Channel Map¶

The canonical source is controlunit/settings.yml — the developer's deliberate in-code data dictionary, written as a memory aid after long gaps in work. Channel names, ADC addresses, gains, descriptions, and conversion function identifiers all live there.

"I did that to not remember the modules. So I DID DOCUMENT THERE INTENTIONALLY." — Arseniy

Hardware documentation (board layouts, signal conditioning, pin numberings, datasheets) lives in the external hub: aklab-howto. Do not expect hardware build notes here.

ADC signal channels¶

Name	Signal	Sensor / notes
`Ip`	Plasma current	Hall-effect sensor on channel 0, on the preanode supply's wire (owner statement 2026-10-07: "plasma current IS preanode current presently"): `25 * (Ip / Vhall - 0.4961) A`; supply-corrected since 4.22.0, with the existing provisional 5 A/V sensitivity at 5 V. Against that supply's panel it reads about 0.70 (0.60 A for 0.9 A on 2026-10-02, 1.54 for 2.2 A on 2026-10-05, a flat 3.53 A where the supply limited at 5 A on 2026-10-06), so the true sensitivity is nearer 7 A/V until the known-current calibration says the number. The resting ratio was the ideal 0.5 until 4.22.2 (files up to 2026-10-02 read 0.098 A low); each file's header carries the formula it was written with.
`Vhall`	Hall supply reference	Channel 1, read immediately after `Ip`, 10 V range (the 5 V range clips at 5.0176 V). Raw volts retained for diagnosis; no additional display or curve.
`Pu`	Upstream pressure	Pfeiffer single gauge PKR251 (currently operating in Pirani mode — cold-cathode discharge not igniting)
`Pd`	Downstream pressure	Ionization gauge, with `Place: "downstream"` in the settings for the Control dock and the web Gauges card to show beside its short name. Its controller's mode (0 Torr linear, 1 Pa log, or 2 off — the gauge switched off at its controller and declared so on the Control dock or the web Ion gauges card, when the raw volts are kept and `Pd_c` is NaN) and exponent are recorded per row as `IGmode` and `IGscale`, the columns the file has always had.
`Pu2`	Upstream pressure	Second ionization gauge, wired to channel 16 on 2026-09-15, with its own selector on the Control dock and the web Control tab, and `Place: "upstream"` in the settings for both to show beside its short name. Its mode (the same three codes as `Pd`'s) and exponent are recorded per row as `IGmode_Pu2` and `IGscale_Pu2`, appended after the original columns so nothing an old reader counts on moves.
`Bu`	Upstream Baratron	MKS 627, FS = 1 Torr
`Bd`	Downstream Baratron	MKS 628B, FS = 0.1 Torr
`MFC1`	H₂ flow	20 SCCM range
`MFC2`	O₂ flow	10 SCCM range
`Ci`	Cathode current	Channel prepared in ADC map and on PCB; never actually measured. Conversion function exists for consistency.
`Cv`	Cathode voltage	Same as `Ci` — prepared but not deployed
`T`	Membrane temperature	Now read off-Pi (NI on Windows). Channel definition remains.
`QMS_signal`	Sync trigger state	Boolean — logs whether GPIO sync is active during each row

Hall supply correction (4.22.0)¶

The reference must measure the supply feeding the Hall sensor, against the same ADC ground. Queezz connected the PSU voltage on 2026-09-30. The prior board inspection identifies the neighbouring terminal as channel 1 beside Ip on channel 0; verify those voltages on the rig before deployment.

Ip names Supply Channel, Nominal Supply, Zero Ratio, and Amperes Per Volt in settings. The values are 5 V, 0.4961 and 5 A/V; the ratio is where the sensor was measured to rest with no current on 2026-09-30 and 2026-10-02, the same within 4 mA across both days. This cancels proportional supply movement in both the zero and the sensitivity; it does not replace a known-current calibration or correct ground offsets. The recorded Ip_c, existing current readouts, and PID use the same correction. Raw Ip and Vhall remain volts, and the CSV header records the formula. Readers must use named CSV columns, as the new signal changes column positions. Slow sampling converts the paired means stored in the row.

A missing, non-finite or out-of-range supply (configured limits 4–6 V) makes Ip_c unavailable (NaN); the event is logged once and recovery is logged too. No stale reference or fixed-voltage fallback is substituted. The PID holds its existing command during that interval. Settings without a supply channel retain 5 * (Ip - 2.52), with optional Zero Voltage and Amperes Per Volt overrides. Settings version 1.5 activates the new canonical map; older local settings need their customizations merged before reuse. The rig had no local settings file on 2026-09-30.

The noise investigation explains why the supply was added. Filters and longer baseline windows are separate work; reassess noise with the newly wired reference first.

DAC outputs¶

Device	Signal	Notes
DAC8532 ch1	MFC1 (H₂) setpoint voltage
DAC8532 ch2	MFC2 (O₂) setpoint voltage
MCP4725	Plasma current (Ip) DAC	Behind galvanic I²C isolator since Apr 2026

Conversion functions¶

Conversion polynomials are defined in controlunit/devices/conversions.py with LaTeX in docstrings — physics inline with the code that uses it.

The polynomials come from external datasheets. The PKR251 datasheet, for example, lives in aklab-howto.

Dispatch is settings-driven: each channel in settings.yml carries a Conversion Function key that maps to a function in conversions.py via AdcChannelProps.set_conversion_function(). Adding a new instance of an existing sensor type means editing YAML, not code.

Settings versioning¶

The repo carries controlunit/settings.yml as canonical schema. The live rig runs ~/.controlunit/settings.yml with its own channel map. A Settings Version key prevents silent drift after a schema bump:

# controlunit/readsettings.py
if local_config["Settings Version"] == config["Settings Version"]:
    config = local_config