How Thermostats and Controls Fail (Generic)
Why this matters
A control tells the equipment when to run. When it fails, the symptom looks exactly like a failure of whatever it controls: the heat will not come on, the pump will not start, the cycle will not advance. Techs replace the expensive thing the control commands, then watch the same symptom return because the cheap brain was the real fault. Learning how controls fail, and how to prove the control rather than the load, saves you from chasing the wrong part on a huge share of no-run calls.
What a control is
Lump three things under "control" because they fail in similar ways and you test them the same way: the sensing element (a thermostat, a switch, a sensor that reads a condition), the logic (a control board or sequencer that decides), and the output (the relay or contact that actually switches power to the load). A call to run flows sense, then decide, then switch, then power the load. A break anywhere in that chain stops the equipment, and your job is to find which link is open.
How sensing elements fail
- Stuck or drifted reading. A temperature, pressure, or level sensor that reads a value that no longer matches reality. It may always read too high or too low, so the equipment cycles at the wrong point or never starts. Confirm by comparing the sensor's reading to a measured actual condition.
- Open or shorted. The sensing element fails to a dead open (no signal) or a dead short. Resistance against spec, with the part isolated, condemns it.
- Contacts pitted or stuck. A mechanical switch (a limit, a float, a pressure switch) burns or corrodes its contacts so it no longer makes or breaks cleanly. It may chatter, stick closed, or fail to close.
- Calibration loss. Older mechanical sensing drifts over time and starts triggering early or late.
How logic fails
- No output despite good inputs. The board receives the call and the safeties are satisfied, but it never energizes the output. The brain is dead or a section of it is.
- Burned or swollen components on the board. A scorched relay, a bulged capacitor, a cracked solder joint. Visual and smell inspection catches a lot of board failures before you ever meter them.
- Lockout that will not clear. Many boards lock out after repeated failed attempts and need a reset, then re-lock if the underlying fault persists. A board in lockout is often reporting a real downstream problem, not failing itself.
- Intermittent from heat or vibration. A board that works cold and quits hot is the heat-or-cool test's home turf.
How outputs fail
The relay or contactor that switches power is a wear part and a frequent failure point.
- Contacts welded closed. The load runs and will not stop, because the switch fused shut under arcing.
- Contacts burned open. The load will not start because the contacts no longer make. Pitted, blackened, high-resistance.
- Coil failure. The relay's coil opens or shorts so the contacts never move. Read the coil against spec, isolated.
Proving the control, not the load
The whole game is to test the chain in order and find the open link before you condemn the load.
- Confirm the call is present. Is the control actually being told to run? A bad thermostat or a satisfied setpoint means nothing should run, and the equipment is innocent.
- Confirm the safeties are satisfied. Limits, pressure switches, and interlocks open the chain on purpose. An open safety stops everything and is usually pointing at a real problem (overheat, low pressure) you must address, not bypass.
- Confirm the logic produces an output. Does the board energize its output terminal when it gets the call and the safeties are closed? Back-probe the output while it should be firing.
- Confirm the output reaches the load. Does switched power actually arrive at the load? If the board fires but power does not reach the load, the relay or wiring between them is open.
- Only then test the load itself. If the load gets correct power and still does nothing, now the load is the suspect.
The bypass trap
It is tempting to jumper across a control to "see if the load works." Done carelessly, you can run a load past a safety that opened for a real reason, and damage equipment or hurt someone. Jumper only across a control you are deliberately testing, only briefly, and never across a safety device whose job is to stop a dangerous condition. A safety that keeps opening is doing its job; find out why.
References
- Manufacturer service documentation for control sequence of operation and component test values
- Trade-standard control diagnosis (sense-decide-switch chain testing)
- See related: Confirming a Part Is Actually Bad; How Connectors and Terminals Fail