How to spot a failing motor before it fails

Why this matters

Motors give warning. They almost always do.

Bearings do not fail on a Tuesday out of a clear sky — they degrade, from smooth to slightly rough to gritty to notchy to seized, over dozens of flights. Magnets do not leap off the bell — the adhesive softens under heat, the magnet shifts a fraction, the airgap changes, and the motor gets hotter and draws more current for weeks first. Windings do not short overnight — the enamel is chafed or cooked, the insulation degrades, and the motor runs hot for a long time before the short is complete.

The failure that drops an aircraft out of the sky is nearly always the last event in a process that started weeks earlier and was visible the entire time, to anyone who knew where to look.

That is the part worth internalising, because the stakes are not symmetric. On a 250 g quad, an in-flight motor failure costs you a quad. On a 5 kg aircraft it is, in the general case, unrecoverable — a quadrotor has exactly four actuators and needs all four, and what comes down is a heavy object with no control authority. It can hurt somebody. So the inspection below is not fussiness. It is the only part of this that you actually control.

The hand test

Props off. Always props off. Then spin each motor by hand, slowly, and pay attention.

A healthy motor spins freely under a flick and coasts to a stop rather than stopping dead. It sounds smooth — a faint, even whirr, with no texture in it. It has no detectable play when you try to rock the bell. Every one of the four feels the same.

Now, the faults, and what each one feels like under your fingers:

• Grinding or gritty. Contamination in the bearing — dust, sand, grass seed, the fine grit that lives in every field you have ever flown from. It feels like turning a shaft through sugar. This never improves. A contaminated bearing is an abrasive, and every revolution from here makes the fault worse.
• Notchy, or catching at one point. You turn the bell and it hesitates, then releases. That is a bearing whose races or balls have started to pit, or debris sitting in the airgap, or a shaft that is no longer straight. All three are progressive.
• Play. Hold the stator, grip the bell, and try to move it — up and down along the shaft (axial), and side to side (radial). A healthy motor gives you essentially nothing. If you can feel the bell move, the bearing has worn, and a motor with detectable play is on borrowed time. This is the check people skip, and it is the one that most reliably predicts a failure.
• A bell that scrapes at one point in the rotation. You will hear it as a tick once per revolution and feel it as a rub. Bent shaft, or a bell knocked out of true — almost always after a crash. It also means the magnets are passing the stator at an uneven gap, which is a heat problem as well as a mechanical one.
• Stiff. Higher drag than the others. Not grinding, not notchy, just harder to turn. This is the subtlest fault and the one that needs the comparison below, because "stiff" is meaningless in isolation and obvious when you have three others in your hand.

The comparison principle

Here is the idea the whole article rests on.

You do not need to know what a good motor feels like in the abstract. You do not need a reference, a spec, or experience. You have four of them, they came from the same batch, they have flown the same flights, and they should feel identical. The one that is different is the diagnosis.

This is why you check all four, every time, one after another, rather than checking the one you are suspicious of. A single motor in isolation tells you very little — your hands are not calibrated and your memory of last month is worse than you think. Four motors in sequence tells you a great deal, immediately, with no equipment.

It is also why a motor that has been swapped in from a different set is harder to assess, and why replacing motors in matched sets is not merely tidiness.

What the aircraft tells you

The hand test catches most of it. The aircraft catches the rest, and it has been telling you for weeks.

