Run-on Starter?

[MikeMl]

Automotive, compound-wound, segmented-commutator, 12V DC, starter motor, the kind with a Bendix drive, not the kind that uses a solenoid to engage to the ring gear.

If the Bendix fails stuck in, and the starter is back-driven by the engine after it starts, does the starter motor produce any open-circuit voltage (as a generator)?

Since there no current can flow back to the battery because I'm assuming that the starter relay released, there might be no path for field current to flow, so does that mean to act as a generator, there would have to be some residual magnetization of the pole pieces?

I'm trying to determine if there is a reliable way of knowing that the Bendix stuck.

 

[Nigel Goodwin]

I would have thought the noise it makes would be instantly obvious?.

 

[MikeMl]

Not in an airplane.

 

[JimB]

I would be surprised if there was not some remanent magnetism and the starter would act as a generator.

Connect a voltmeter and see.

If there is no volts with the engine running - bendix probably not stuck.

If there is volts which vary with engine speed - bendix definitely stuck.

JimB

 

[MikeMl]

Kinda hard to do with a swinging prop about 6" forward of the ring gear. Also, I don't want to bugger the Bendix, or for that matter, remove the starter from airplane...

I'd kinda like to generalize the answer without having to test a few hundred starters

 

[JimB]

Help me understand a few things here.

Do you have an aircraft with a suspected stuck bendix?
Or is this just some thought experiment to find a way to diagnose the problem if it should occur at some thime in the future?

Why do you have to test a few hundred starters?

I appreciate that the HSE, OSHA and comonsense dictate that probing around with a multimeter in close proximity to a rapidly propellor can be detrimental to your health.
However, it is surely possible to make a secure connection to the starter terminal while the engine is stopped and run some suitable leads to a meter which is in a safe location before starting the engine.

Also, instead of running the engine, why not ensure that the ignition system is isolated suche that the engine cannot start, maybe remove the plugs to make the compression zero and then turn the engine with the propellor whilst monitoring the starter motor volts?
That way you will not bugger the bendix or the starter motor.

But, what makes you suspect that the bendix is stick in the first place?
You have already inferred that it is impossible to tell whether the bendix is stuck when the engine is running.
If the engine has been run for a significant time with a stuck bendix I suspect that the starter motor will be well and truely buggered, the bearings the commutator and the bendix. And the ring gear may not come out of this very happly either.

So, do you have an ailing aeroplane or not?

JimB

 

[MikeMl]

Neither of my airplanes is ailing. The Cessna 182 (see my avatar) is powered with a Continental engine which does not use a Bendix starter drive. It uses what is called a Sprag clutch, which does not have the same failure mode as a Bendix. It is extremely unlikely that the starter drive on the Continental engine can fail in a manner where the engine would back-drive the starter.

My other airplane is powered with a Lycoming engine which uses a ring gear on the back of the prop, and a Bendix engagement. However, the original compound-wound segmented-commutator starter motor was replaced with a modern, lightweight starter with permanent magnet field. In the thirteen years I have owned this one, the Bendix failed once, leaving the starter gear engaged to the ring gear, but I knew instantly that I had a problem.

The reason I knew I had a problem is because I had previously wired a LED indicator back to the cockpit. The LED is intended to sense the voltage applied to the starter via the starter relay. Another common failure mode is welded contacts in the starter relay or a bad key switch, where the starter remains powered via the start relay even though the start key switch is released. In the case of the stuck Bendix, the back-driven permanent magnet starter ran as a generator so even though the starter relay broke the connection to the battery, the voltage produced by the starter lit the LED after I release the start switch. I shut down the engine, and found the frozen Bendix.

(See attachment, below)

I participate in several aircraft owners forums. On one of them, one of the posters just had a stuck Bendix on a Lycoming engine but he has the older wound-field type starter. That resulted in a catastrophic failure of the starter, which cost big bucks to repair. He did not have my LED indicator, but the question came up, would my little circuit have indicated his problem. It depends on if there is sufficient residual magnetism in the starter for it to produce a few volts at a few mA to light the LED?

I am going to borrow an older starter from an aircraft mechanic that lives further down the runway. I will spin it up on the workbench using a battery, with my scope connected to the starter terminal. This should show me what happens during the time that it takes the starter to spin down after I disconnect the battery. Only problem is, even if the one I test does it, would they all do it? Would they all produce the same polarity?

 

[alec_t]

Would they all produce the same polarity?

Build a monitoring circuit which doesn't care about polarity and will sense just a few mV.

 

[RCinFLA]

Starter motors have their field winding in series with their armature so once the circuit is broken there would be little excitation field to produce back voltage.

The series connection of the field winding gives them more power torque the greater the mechanical load. You don't want to test a starter with no mechanical load as its rpm's can get destructively high. A short burst is okay but expect the starter to do back flips if not held down.

 

[MikeMl]

Starter motors have their field winding in series with their armature so once the circuit is broken there would be little excitation field to produce back voltage.

