PowerLite starter motor.

Presumably it's the inertia type rather than a pre-engagement starter with a solenoid and fork?

The pinion is on a spiral on the shaft and, when the shaft suddenly starts rotating, the pinion's inertia causes it to move along the spiral and engage with the starter ring on the flywheel. If it doesn't move freely on the spiral, it can spin up at the same rate as the shaft, and fail to engage. Sounds like that's what is happening.

For the pinion to move on the spiral, there needs to be some light lubrication on it. If the garage cleaned *all* of the oil off it, that might be what's causing the problem.

no it is a pre-engaged type and it sounded rough when it went in.

All this design are pre-engage.

Not a Bendix drive. I'd be surprised if that is used anywhere these days. But the original Lucas was pre engage too.

Ah - forgot you'd once heard it.

But that one was replaced under warranty years ago and has been fine up until now. Although some do say they sound odd in operation. I'd say just different.

In that case, it seems odd that it spins without engaging. My recollection is that this type only turned the motor on after the solenoid had engaged the pinion.

the starter pinion has a one way clutch in it (so that if you hold the starter engaged it does not spin the starter too fast.) If that one way clutch fails or is gummed up, then it is quite possible for the pinion to be thrown in and the motor side to run, but it just doesn't turn the pinion. That could be quite random as to whether it starts the car, and may operate normally after a few tries (and the pinion warms a little) It would also match well with your mention of the oil leak, particularly if they dumped the starter in cleaner. (the guy that used to fix starters/alternators for me would not accept any for repair if they had been washed by the customer !! unfortunately he has now retired)

I don't quite understand what you're saying.

We've now ascertained that it's a pre-engage type starter. This type does indeed have a one way clutch - which allows the pinion to turn faster than its shaft - but not vice versa. This prevents the engine from driving the starter motor at 40,000 RPM or whatever if the pinion remains in engagement after the engine has started. [Not a problem with the earlier Bendix type because the engine in that case would wind the pinion back down its spiral - causing it to disengage].

But that doesn't really explain how a pre-engage starter can spin without rotating the engine. Unless the one-way clutch can fail in a way which allows the shaft to spin *without* the pinion turning?

that is exactly it, just as is often found on modern alternators, one way clutchs do fail, I have seen three in the last two years on alternators. I would guess that when starter one way clutches fail then they get changed with the whole starter on spec. when it won't start, rather than diagnosed and the actual part replaced. This would often be done at the roadside, while the alternator one way takes longer to develop as the vehicle still starts and runs, the only clue is some squealing initially which fades as the load reduces or it warms up and locks as it should.

As far as I know, all they did was steam clean the underside of the engine. Before replacing the sump gasket - which was what it was in for. (I'm getting too old to work under a car in the street)

They did mention the starter was soaked in oil, though.

Makes sense. Is it a possible DIY job to replace the one way clutch - assuming it can be bought as a spare?

I'm still just about up to working on things on the bench - unlike under the car. ;-)

I doubt that it would be available easily, although the starter should be easy enough to take apart to replace it.

You could always go for a wosp 2kw version, only 300 sobs