Aircon start relay: the part that gets out of the way
A compressor that hums, heats and stops looks finished. The same behaviour comes from a small component upstream of it, and the two behave alike from outside. Which one it turns out to be decides the size of the bill.
By Team Snowflake | Updated 8 Aug 2026
Why a single-phase motor will not start itself
Three-phase power starts a motor on its own. The three supplies arrive out of step with each other, so the magnetic field they make travels around the motor. A rotor sitting inside a travelling field is pulled along with it, and nothing has to be added.
Single-phase supply has no second and third leg to work with. One alternating current builds a field that grows and collapses along a single line, without turning. A rotor standing still inside that field is pulled equally hard in both directions, so it stays where it is and hums.
The way out is to build the missing offset into the motor itself. A second winding sits at an angle to the main one, and the current fed to it is held slightly out of step with the main current. Those two together give the field somewhere to travel and give the rotor a direction to follow.
Copeland describes its single-phase compressor motors as carrying one main and one auxiliary winding, with a capacitor and relay assembly supplied so the machine reaches the starting torque a three-phase motor produces unaided. The pair of windings and what each does is set out under compressor windings. What sits on this page is the switch that governs the second one.
Where the offset comes from, and what governs it
The capacitor sets the size of the offset. It shifts the current in the start winding away from the current in the run winding, and a larger shift means more turning force available at standstill. A capacitor weakening hands over a smaller shift, and the machine has less to break away with.
The relay decides how long the offset lasts. Producing the push and ending it are two different jobs, carried by two different parts, and they fail in two different ways. Owners are usually told about the first one and almost never about the second.
What happens in the moment the compressor breaks away
Standstill is the hardest instant in the working life of a compressor motor. A rotor that is not turning generates nothing back against the supply, so current pours in with little to oppose it. Copeland puts that inrush well above the figure the same machine draws once running, and it holds until the rotor begins to move.
The start path exists to shorten that instant. More turning force at standstill means the rotor breaks away sooner, and the heavy current ends sooner with it. Everything the starting device does is aimed at that one moment and nothing beyond it.
Once the rotor is turning, the same path turns into a liability. The motor now pushes back against the supply, current settles to the running figure, and the second winding has stopped contributing anything the machine needs. Left connected, it only makes heat.
Three ways the handover gets made
The device managing that handover comes in three broad forms, and they differ in what they watch. A current relay watches the line. Its contacts sit open at rest, the inrush pulls them closed, the start winding joins the circuit, and the contacts fall open again as current drops away with the motor gathering speed.
A potential relay watches the start winding instead. A turning motor generates a voltage in that winding, and the size of it tracks how fast the machine is going. Copeland's own wording is that the relay opens only once the motor has started and is approaching normal running speed.
A PTC device watches nothing and leans on its own physics. It is a ceramic resistor that sits low in resistance while cold. Current through it at switch-on is what supplies the assist, that same current heats it, and the heating drives its resistance up until almost nothing passes. Copeland's electrical handbook notes it stays wired in afterwards while having no bearing on how the compressor runs.
| What the device is | What ends the assist | What goes wrong with it |
|---|---|---|
| What the device isA relay watching current in the supply line | What ends the assistCurrent falling as the rotor comes up to speed | What goes wrong with itContacts that stick shut leave the start path live |
| What the device isA relay watching the start winding's own voltage | What ends the assistVoltage rising as the motor nears running speed | What goes wrong with itA tired coil or contact set holds in, or lets go early |
| What the device isA ceramic resistor that heats itself | What ends the assistIts own climbing resistance choking the current off | What goes wrong with itStill warm from the last attempt, so it cannot repeat it |
Why the assist has to be withdrawn
The start path is built for a moment, not for a shift. It carries heavy current for the instant it is wanted, and nothing along it is sized to hold that current continuously. Getting out of the way is therefore half the job, equal in weight to supplying the push.
Compressor makers name the failure that follows when it stays connected. Copeland's motor-burn guide lists a fault in which the start winding alone is uniformly overheated, with everything around it untouched, and points the technician at the starting capacitor and the starting relay. A burn that even and that confined says the winding was carrying current nobody designed it to hold.
That is the quiet version of this fault, and the costly one. The compressor keeps running. The room keeps cooling. Damage accumulates inside a sealed part that cannot be opened, and the bill arrives later as a compressor replacement rather than a relay.
A hot PTC has nothing left to give
A PTC that has just done its work is hot, and hot is its high-resistance condition. Asked to start the compressor again immediately, it cannot repeat what it delivered the first time. Secop's compressor instructions state the device needs a cooling period before it can restart a machine with full starting torque, and that brief cuts to the supply can leave a unit unable to start well after power returns.
