Alternator Noise

On Tue, 07 Dec 2004 03:45:23 GMT, zatatime wrote:

On Tue, 07 Dec 2004 01:00:19 GMT, "Jon Woellhaf"
wrote:

All of which is much less likely if you've disconnected the ground first.
That way, you'll have to hit both battery terminals simultaneously to
experience John's scenario.


It can be done:-))

Roger Halstead (K8RI & ARRL life member)
(N833R, S# CD-2 Worlds oldest Debonair)
www.rogerhalstead.com
Jon

Yep. Got it. I missed the point and thought he was talking about the
alternator, not the battery.

Thanks for the reply.
z

zatatime writes:


Thanks for the explanation. I was confused and thought you meant to
disconnect the ground from the alternator before the field. I do
understand to disconnect the battery negative first and reconnect it
last for the reasons you've given. I've been that 'hot shot' before
and startled the hell out of me! Luckily nothing serious happened,
just a little black spot on the battery box.

Fortunately, typical GA starting batteries have limited current
capacity; something like http://www.trceng.com/services/coe.htm
will turn a 12" Crescent wrench bright orange before you can say
"Oh s&*8"


If I may indulge you, when disconnecting/reconnecting the terminals at
the alternator, is there any particular order, or precautions one
should take?


With the battery ground off; the order does not matter. What DOES matter
is getting each wire back to the correct spot. You may wish to tag them,
take a picture, and then remove them.

The biggest way to have grief is grounds. The case of the alternator
may have a grounding wire. (I'd hope so, but...) The VR will be on
the firewall and typically should be grounded via its mounting bolts.
But corroded/dirty ground connections are more common than ants at
a picnic. Shine things up bright before reassembling. Be sure there
are star lockwashers to bite into the joining surfaces.

It's also possible there are insulating washers/spaces behind certain
bolts. Don't lose them or mix up where they go.
--
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

(John_F) writes:

Find or purchase a clamp on amp meter that will read both AC and DC
amps. Turn on enough loads so that the alternator is generating at
least 50% of it's rated load. Measure the DC amps and the AC amps of
the alternator output. The AC current should not exceed 15% of the DC
amps. If it exceeds 15% then you have a bad diode or stator winding.
In many cases you can not get 50% of rated output if you have a bad
diode.

Interesting approach. I'm not sure how well it will work with the
tools a GA pilot has.

A real clamp-on DC ammeter is a rather specialized tool. The usual
ones are AC -- the clamp makes a 1-turn transformer out of the wire
you are around. It may or may not read on dc-dominant paths; I can
see the trasformer getting saturated and/or only reading accurately
on balanced waveforms.

Hmm, it should be possible to use a DVM in AC mode on the
alternator output. Fair-good DVM's are really inexpensive.
A True RMS one would tell you in a second. But it would take
some experimenting to say what AC voltages would mean "yes,
you DO have a bad diode."...

--
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

The clamp on amp meters that read DC and AC current use a Hall effect
IC. I bought one on Ebay for $60 over a year ago.
Most run of the mill clamp on amp meters use a transformer which will
only read AC.

On Tue, 7 Dec 2004 06:16:02 +0000 (UTC), David Lesher
wrote:

(John_F) writes:

Find or purchase a clamp on amp meter that will read both AC and DC
amps. Turn on enough loads so that the alternator is generating at
least 50% of it's rated load. Measure the DC amps and the AC amps of
the alternator output. The AC current should not exceed 15% of the DC
amps. If it exceeds 15% then you have a bad diode or stator winding.
In many cases you can not get 50% of rated output if you have a bad
diode.

Interesting approach. I'm not sure how well it will work with the
tools a GA pilot has.

A real clamp-on DC ammeter is a rather specialized tool. The usual
ones are AC -- the clamp makes a 1-turn transformer out of the wire
you are around. It may or may not read on dc-dominant paths; I can
see the trasformer getting saturated and/or only reading accurately
on balanced waveforms.

Hmm, it should be possible to use a DVM in AC mode on the
alternator output. Fair-good DVM's are really inexpensive.
A True RMS one would tell you in a second. But it would take
some experimenting to say what AC voltages would mean "yes,
you DO have a bad diode."...

(John_F) writes:

The clamp on amp meters that read DC and AC current use a Hall effect
IC. I bought one on Ebay for $60 over a year ago.
Most run of the mill clamp on amp meters use a transformer which will
only read AC.


