Control Board Replacement and Configuration
Purpose
A control board almost never dies of old age. Something took it out: a motor winding leaking to its housing, a heater shorted to a chassis, a chafed harness, water tracking across a connector, or a load drawing past what its driver was built for. Install a second board into the same fault and you have bought two boards, and the second failure is the one the customer blames on the shop.
This procedure guarantees three things before a board goes in. The board is proved faulty with a measurement rather than by elimination. What killed it has been found and corrected, or explicitly recorded as not found. And the new board is configured for that exact model and revision, because a generic service board fitted and never configured runs the wrong cycles and reads as a second board failure.
Scope
Replacement of electronic control boards, user interface boards and inverter or drive boards on residential appliances: laundry, dishwashers, refrigeration and cooking equipment.
It does not cover diagnosis up to the point of indicting the board, which the diagnostic visit intake SOP owns and which is a precondition here, nor board-level component repair. Microwave and induction inverter boards are in scope for replacement only, and only after the high-voltage capacitor is discharged.
Roles and the handoff between them
| Role | What they own | What they hand off |
|---|---|---|
| Technician | The indicting measurement, the upstream-cause search, configuration, and the post-install verification | A board record naming the cause found or explicitly stating none was found |
| Parts desk | Board number for the exact model and revision, supersession history, and whether a paired interface board is required | Both numbers where the board is paired, so nobody arrives with half a set |
| Lead tech or owner | Any board installed with no upstream cause identified | A written note accepting that risk, and the warranty position that goes with it |
| Office after the visit | The configuration record and the customer's restored settings | The configured options onto the equipment file, so a future board goes in matching |
Before the board is touched
Prove the appliance dead before any board is exposed. Unplug, or open the breaker on a hardwired unit, then verify at the terminal block with a meter proved live on a known source before and after, per NFPA 70E-2021, 120.5, with work practices at 29 CFR 1910.333(b)(2) rather than 1910.147, which excludes electrical utilization work at its own (a)(1)(ii)(C).
A microwave high-voltage capacitor holds a lethal charge after unplugging. Discharge it through a resistor across its terminals, then confirm at zero, before a hand goes into that cabinet. Do this even if you are only removing a user interface board on the door.
Drive and inverter boards store energy too. Wait the manufacturer's stated bleed time after disconnection and confirm the bus is at zero before handling.
Static destroys a board silently. Leave it bagged until installation, bond yourself to the appliance chassis, and never set a board on carpet or a vehicle seat. A board killed by static works for a week and then does not.
The procedure
Prove the board with an input and output measurement, not by exclusion. Confirm the board has correct supply and its inputs are present, then confirm it fails to produce an output the sequence calls for. Acceptance: a written pair, correct input present and commanded output absent at the board terminal. Wrong looks like "everything else tested good, so it's the board", which is how a board gets sold for a broken harness. Stop rule: no output measurement, no board. This step is energized work: intact lead insulation, one hand, stance clear of the cabinet.
Find what killed it before ordering anything. With the appliance proved dead and each load disconnected from the board, measure every motor, heater, valve and pump for winding resistance against the tech sheet and for any path between winding and housing. Inspect the harness at every edge and clamp, and look for water tracking or corrosion at the board connectors. Acceptance: every load in band and effectively open to ground, harness intact, connectors dry. Wrong looks like a pump whose winding reads correct but shows a path to its housing. Stop rule: a fault found here is corrected first; the new board does not go in until it is.
Record the outcome of the search honestly, including a null result. Write either the cause found or the words "no upstream cause identified" with the list of what was tested. Acceptance: one of those two statements present. Wrong looks like a blank field, which reads later as though nobody looked. Stop rule: a null result requires the lead-tech acceptance named in the roles table, because the risk of a repeat failure is now the shop's to carry, not the customer's to discover.
Confirm the exact board, its revision, and any paired part. Match against the full model and serial, check the supersession chain, and confirm whether the control and interface boards must be replaced as a set or matched by revision. Acceptance: part numbers confirmed for this model, with the pairing question answered yes or no. Wrong looks like a board that physically fits and communicates with nothing. Stop rule: an unanswered pairing question means do not order; a half set is a second trip.
Photograph and label before anything is unplugged. Image every connector in place, then label any two that could be interchanged. Acceptance: a photo set that lets a stranger reassemble the harness. Wrong looks like relying on connector keying that turns out not to be unique on that board. Stop rule: interchangeable connectors get labelled before removal. Cut-resistant gloves behind console panels.
Configure the new board before or during installation, per its instructions. Set jumpers, dip switches, model-select entries, or load the personality or software the service instructions specify, and transfer any plug-in memory module the design uses. Acceptance: configuration set and confirmed by the method the instructions give, usually a displayed model or version code. Wrong looks like a generic service board installed unconfigured, which runs default cycles and mimics a fresh board failure. Stop rule: a board that cannot be configured on site is not left installed.