Roger, that is why I'm interested if residual magnetism will do anything. I'm only trying to light a LED.

The series connection of the field winding gives them more power torque the greater the mechanical load. You don't want to test a starter with no mechanical load as its rpm's can get destructively high. A short burst is okay but expect the starter to do back flips if not held down.

Roger, I've been there, done that! I'll just blip it with battery voltage.

Since I can likely borrow one which is not currently mounted in an aircraft, I might be able to rig it so I can back-drive it with a drill-press, or a lathe....

 

[jpanhalt]

How many field coils are in the starter in question? I am aware of 2-coil, 4-coil, and 3-coil configurations. (That is in reference to older Ford starters.) The 3-coil version uses the third coil as a regulator. Getting to its connections in the Ford starter is relatively easy. If you have that configuration, you might use a small current to the drive coils or to the control coil to induce a voltage in the other coil, if the armature is turning.

John

 

[MikeMl]

I don't own one of the starters in question. Nor can I modify the normal wiring or operation of a starter in an FAA Certified Aircraft. I am trying to work within the constraints imposed by the FAA on what they consider a minor modification, i.e. adding an LED circuit to indicate the status of the existing equipment...

 

[alec_t]

I knew it would come in useful one day..... I've just dug out an old 12V Ford car starter motor, Bendix type, which I've been hoarding unused (as you do) for the last 20 years or so. Rotating the shaft somewhat jerkily, and obviously slowly, with a finger on the cog gives about 2mV output. Clearly there's some remnant magnetism after all that time (or else the Earth's field is doing the trick). So my guess is any starter motor would give a suitable output voltage for monitoring the Bendix.
For what it's worth the label on the motor reads:
Ford
81AB11000AB___ 012109
8M90 PE 12V____ 25326

 

[jpanhalt]

It looks like alec_t may have answered the question empirically. If that doesn't work out for the aircraft starters in question, then maybe a trickle nudge to the coils may be worth pursuing. I understand the problem of actually making new connections to the coils in a certificated airplane. I was thinking it might be a homebuilt.

John

 

[MikeMl]

I knew it would come in useful one day..... I've just dug out an old 12V Ford car starter motor, Bendix type, which I've been hoarding unused (as you do) for the last 20 years or so. Rotating the shaft somewhat jerkily, and obviously slowly, with a finger on the cog gives about 2mV output. Clearly there's some remnant magnetism after all that time (or else the Earth's field is doing the trick). So my guess is any starter motor would give a suitable output voltage for monitoring the Bendix.
For what it's worth the label on the motor reads:
Ford
81AB11000AB___ 012109
8M90 PE 12V____ 25326

Thank you for doing this. I will proceed with some local testing where I intend to spin a similar starter at the RPM as driven by a ring gear from an aircraft engine at slow idle. I will count the teeth on the ring gear and on the bendix to see what the speed up is.

 

[MikeMl]

...then maybe a trickle nudge to the coils may be worth pursuing...

By that, I think you mean back feeding the starter a small current pulse from my circuit, and then seeing if that excites some output that lasts longer than it would if the starter were not spinning?

 

[alec_t]

I will proceed with some local testing where I intend to spin a similar starter at the RPM as driven by a ring gear from an aircraft engine at slow idle.

If your motor is anything like mine you will need a gear that mates with the Bendix cog to drive it. The shaft on mine is not accessible, nor is half the cog circumference because of the Bendix housing (which also includes the bearing for the shaft end).

 

[MikeMl]

If your motor is anything like mine you will need a gear that mates with the Bendix cog to drive it. The shaft on mine is not accessible, nor is half the cog circumference because of the Bendix housing (which also includes the bearing for the shaft end).

Yes, you are right. The front bearing of the starter motor is in the bendix dome. However, I would have to jam the bendix cog so that the starter could be back driven. If the Bendix is normal (not failed), then the little cog spins freely when you try to back drive the starter. That is the entire premise of the Bendix drive.

I guess my only option will be to spin the starter with a battery, and then use my digital scope to capture the voltage decay as the battery is disconnected while the starter coasts down. I can try this at various smalll load currents.

To detect a stuck Bendix, the good news is that there would have been a start attempt (applying battery voltage to the starter coils) just before the engine started and the Bendix failed. If the magnetization is not long lasting, at least it will have been refreshed just before I'm trying to detect it.

 

[alec_t]

Beware the back-emf when you disconnect the battery. Don't kill your scope.

 

[jpanhalt]

By that, I think you mean back feeding the starter a small current pulse from my circuit, and then seeing if that excites some output that lasts longer than it would if the starter were not spinning?

That is basically it. I would have to do experiments to form a solid suggestion and am simply thinking out loud. In a non-magnetized system, if the armature is turning and you add a small, pulsed current to the field. My guess is that the back emf from that current will be different than if the armature is still. I have seen motor speed controls based on a similar principle. The emf measured between PWM pulses is proportional to rpm. It was more than 10 years ago that I saw that paper, so details are fuzzy.

John