Control Boards & Electronic Controllers

A board is four things, not one

Every control board consists of a power supply section that derives low-voltage rails from line power, an input section that reads switches and sensors, a logic section that decides, and an output section with relays or triacs that drive loads. Diagnosing 'the board' as a single object is what makes board diagnosis feel like guesswork. Diagnosing the four sections is a normal, tractable job.

Start at the power supply every time. A board with a sagging 5 V or 12 V rail behaves erratically in ways that mimic every other fault in the machine: random resets, dropped outputs, phantom key presses, and fault codes that make no sense together.

• Rails are typically 5 V for logic, 12 V for relay coils, sometimes 3.3 V on newer designs
• Erratic multi-symptom behaviour points at supply, ground, or connections before logic
• Visible damage — bulged capacitors, burnt traces, relay discolouration — is real evidence; note it

Inputs: what the board can see

Boards read switches by pulling a line to a reference and watching whether an external contact changes it. That is why a door switch fault, a corroded ground, or a shared-return harness problem can make the machine behave as if the user did something they did not. When a machine will not start, prove the board actually sees the interlocks: check the input voltage change at the connector as you actuate each switch.

Service modes matter here. Most manufacturers provide a diagnostic mode that displays live input states — that single feature converts an hour of measuring into two minutes of reading, and it also shows you what the board believes, which is the fact you need.

• Use the manufacturer's service mode to read input states before measuring individually
• A shared return or ground fault produces multiple unrelated input errors at once
• Input reads correct at the switch but not at the board = harness, not board

Outputs: what the board can command

Output diagnosis is a two-part question: did the board command the load, and did the load receive it. Measure at the board's output terminal with the machine calling for that load. Output voltage present and the load dead means the fault is downstream — harness, connector, or the load itself. No output voltage with all input conditions satisfied is where a board becomes a genuine suspect.

Watch for triac-driven outputs: they can leak enough voltage when off to read as live on a high-impedance meter while delivering no usable current. Use a low-impedance (LoZ) meter setting or confirm with a clamp before you act on that reading.

• Output present, load dead → downstream fault
• No output with all permissions satisfied → board suspect
• Use LoZ or a current clamp to defeat phantom voltage on triac outputs

A board driver can fail from a shorted load. Always test the load before installing a new board, or you will kill the new one too.

Condemning a board honestly

Three conditions justify a board replacement: all inputs required for the function are proven present at the board, no output is produced at the board terminal, and the downstream load and harness have been tested good. Anything less is a guess with an invoice attached.

Two extra habits protect you. Photograph and label every connector before removal. And before fitting the new board, retest the load that the old board was driving — a shorted valve coil that killed one driver will kill the next one within a cycle.

Start at the rails, always

A control board is a power supply with a computer attached. Line voltage enters, a switch-mode supply produces the low-voltage rails the logic and the relays need, and everything the board does afterwards depends on those rails being clean. Typical rails are a logic supply around 3.3 V or 5 V, a sensor reference at 5 V, and a relay or valve supply at 12 V or 24 V — AC or DC depending on the platform.

Rail faults produce symptoms that look like anything but a power supply. A sagging logic rail causes resets mid-cycle that the customer describes as random. A weak relay rail lets the board command an output and the relay chatter or fail to pull in, which looks like a failed valve. A degraded sensor reference makes multiple sensors read wrong at once. Measuring three voltages takes two minutes and rules out an entire family of misdiagnoses.

Look at the supply section physically as well. Bulged or vented electrolytics, discoloured resistors around the switcher, and the smell of cooked flux are all findings. So is a board that is clean and cool: it means the fault is probably not in the board.

• Logic rail: steady within a few percent, no dips as loads engage
• Relay or valve rail: holds under load when the output is commanded
• Sensor reference: steady 5 V measured at the sensor connector, not just at the board

Primary-side capacitors in the board's supply can hold charge after disconnect. Treat the primary side as live until measured.

Input, output, or interpretation

Every board fault reduces to one of three things: the board received bad information, the board failed to act, or the board acted on good information in a way you did not expect. Deciding which one you are looking at is the whole job, and the board will usually tell you if you ask correctly.

For inputs, compare what the board reports in service mode against what you measure at the sensor or switch. Agreement means the input path is sound and the fault is downstream. Disagreement puts the fault between the sensor and the board's own reading — harness, connector, or input circuit.

For outputs, command the output through service mode and measure at the board terminal, then at the load. Voltage at the terminal and nothing at the load is a harness or connector fault. No voltage at the terminal with the output commanded, and a healthy rail behind it, is a genuine output stage or relay fault. Voltage present and the load still dead is a load fault, and the board is innocent.

For interpretation, read the sequence. A board that refuses to advance is usually waiting for a permissive it never received — a level signal, a door switch, a thermal cut-out. That is not a board fault; that is a board doing its job in a machine that has a problem elsewhere.

• Board reports what you measure = input path good
• Voltage at the board terminal, none at the load = harness or connector
• No output with a good rail and a commanded state = board output stage
• Refusing to advance = missing permissive, not a failed board

Communication lines and what they look like when they fail

Modern platforms split control across boards: a main control, a user interface, an inverter or motor control, sometimes a dedicated sealed-system or dispenser board. They talk over a serial link, a CAN bus, or a proprietary two- or three-wire bus. Communication faults are the most misdiagnosed category on the bench because the symptom appears at one end and the cause is usually in the middle.

The classic pattern: the interface goes blank or unresponsive while the machine is otherwise fine, or the machine runs but ignores the panel. Both are communication symptoms. Before replacing either board, verify that both have power and ground, that the bus wiring is continuous with no shorts between the bus conductors, and that the ground reference between the boards is solid. A degraded ground between boards is a genuine and frequent cause of bus faults, and no board replacement fixes it.

CAN and similar differential buses have termination requirements. On commercial controllers, a missing or duplicated terminator produces intermittent faults that follow load and temperature, not any single device.

• Both boards powered and grounded before suspecting either one
• Bus continuity end to end, and no short between bus conductors or to ground
• Shared ground integrity between boards — a common cause with no part number
• On commercial buses, confirm termination is correct and not duplicated

The diagnose-before-replace standard

Boards are expensive, frequently non-returnable, and often the wrong part. The professional standard is that you can name the specific evidence that condemns the board, and that evidence is a measurement, not an elimination. 'Everything else checked good' is not evidence; it is the absence of one.

Also ask what killed it. A board that failed from a shorted valve coil, a welded relay driven by a stalling motor, or a surge from a poor supply will fail again unless the cause goes with it. Fitting a board without addressing the cause converts a repair into a warranty problem.

• Rails measured and either good or explained
• The specific input or output proven faulty at the board terminal
• Load and harness cleared, so the fault cannot be downstream
• The cause of failure identified where the board shows damage

Document the measurements that condemned the board. It protects the customer, your employer, and you.

Failure modes and what confirms them

Test procedures

Manufacturer differences

Safety and professional boundaries

• Board primary sides carry line voltage and stored charge — measure before contact, every time.
• Component-level board repair is outside the scope of most field service agreements; replace at the assembly level unless your employer's policy says otherwise.
• Do not defeat or bypass a permissive input to make a machine run. If a board is waiting on a safety input, it is right to wait.

Expected readings and what they mean

Field scenarios

Knowledge check