I've not seen a Hall Effect meter near that cheap. [When I last used
on, such was a multi-thousand $$ Tek probe...] That will work on DC
as well as AC.
--
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

David Lesher wrote:
Jonathan Goodish writes:

I know that this is another frequent topic, but I'm hoping that
someone
can help narrow this down for me.

I have a rebuilt Chrysler alternator (from Electrosystems) on my
Cherokee that was installed about 400 hours ago. However, I am
experiencing alternator whine that varies with engine speed, but
which
disappears completely when I turn off the alternator side of the
master
switch. In addition, increasing electrical load seems to make it
worse.

You could have lost a diode. There are 6 main diodes, usually molded
into a unit.

If you have a friend with a portable oscilloscope, it's an easy test:
load the alternator and look at the waveform. You'll see the chunk
missing as compared to http://www.medar.com/images/3phase5.gif

Or pull the alternator [disconnect the battery ground FIRST,
please..]
and take it to an alternator shop. [NOT a chain auto parts store
who has an alleged tester.] They should be able to spin-test it.
AFAIK, having a test done by a non-FAA-anointed shop would be legal.
(Having them fix it, of course will get you sent to Gitmo..)

Alternator noise is often caused by an electrical "ground loop," as
the engineers call it, and it's not related to the taildragger
groundloop some of us are familiar with. The alternator generates noise
even if all the diodes are working, and in some airplanes even if all
the filters are in place. If there's a bad ground anywhere in the
alternator or power supply circuitry or radio and intercom circuitry,
resistance forms at that point and the ground current seeks another
path for a portion of the load. If this alternate path happens to be
through avionics, the noise will be picked up.
For example, if the grounding of the engine mount is imperfect, and
the alternator has no separate ground wire from it to the firewall,
some of the alternator's noisy current might seek a path through, say,
the electric oil temperature wire and through the gauge to the panel.
Now we have electrical noise through the panel to the rest of the
airframe, and radios grounded to various points of the panel, and since
gound connections between sections of the panel might be old and dirty
we could get voltage drops between those points and the radios pick
them up. I have also seen erratic oil temperature readings caused by
the spurious current.
That's my take on it anyway, and I have managed to eliminate noise
on occasion by cleaning all the ground connections I can find and
bridging suspect panel sections with ground wires to eliminate the
voltage drops. If you have a portable intercom, try making sure it's
NOT grounded except for its power supply. I've seen noise picked up as
a result of different potentials between the power line ground and a
screw that was holding the intercom to the pedestal. Velcro fixed that
one.
Some alternator noise comes from the field. There is magnetic
interaction with the stator when the thing is working, and noise can
come back up the alternator field line into the bus. Many alternators
have a filter capacitor on the field terminal.

Dan

"David Lesher" wrote

Hmm, it should be possible to use a DVM in AC mode on the
alternator output. Fair-good DVM's are really inexpensive.
A True RMS one would tell you in a second. But it would take
some experimenting to say what AC voltages would mean "yes,
you DO have a bad diode."...

Take the alternator off, and take it down to auto zone, and they will put it
on a machine to test the output and the diodes, for free.

Another possibility is that there is a ground loop in the audio system.
(using the frame as the ground to carry the audio signal)
--
Jim in NC

writes:






Alternator noise is often caused by an electrical "ground loop," as
the engineers call it, and it's not related to the taildragger
groundloop some of us are familiar with. The alternator generates noise
even if all the diodes are working, and in some airplanes even if all
the filters are in place. If there's a bad ground anywhere in the
alternator or power supply circuitry or radio and intercom circuitry,
resistance forms at that point and the ground current seeks another
path for a portion of the load.


I'm the first person to suspect ground loops. Copper-Aluminum
connections are infamous for this; Al in general enjoys oxidizing
to make our lives miserable. Add the vibration of an aircraft...

[I've wondered a CAD-weld type connection to the frame
for the battery ground, but there are too many minuses...and
that's only one ground...of many.]

But there's a clue in this line:

In addition, increasing electrical load seems to make it worse.

as that points toward diodes. At low loads you can limp along. At
full load, that {waveform} hole gets to be really important.

But, in any case, while you have the alternator out; clean up the
grounds...



--
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