EPR spectroscopy endstation

EPR spectroscopy endstation is a multi-functional setup located at the Novosibirsk Free Electron Laser Facility. The endstation consists of 3 EPR machines, namely (i) a pulsed X-band EPR; (ii) a continuous wave X-band EPR; (iii) a time-resolved X-band EPR equipped with a Lotis TII Nd:YAG laser. To be continued...

Hardware

Pulsed machine

Component	Model
Pulse generator / ADC / DAC	Insys FM214x3GDA (312.5 MHz TTL, 2.5 GHz ADC, 1.5 GHz DAC)
Magnetic field controller	BH15
Microwave bridge	Micran X-band v2
Temperature controller	Lakeshore 335

CW EPR machine

Component	Model
Lock-in amplifier	Stanford Research SR-850
Frequency counter	Agilent 53131A
Temperature controller	Lakeshore 335
Magnetic field controller	BH15

TR EPR machine

Component	Model
Oscilloscope (primary)	Keysight DSOX3034A
Oscilloscope (secondary)	Keysight DSOX2012A
Frequency counter	Agilent 53131A
Temperature controller	Lakeshore 335
Magnetic field controller	BH15

Other instruments

Component	Model
NMR gaussmeter	Sibir 1

Automated tune-up and measurement

The pulsed machine can be driven unattended by epr_auto, a YAML protocol runner that ships with the endstation's Atomize_ITC build: it tunes the spectrometer (vane power, working field, integration window, phase, pulse calibration, repetition rate), runs T2 / T1 measurements with SNR- and time-budget-driven scan counts, judges every result and records a full run manifest. See the epr_auto overview for the manual.

Saving the full 2D data

The two tools that acquire a full 2D array can write it as a single HDF5 file instead of comma separated text. The 1D result of an experiment stays CSV in both of them; only the 2D arrays change format, and reading either format stays supported everywhere.

In the AWG phasing tool the Settings tab carries the pair of checkboxes. "Save 2D" saves the full 2D array next to the integrated 1D result, as before; "Save 2D as HDF5" makes every 2D dump of the run an .h5 file — the _2d companion, the primary dump when the phase correction is off (so the 2D array is the result itself), and the per-cycle files of an ESEEM tau average, which stay one file per cycle. The 1D result files are unaffected.

In the TR EPR tool a single "Save as HDF5" checkbox picks the format for the whole run: the save dialog then offers .h5, and the derived _osc2 and _pulse files follow the name you choose. With "Save Each Scan" on, an HDF5 run no longer writes a whole new _{j}_scans file per scan — the cumulative average after each scan is appended as one slice of a scans dataset inside the same file, which is both far faster inside the acquisition loop and far smaller on disk.

The main window's "Open 1D Data" / "Open 2D Data" / "Open TR Data" actions and both Data Treatment windows read .h5 as well as .csv. A 2D file holding both quadratures opens as I and Q together, and the axes stored in the file are used to set the plot scales.

Per-tool working directories

The endstation runs as several independent Qt processes — the main window plus each control-center tool in its own QProcess. Every open/save dialog uses the last-opened-directory helper so it reopens where you left off, remembered independently per tool and across relaunches.

On top of the common script and data categories, the endstation build adds one category per preset tool, so each remembers its own folder independently:

Key	Tool	File type
cw	CW EPR control	*.cw
tr	TR EPR control	*.tr
tune	resonator tune preset	*.tn
phase	RECT phasing	*.phase
phase_awg	AWG phasing	*.phase_awg
phase_cor	phase correction	*.csv

The data category here also covers the TR data-open dialogs in addition to the 1D / 2D ones.

On the endstation build these <key>_lastdir.txt notes are stored under the repo's libs/ runtime directory, next to the other runtime-IPC files (these files are git-ignored), rather than under the per-user config directory used by plain Atomize.