NEW)
• Ticket7975 #8038

To test

• Change $(P) to IN:RIKENFE: in the following lines:
  • https://github.com/ISISComputingGroup/EPICS-Rknmntr/blob/282bcd7006ef54ee3e0ab27bea0d3d7455cbd28f/rknmntrSup/RIKEN_TEMP_CALC.template#L7
  • https://github.com/ISISComputingGroup/EPICS-Rknmntr/blob/282bcd7006ef54ee3e0ab27bea0d3d7455cbd28f/rknmntrSup/RIKEN_TEMP_CALC.template#L43

Notes

• There is an upcoming meeting to discuss data persistence. For now archiving is enough.
• No simulation records because there is no device to simulate.
• Notes added to wiki here

[esmith1729]

Was seeing some strange behaviour when magnets are turned off, NaNs and +/- Infinity for MOhms:

[esmith1729]

It would also be good to write some system tests for this, even if a bit trivial.

[zsoltkebel]

I now added some simple tests because it is not possible (I think) to simulate calc records. I also added edge case handling in the calculations to guard against division by 0.

[zsoltkebel]

New system tests have been added. Changes are made in support, ioc, and testframework. Those are important to pull before testing.