Alternator failure modes (long)

List message posted by: Scott Bilinski

Is there any way to tell if your alternator is on its last leg?
What are the signs?

Stator: Failure in Stator rectifiers. I have had diodes in the
stator rectifier fail open-circuit, meaning that the alternator
delivers only one-of-three or two-of-three phases. This causes a
reduction of max potential output current, so that the
alternator may not be able to carry the normal load.

The charging system still sort-of-appears-to-work, making
diagnosis difficult. If you monitor bus voltage with an
accurate dc voltmeter measured with engine running 1500rpm,
most of the loads switched off (light load), then the bus
voltage would be normal (14.2-14.5V). The Voltage Regulator (VR)
is still doing its job by chopping the field current down to a
duty cycle of 5-50%. A +- ampmeter would show no net charge or
discharge in the steady state (proper indication).

If you turn on all of the loads (especially landing/taxi/nav
lights), then the bus voltage will sag due to the inablility of
the alternator to supply these loads. Under these conditions,
the bus voltage may sag to 12.6-13.5V, and the ammeter would
show a net discharge from the battery in the steady state. The
VR is likely doing its job by applying full (100%, unchopped)
field to the rotor, but with one third to two thirds of
the stator gone, the alternator cant keep up with the load...
Eventually, the battery will be chronically undercharged,
leading to cranking problems, especially in cold weather.

There is also a marked increase in bus ripple. If you normally
hear a (hopefully faint if you did your audio grounds right)
alternator whine in your headphone audio, then after the
stator/diode failure, the whine gets a lot louder, and lower
pitched.

Most stator failures are caused when the soldered electrical
connection between the stator wire and the diode stack comes
apart due to heat/vibration. Sometimes this is the only
problem, and can quite easily be repaired, but this requires
total disassembly of the alternator. You have to use a "hard
solder" to repair these connections (not Pb/Sn electronic
solder).

Rotor: The brushes can wear down to where they no longer "ride"
on the rotor slip rings. As they approach the wear limits, the
contact pressure between the brush and the slip ring is
insufficient to keep the slip ring clean, meaning that the brush
contact to the slip ring gets intermittent and electrically
"noisy". This reduces the available field current.

This manifests as "reduced max output current", with similar
behavior of the dc voltmeter and ammeter when all of the loads
are switched on. The noise in the headsets may sound different,
however. As before, this is not the VRs fault. It will probably
still try to do its job of keeping the bus voltage at 14.25V by
applying 100% field duty cycle, but the high resistance of the
brushes against the slip rings prevents the alternator from
developing much field current, thereby reducing output. This can
also manifest as the "pulsating ammeter syndrome", along with
unnaturlly high resistance in the external field wiring...

The alternator whine in headset audio will sound less musical,
more like hash (bacon sizzling). If you have an ADF or AM
radio, tune to the low end of the band (200Khz or 540Khz,
respectively) and you might hear the alternator hash. If you do
remember to this before you have alternator problems, the
increase in direct audio and/or ADF/AM radio RF noise as the
brushes wear down can give you a clue as to what is going on.
I have seen Lorans quit as the RF hash from the alternator
increases over time and finally overwhelms the Loran signals at
100KHz.

The fix is disassembly of the alternator, installing new
brushes, cleaning or possibly turning the rotor slip rings on a
lathe. This is part of a normal alternator overhaul done by
overhaul shops.

Mechanical: As bearings wear, the sideways pull of the belt can
cause the rotor to hit the stator pole piece causing
catastrophic failure and total destruction. Bearings are
normally replaced during overhaul. At annual, remove the belt,
and see how much play is in the bearings.

MikeM
Skylane '1MM
Pacer '00Z

mikem writes:

List message posted by: Scott Bilinski

Is there any way to tell if your alternator is on its last leg?
What are the signs?

This is a good post to file away. I agree with all of it,
with minor qualifiers..

Stator: Failure in Stator rectifiers. I have had diodes in the
......

You can go a long time minus one diode and never know it..

There is also a marked increase in bus ripple.

WhatHeSaid. The best test is a portable oscilloscope on the alternator
output. You can SEE the hole in the waveform.

Most stator failures are caused when the soldered electrical
connection between the stator wire and the diode stack comes
apart due to heat/vibration.

Here I'll mildly defer. The {auto} ones I've seen are always dead
diodes, not bad solder joints. But it never hurts to try. Note some
diodes do fail only when hot, and "repair" themselves when cool.

Rotor: The brushes can wear down to where they no longer "ride"
on the rotor slip rings.

GM alternators used to have a "full field test tab" to bypass
the regulator. Ground it & the output voltage SHOULD go up
to 15++ volts and the engine/belt alternator all GROAN. If you
are doing a full-field test in-situ, be SURE all the avionics
are off-line.

Some brushes can be replaced without major work. Even if the
rings are dirty, with more brush pressure they often clean
up.

At annual, remove the belt, and see how much play is in the bearings.

And SPIN the pulley! It should be smooth and sweet.


BTW, there is an alternator parts supplier in most major cities.
These are the folks who sell parts to other rebuilders and work on
marines, 300A 28VDC truck units etc. (NOT an auto-parts retailer in
a strip mall..) If you don't require FCC blessed parts [1], you
might seek such out. They stock and/or can order brushes, diode
arrays, etc.

One last rant! Grounds, grounds, grounds!!! Inspect/clean the ground
bonds, be they engine to alternator, alternator to firewall, whatever.
Grounds are Murphy's way of making troubleshooting electrical issues
an adventure.


1] Your '540 is used in an airboat...
--
A host is a host from coast to
& no one will talk to a host that's close........[v].(301) 56-LINUX
Unless the host (that isn't close).........................pob 1433
is busy, hung or dead....................................20915-1433