• Uneven motor temperature after a flight. Put your hand on all four bells the moment you land. One materially hotter than its siblings, on the same aircraft, on the same flight, is either mechanical drag or a weakening motor — and heat is the energy that did not become thrust.
• One motor consistently commanded harder in the log. Look at motor[0..3] in a steady hover. This is an extremely informative trace and almost nobody looks at it. Be careful about what it means, though: an output offset is usually the FC correcting for the airframe, not the motor, so read that first before you condemn anything. Pulling the log is a two-minute job and it is the only objective record you have.
• Rising current draw for the same flight profile, over weeks. A motor that is getting harder to turn burns more current to do the same work. If your hover current has crept up across a month on an unchanged aircraft, something is dragging. Trend it. This is the earliest signal available to you and it is available for free in every log you take.
• New vibration that was not there before. A bearing on the way out is a vibration source, and it will show up in the gyro trace long before you can feel it — see why your quad is shaking. Vibration is also self-amplifying: the bad bearing shakes the motor, and the shaking finishes the bearing.
• An RPM trace that is worse on one motor. With bidirectional DShot you get per-motor RPM back from the ESC. A motor that is noisier in that trace, or slower to accelerate to a commanded change, is telling you it is mechanically or electrically compromised — and a motor that struggles to follow commands is also the motor most likely to desync.

Things that kill motors early

Prevention beats detection. In rough order of how often it is actually the answer:

• Water and grit. Bearings are not sealed against the world. Landing in wet grass, flying off sand, spraying the motors down with a hose after a muddy session — all of it puts abrasive or corrosive material where the balls run. If you fly in dirt, expect bearings to be a consumable.
• Crashes. A hard impact bends the shaft, knocks the bell out of true, or shifts a magnet. A motor that survived a crash and now feels different is not the motor you had before the crash.
• Heat. A cooked motor is permanently weakened. Once the magnet adhesive has softened and the magnets have lost flux, the motor draws more current for the same thrust for the rest of its life — and therefore runs hotter, forever. Heat damage does not heal.
• Motor screws that are too long. They bottom out into the windings, chafe the enamel, and eventually short it. This is entirely avoidable and it is created fresh every time somebody uses the screws that came in the bag rather than the ones that came off. Check the length against the mounting depth on every swap.
• Living at the thermal limit. Over-propped, over-loaded, hovering at two-thirds throttle. The motor is not failing today, but it is being consumed steadily, and it will fail earlier than it should.
• Vibration from an imbalanced prop. A prop that is chipped, repaired or simply out of balance shakes the motor it is bolted to, at the frequency the bearings least enjoy, for the whole flight. An imbalanced prop is a bearing-destroying machine, and it is the cheapest thing on the aircraft to replace.

When to retire a motor

Be decisive about this. Retire the motor if any of the following is true:

• Any detectable play, axial or radial.
• Any grinding or notchiness, however slight.
• Any motor that runs materially hotter than its siblings on the same aircraft, on the same flight, once you have ruled out the airframe causes.
• Any motor that went through a hard crash and now feels different from the other three.
• Any motor from a set where one has already failed. They have the same hours, the same abuse and the same manufacturing batch. One failing is information about the other three.

Bearings can be replaced on many motors — the bell comes off, the old bearings press out, new ones press in — and on a larger motor with a decent price tag that is often worth doing. It is not always straightforward: some motors are not designed to come apart, the C-clip is easy to lose, and pressing a bearing in crooked leaves you worse off than you started. If you are going to do it, practise on a motor you have already written off.

But be honest about the arithmetic. A motor is a cheap part. An airframe is not, and a person is not. On a heavy platform that calculus is not close, and it never has been. A motor you are unsure about is a motor you replace. The uncertainty itself is the answer.

A maintenance rhythm

None of this works as a one-off inspection. It works as a habit, and the habit is short:

• Before every session: props off, spin all four by hand, compare them. Thirty seconds. This is part of your preflight, not a separate ritual.
• After every flight: put your hand on all four motors. You are looking for the odd one out, not an absolute temperature.
• Every flight, passively: record the log. You are not going to read most of them. You are building the baseline that lets you notice, in three months, that the hover current has crept up by 10%.
• On a schedule, not on failure: for anything doing real work — heavy lift, commercial, anything carrying a payload over anything you care about — replace bearings or motors on a service interval you decide in advance, rather than waiting for the aircraft to decide for you. The correct interval depends entirely on your motors, your environment and your loading, so set it from your own logs and your own inspections, and then actually hold to it. A component replaced early cost you a part. A component replaced late cost you the aircraft.