Timepix hardening + delay laser with calibration in BEC #47

2026-10-08T23:09:05+02:00

wyzula_j commented

2026-10-08 23:09:05 +02:00

TimePix and timepix_scan: reviewed and hardened + tests.
Laser delay (dev.laser_delay): implemented according to the DelayLaser.py. The original script is kept in the repo as a reference; the tests check that the device writes the same PV values.
Simulation: SimLaserDelay (with a simulated scope) for local development without hardware.
Calibration is kept in autosaved PVs (X10DA-LAS-DELAY:*), so it survives BEC restarts and session changes.
- Calibrate in the BEC client with the scope reading (X-ray minus laser, in ns): dev.laser_delay.calibrate()
- Show it with dev.laser_delay.get_calibration(), or get_calibration_info() for the date and scope value.
- The calibration values can't be set individually from BEC, only through the calibration command.
- They are written to every HDF5 file as configuration signals of laser_delay.
Safety:
- Phase shifter voltages are capped at 5 V and ramped slowly, as in the script.
- Only the device itself writes the delay card and the phase shifter.
- If the hardware no longer matches the stored calibration (laser restart, changes outside BEC, an interrupted move), moves are refused until the next calibrate.

Tested at SuperXAS with Grigory and Alvin; this branch is what the beamline runs now.

Open questions (from the original script logic, not blocking to proceed with the PR atm, but we can discuss them later): can the delay card's FINE value be 200 (BEC currently writes the next coarse unit instead)? Are the RF and cavity periods 2.0 / 12.01 ns as in the script? Is a phase shifter amplitude of 1.0 V right?

- TimePix and timepix_scan: reviewed and hardened + tests. - Laser delay (dev.laser_delay): implemented according to the DelayLaser.py. The original script is kept in the repo as a reference; the tests check that the device writes the same PV values. - Simulation: SimLaserDelay (with a simulated scope) for local development without hardware. - Calibration is kept in autosaved PVs (X10DA-LAS-DELAY:*), so it survives BEC restarts and session changes. - Calibrate in the BEC client with the scope reading (X-ray minus laser, in ns): dev.laser_delay.calibrate(<ns>) - Show it with dev.laser_delay.get_calibration(), or get_calibration_info() for the date and scope value. - The calibration values can't be set individually from BEC, only through the calibration command. - They are written to every HDF5 file as configuration signals of laser_delay. - Safety: - Phase shifter voltages are capped at 5 V and ramped slowly, as in the script. - Only the device itself writes the delay card and the phase shifter. - If the hardware no longer matches the stored calibration (laser restart, changes outside BEC, an interrupted move), moves are refused until the next calibrate. Tested at SuperXAS with Grigory and Alvin; this branch is what the beamline runs now. Open questions (from the original script logic, not blocking to proceed with the PR atm, but we can discuss them later): can the delay card's FINE value be 200 (BEC currently writes the next coarse unit instead)? Are the RF and cavity periods 2.0 / 12.01 ns as in the script? Is a phase shifter amplitude of 1.0 V right?

wyzula_j self-assigned this 2026-10-08 23:09:05 +02:00

wyzula_j added 2 commits 2026-10-08 23:09:05 +02:00

Introduction of new timepix scan 3b0017ca99

Added ROI parameters 200c2cdfdb

wyzula_j force-pushed feat/laser-delay-calibration-pvs from 5284f55e9d to dd9d26085f

2026-10-12 12:23:36 +02:00

Compare

wyzula_j added 3 commits 2026-10-13 16:00:29 +02:00

fix(timepix): harden timepix_scan and repeat acquisitions with lost frames 5bbddfa6aa

feat(devices): add LaserDelay positioner for the delay card and RF phase shifter 422fe7edfc

feat(devices): keep the laser delay calibration in soft records on an IOC

CI for superxas_bec / test (push) Successful in 2m3s