Trace the proposed PID retrofit through the Casa IV’s original controls, heater safety chain and the exact ProS3 header pins. Select a mode or click any component for its role, terminals and verification gate.
Design architecture · survey not complete
Interactive system architecture
Active Neutral Protective earth Selected heater path 5 V SELV Logic / sensor Solid = OEM/retained Dashed = added
Build gates · test before advancing
Eight controlled phases
Each phase ends with recorded evidence and a stop/go gate. Phases 0–3 are de-energised or SELV-only. The stock machine is not altered electrically until its sensor baseline and every low-voltage failure path have passed.
You already have the first tool
Digital multimeter: useful for unplugged continuity/resistance and live SELV measurements. It is not enough to prove boot-time glitches, appliance insulation or protective-earth integrity, and this plan does not ask you to probe inside an energised machine.
Buy or borrow for the bench
Current-limited 0–6 V bench supply, rated for at least 1 A; a known-good USB supply or power bank; insulated clip leads; a reference thermometer; and optionally precision PT100 substitution resistors. Set the ProS3 rail to 5.0 V and never above 5.2 V.
Borrow rather than improvise
Two-channel oscilloscope for the heartbeat/SSR-input proof and a contact thermocouple logger for the heatsink. Use a normal grounded scope only on the SELV bench while it is physically disconnected from mains. The electrician supplies appliance-safety and mains test instruments.
0Survey and quarantineNo power · establish facts before design freeze
Build
Photograph and label the untouched loom; trace brew/steam topology; record nameplate, thermostat markings, heater resistance, terminal sizes, enclosure space and candidate non-pressure-boundary probe location. Inspect every ordered part and quarantine anything inconsistent with its datasheet.
Tools
Multimeter, calipers, camera, labels and notebook. Resistance/continuity only after unplugging, proving dead and isolating the measured part where parallel paths would corrupt the reading.
Verify
Calculate heater power from measured cold resistance; confirm all three OEM thermostats and protective earth remain; produce an as-found schematic with wire IDs and photographs.
Gate
Do not buy final mounting hardware or design the mains PCB until the trace and dimensions are complete and independently reviewed.
1PT100 and MAX31865 benchSELV only · make temperature measurement trustworthy
Build
Wire the ProS3, MAX31865 and PT100 only. Use LDO1 3V3, the documented SPI pins, the measured RTD wire count and the matching board jumpers. Configure the 50 Hz rejection filter.
Tools
Approved USB supply/power bank or 5.0 V current-limited bench supply, multimeter, ice bath, boiling-water bath, reference thermometer and insulated clips.
Verify
Measure the probe resistance before connection; compare ambient, stirred ice bath and boiling-water readings with the reference; unplug and short the RTD leads to confirm explicit faults. Record reference uncertainty and target agreement within 1 °C rather than silently applying an offset.
Gate
Stable readings, correct fault reporting and a confirmed surface-mount retention plan. Reject the generic probe if its wiring, insulation or temperature rating cannot be proven.
2Firmware with a fake heaterSELV only · LED/resistor replaces the SSR
Build
Implement the local state machine, 3–5 s slow-PWM window, mode-specific limits, NVS validation, RGB annunciation and an output interlock. Represent heat demand with an LED and resistor; simulate IO14 steam state with a SELV jumper or switch.
Tools
USB or bench supply, multimeter, LED/resistor, jumper switch and optional precision resistors or a firmware sensor simulator.
Verify
Power-on defaults to heat disabled; Wi-Fi/API loss cannot force heat; open/short/stale/plausibility, no-rise, excessive-rise and hard-max tests all remove demand and latch the red fault pattern; steam state suppresses PID demand without a brew over-temperature latch.
Gate
A written test table passes after cold boot, reset and repeated network loss. No SSR is connected yet.
3Hardware fail-off and mode benchSELV output side · scope the states a meter can miss
Build
Add the independent IO15 heartbeat permit, IO16 demand gate, BC337/4.7 kΩ driver and only the G3NB input terminals 3+/4−. Bench the isolated steam detector's SELV output; its mains-facing input design and test source require qualified review.
Tools
5.0 V current-limited bench supply, multimeter and preferably a two-channel oscilloscope. A logic analyser can confirm digital timing but cannot replace the analogue SSR-input trace.