Energize once, deliberately, and verify each output. Reconnect, confirm ground and neutral are landed, restore power, clear stored faults, and run the tech sheet's service test, checking each output and re-reading each load's current. Acceptance: every output commanded and observed, load currents in band. Wrong looks like a board powered up with a load still disconnected, which posts a new fault and sends the tech chasing it. Stop rule: an output that does not respond, or a load current out of band, stops the job before the customer's cycle runs.
Restore the customer's settings and run a full cycle. Set temperatures, cycle defaults, water hardness, and any accessibility or sound options the customer had. On gas equipment the fuel configuration is not a preference: set it to match the appliance rating plate and the gas actually supplied, verify it by the method the service instructions give, and before any full cycle runs, leak-check every gas connection disturbed and confirm the burner lights and proves without rollover with a personal CO monitor in the room. A fuel configuration you cannot verify against the rating plate means the appliance does not get fired. Acceptance: a full cycle completed end to end with no fault posted and the customer's settings as they were. Wrong looks like a machine handed back on factory defaults, which the customer experiences as a second fault. Stop rule: a fault that re-posts during the full cycle means the board was not the whole story; return to step 2 rather than ordering a second board.
The record this produces
A board record on the ticket, and it is the field the next tech will most want.
- The indictment: the input measured present and the output measured absent, at which terminal, during which commanded step.
- The upstream search: every load tested with its reading, the harness result, connector condition, and either the cause found or the explicit null result with its lead-tech acceptance.
- Parts: board number installed, revision, superseded-from number, and whether a paired interface board went in with it.
- Configuration: every jumper, switch, model-select code or software version, exactly as entered. The field that matters most in two years, because the next board is configured the same way and nobody will remember.
- Verification: service test outputs confirmed, load currents re-read, full cycle completed, faults cleared.
- Settings restored: the customer's own options.
The configuration line and the cause line earn this record its place. Without configuration a future board goes in guessing, and without the cause line a repeat failure in six months is indistinguishable from a bad part.
One run of this procedure, filled in
Front-load washer, no drain, control posts a drain fault, board already suspected by a previous shop.
- Step 1: board has correct supply. Control commands drain and the drain output at the board terminal stays dead through the command window. Input present, output absent. Board indicted.
- Step 2: appliance proved dead, drain pump disconnected from the board. Pump winding resistance reads in band, but there is a measurable path between the winding and the pump housing where the tech sheet requires effectively open. Fail. Stop rule taken: the new board does not go in. Pump replaced. Harness intact end to end, board connectors dry.
- Step 3: cause found and recorded as a pump winding leaking to its housing, with the reading. No lead-tech acceptance needed, because the search was not null.
- Step 4: board confirmed against the full model and serial. Current number supersedes the original. Interface board not paired on this platform, confirmed with the parts desk rather than assumed.
- Step 5: nine connectors photographed in place, two similar plugs labelled.
- Step 6: service board arrived generic and required a model-select entry. Entered per the instructions, confirmed by the displayed code.
- Step 7: reconnected, ground and neutral landed, faults cleared, service test run. Drain output commands and the new pump runs. Load currents re-read in band.
- Step 8: customer's cycle defaults and temperature preferences restored, full cycle completed with no fault posted.
The previous shop had the diagnosis right at step 1 and would have installed the same board. What separated the two visits was step 2, which took about 10 minutes and found the pump that had killed the first board. That board was not defective in any useful sense; it was a casualty, and replacing it alone would have produced an identical failure on a schedule nobody could predict.
When the job cannot run as written
No upstream cause can be found. A surge or a one-time transient leaves no evidence. Take the lead-tech acceptance under step 3, tell the customer plainly the cause was not identified, and say what happens if it recurs. Invent nothing to fill the field.
The board is discontinued. A superseded chain that ends with no current part sends the job to the repair-versus-replace standard, not to a used board from an unknown source. Where a used board is the only option, say so in writing and attach no parts warranty to it.
The board must be programmed by the manufacturer's tool and the tech does not have it. Do not install it. An unconfigured board left in a customer's machine is worse than the original fault, because the machine now runs and runs wrongly.
The paired interface board is on backorder. Install neither. A mismatched pair communicates badly rather than not at all, producing intermittent faults nobody traces back to the pairing.
The customer wants the board swapped without the upstream search. Explain what a casualty board is and what the second one costs them. Where they insist, record the refusal, take the lead-tech acceptance, and state that the board carries no warranty against a repeat.
References
- 29 CFR 1910.333(b)(2) for electrical work practices; 29 CFR 1910.147(a)(1)(ii)(C) for the carve-out that sends utilization work there; NFPA 70E-2021, 120.5 for the proving sequence.
- Manufacturer service manuals, tech sheets and board instruction sheets, which carry the load resistance values, the configuration method, the service test sequence and the capacitor bleed times.
- See related: the diagnostic visit intake and repair authorization SOP, which produces the indictment this procedure acts on, and the repair-versus-replace recommendation standard.