This matters because it produces something that looks like unreliable electronics. A unit fails to start, gets switched off and on again several times, and every attempt leaves the next one worse off. Backing away and letting everything cool is what breaks the loop.
The same pause serves a second purpose. Pressures across the circuit have to settle before a machine is asked to turn again, and holding a unit back until they do is what compressor lockout is for. A restart forced through that window asks the starting device for more than it was ever going to give.
The two ways it fails, and what each does to the compressor
Every failure of a start device lands in one of two camps. Either the assist never arrives, or it never leaves.
When the assist never arrives, the motor cannot break away at all. Power reaches the windings, the field is there, and the rotor stays put. Current sits at the standstill figure, heat climbs quickly, and the overload protector inside the motor opens to stop the whole thing. From the ledge that is a hum, a click, and silence. Once it has cooled, the protector closes and the sequence runs again.
Repeated attempts are what turn this expensive. Each one holds the motor at its heaviest draw until the protection intervenes, and that is the worst load the windings ever see. A cheap part that failed cheaply becomes a compressor question if the unit is left cycling on it for a stretch.
When the assist never leaves, the compressor starts and keeps running. Nothing in the room suggests a problem. The start winding stays in circuit carrying current it was never sized for, and that heat has nowhere to go except into the coating on the wire. This is the version that gets missed, because everything an owner can observe says the unit is working.
A third pattern sits between the two. A device on its way out succeeds on some attempts and fails on others, so starting looks unreliable and appears to follow no rule. Hot afternoons expose it first, since the machine has more to push against when the air outside is warm.
| What the outdoor unit does | What that points at | What it costs if left alone |
|---|---|---|
| What the outdoor unit doesHums, clicks off, comes back to life after a rest | What that points atNo assist reaching the motor at all | What it costs if left aloneEvery retry puts standstill current through the windings |
| What the outdoor unit doesStarts and runs, but the machine sits hotter than before | What that points atAn assist still connected long after break-away | What it costs if left aloneHeat concentrated on one winding, inside a sealed part |
| What the outdoor unit doesStarts on some attempts and not others, with no pattern | What that points atA device that is marginal and losing its margin | What it costs if left aloneThe failed attempts are the ones doing the damage |
| What the outdoor unit doesRefuses to start right after a brief power interruption | What that points atA self-heating device that has not cooled yet | What it costs if left aloneNothing, provided the unit is left to settle |
Why this reads as a dying compressor, and what to ask
Both faults present as an outdoor unit that will not run, and that is the entire problem. The compressor is the thing an owner can hear failing to start, so the compressor is the thing that gets named. The device meant to start it makes no sound while failing.
What separates the two is access. A starting device sits outside the sealed shell, so it can be assessed on its own terms and ruled in or out cheaply. The compressor cannot be opened, so everything said about it is argued from how it behaves and from what its terminals give up. Anyone working carefully clears the accessible part before condemning the sealed one.
Temperature is the trap on the compressor side. A machine that has just stopped on its own internal protection reads as a broken circuit while that protection is out, and heavy metal holds heat long after the casing feels cool to the hand. A reading taken in that state can condemn a compressor that is perfectly sound. What such a reading settles and what it does not is covered under motor windings.
Two adjacent things are worth naming without wandering into them. A starting aid can be fitted on purpose to an installation that shipped without one. That is done where a weak supply makes switch-on hard on everything nearby, and it counts as a design decision instead of a repair. Separately, what a machine pulls at the moment of break-away is readable from outside with a clamp meter. No other single figure settles as much of this.
Before agreeing to a compressor replacement
A quote for the most expensive part in the system should show its work on the cheaper parts first. These are reasonable things to ask, and a competent supplier answers them without friction.
- Whether the starting components were assessed at all, and whether that happened before the compressor was
- Whether the compressor was hot at the moment anything was read from it, and whether a second reading followed after it cooled
- How many times the unit had been switched off and on before the visit, since those attempts are part of the history
- What current the machine pulled while it was trying to break away, since a stuck motor and an unassisted one draw differently
- Whether the run and start components were considered together, because they set up the same push between them
Where the answer goes next
A confirmed start-device fault is a cheap ending, and what matters next is what put it there. Units that short cycle, restart against unsettled pressure, or run with a weak capacitor all work their starting components harder than the design intended. Replacing the part without changing any of that buys a repeat.
A confirmed compressor fault opens a different set of questions, and none of them are about the relay. What an electrical ending leaves behind in the pipework belongs to compressor burnout, and the earlier warnings of a compressor losing performance are described on their own page. This one stops at the push and the part that ends it.
Ready to get started?
Tell us what’s going on. Symptoms, setup, photos, anything we should know. We’ll assess and come back with the right next step.