Verify
Capture power-up, reset, firmware flash, IO15 stuck high/low, heartbeat disconnect and IO16 stuck high. SSR input must remain below 1 V when off, exceed 4 V only with valid heartbeat plus demand, and lose drive within the commissioned timeout.
Gate
Saved scope captures prove there is no trigger pulse and every single fault fails off. Do not connect the SSR output to AC on the hobby bench.
4Stock-machine thermal baselineSensor only · heater wiring remains factory stock
Build
Mechanically retain the PT100 and its independent strain relief, then reassemble the machine. Run the logger from a power bank in a protected SELV enclosure; use Wi-Fi for logs and keep ordinary USB/laptop leads disconnected while the machine is on mains.
Tools
Power bank, reference/contact thermometer, Wi-Fi logging and camera. No exposed internal parts during powered operation.
Verify
Record at least three brew-stat cycles and three steam cycles: shell temperature at thermostat open/close, warm-up time, normal dT/dt and temperature rise under sustained heat. Inspect the cold probe afterward for any movement or lead damage.
Gate
Repeatable logs and a probe that cannot detach. Derive the initial shell setpoint and provisional no-rise/rate-of-rise envelopes from these measurements.
5Mains integrationLicensed electrician · no hobby live probing
Build
The electrician installs the fused IRM-05 PCB, barriers, steam detector, SSR/heatsink, high-temperature wiring, enclosure and service provision while retaining protective earth and all OEM thermostats.
Tools
Electrician-selected CAT-rated instruments, earth-bond/continuity tester, insulation-resistance or appliance-safety tester, leakage test equipment and calibrated termination tools. Your general-purpose multimeter is not the acceptance instrument.
Verify
Visual/clearance review, protective-earth continuity, polarity, insulation resistance using a method safe for connected electronics, fuse coordination, strain relief and leakage/appliance tests. Powered mode-sense checks are performed only under the electrician's controlled procedure.
Gate
Recorded appliance-safety results and covers/barriers fitted before normal mains energisation. An RCD does not replace these tests.
6Controlled commissioningFilled, closed and continuously attended
Build
Prime the boiler, reassemble every cover and start with heat demand disabled. Verify stock/bypass control if provided, then enable the lowest conservative shell setpoint. Keep Home Assistant advisory only.
Tools
Contact thermocouples on SSR/heatsink/enclosure, logger, timer and electrician-operated current measurement if required. No grounded oscilloscope or ordinary USB cable is attached to the energised appliance.
Verify
Confirm brew regulation, steam detection with PID demand forced off, 104/135/165 °C hardware devices still in their intended paths, sensor-disconnected-at-boot fail-off, heartbeat reset fail-off and local fault acknowledgement. Measure terminal, SSR and sink temperatures through repeated brew/steam duty.
Gate
No unexpected heat, no loose or hot termination, no thermal limit approached and all installed-system fault tests pass. Stop immediately on smell, noise, leakage, unstable temperature or unexpected current.
7PID tune, soak and handoverOnly after safety behaviour is already fixed
Build
Tune around the measured OEM-equivalent shell temperature with the 3–5 s output window. Calibrate against cup/group behaviour, then lock commissioned setpoint limits, thresholds and the offline fallback.
Tools
Temperature logger, fast cup-water thermometer or thermocouple, scale/timer and contact thermocouple on the SSR sink. A thermal camera is optional; shiny metal emissivity makes it secondary to a contact probe.
Verify
Run at least five brew cycles, three steam recoveries and a one-hour attended soak. Record overshoot, recovery, brew-water repeatability, SSR temperature and any operation of the retained 104 °C thermostat. Repeat fault checks after the final firmware image.
Gate
Configuration, schematic, photographs, test results and rollback instructions are archived. Normal PID operation should not rely on the 104 °C thermostat cycling.
Exact header reference
ProS3 pin allocation
The labels and physical header order below are transcribed from Unexpected Maker’s ProS3 pinout card. Gold rows are assigned by this project; red-edged IO0 and IO3 are boot-strapping pins and deliberately unused.
Visual parts index
Component atlas
Purpose-drawn vector illustrations preserve the important physical distinctions and terminal labels without pretending an AliExpress listing photo proves the delivered part’s identity. Click a card for its acceptance checks.
Source boundary
Component identity and physical positions come from locally held ECM Casa IV research copies, which are intentionally not republished here. The ProS3 header order and proposed project assignments are checked against Unexpected Maker’s official card; the mains topology remains survey-dependent.