Alternator Controller Mk.I

[Bigglez]

Greetings Kurt,

I bought a DMM today with a DC current clamp sensor and clamped around the 4 gauge battery cable between the unit and the battery bank (2-12v deep-cycles). This is where I got the 220a reading. Worked great!

This is something new, usually the clamp on type
only work on AC. BTW, if you thread the wire through the
clamp you can scale the meter sensitivity (two turns = 2x
reading, etc.). I assume that would also work for the DC
sensor, but perhaps there is a limit due to saturation. It
would be worth experimenting with using a light bulb load
for future reference.

The 18.5v was measured across the battery terminals right thereafter.

This is good data, but it will boil your batteries!

Comments Welcome!

[jwax]

Running that under load for any length of time may need some additional alternator cooling? 110 amps per? Any idea what they are rated at?

[Bigglez]

Greetings John,

Running that under load for any length of time may need some additional alternator cooling? 110 amps per? Any idea what they are rated at?

half throttle with all four field wires connected directly to 12v (full-blast), my readings were: 18.5V / 220A

Spread over four units, that's about 55A each or 660W at 12V.
The alternators are "1970s vintage rebuild Fords". Yes, they
are under stresss, but don't forget there is no controller in play.

Comments Welcome!

[jwax]

Oops! My bad. Thanks Peter- I was thinking two alternators.

[kheston]

With the field on one alternator connected directly to 12v, I read 2.95A with my DC current clamp.

I needed to shut things down, however, bearings were making more noise than I was comfortable with. So, I didn't get a frequency (RPM) reading yet.

I went cheap on the alternators and bought them for $20 each at the junk yard. At least one needs a new main shaft bearing. I'll replace all of them and fire it up again today or tomorrow.

[CeaSaR]

Kurt,

That is what I had envisioned in a previous topic (Feb. 2007). "See this."
Would you be willing to share your wiring diagram? I realize it would differ
for each type of alternator, but a starting point would be nice. Keep up
the good work.

CeaSaR

[kheston]

CeaSar,

Would you be willing to share your wiring diagram?

Count on it!

-
I finally got around to taking some more readings. The frequency counter I have wasn't able to sense the DC ripple, not sure why. So, I tried to fudge something similar by using the stator lead on the alternator and a pulse lead off of the engine's ignition coil. Here's what I got:



The first column is my guess at what the engine rpm was based upon throttle position and sound.

My drive pulley is 4.25" and the alternator pulleys are ~2.5" each. So, the alternators are spinning at nearly twice the engine speed (4.25 / 2.5 = 1.7).

I can only speculate as to why the 3800 rpm readings are lower than the 2000 rpm readings.

Does this change the spec for our testing circuit at all?

[kheston]

Found this while Googling:

http://www.amsterdamhouseboats.nl/voltage_regulator.htm

How hard would it be to replace his manual adjuster with a microcontroller?

[kheston]

Just ordered the components for Peter's test circuit from Jameco. Here goes...

[kheston]

I'm still in electronics kindergarten, but as I understand it, the left side is my power supply. 1 provides power for 3 and 2 provides power for 4. 5 is a pot that allows adjustment of the charge rate.

Putting the cart before the horse...is 3 (or 4) a potential place to splice in a microcontroller?

--------------

I'm looking at running a full Linux kernel on an ARM-based single board PC something like: http://www.embeddedarm.com/products/boa ... ct=TS-7250

The idea is that at some point I'll be able to do remote monitoring/adjustments. Once I get one of these, I'd like to start implementing circuits to measure battery voltage and RPM along with controlling a starter relay and stepper for throttle adjustment. Suggestions? Links?

[Bigglez]

Greetings Kurt,

Just ordered the components for Peter's test circuit from Jameco. Here goes...

I've been in "computer upgrade" for the
past couple of weeks, so my time here has been limited - I'm back!

The design was done on the assumption you might be used
to shopping at Radio Shack, I didn't know you were in the
valley until later. The Mk.I design has no control loop from
output to input, as we had to answer some assumptions about
the alternators first. It is a building block, so don't cramp
your construction. We'll add stuff on later, leading up to a
uC system.

Comments Welcome!

[Bigglez]

Greetings Kurt,

I'm still in electronics kindergarten, but as I understand it, the left side is my power supply. 1 provides power for 3 and 2 provides power for 4. 5 is a pot that allows adjustment of the charge rate.

Correct. The schematic uses symbolic nets to avoid the
clutter of drawing every "wire".

The stacked voltage regulators provide stable 5V and 10V.
The op amp compares the 'demand' from the pot with the
feedback from the output and drives the current amplifier
and the alternator field. Most of the other components
are for protection and safety.

Once we have this working and can map the output of
the alternator against the drive to the field coil we can
work on two feedback loops. (1) Voltage control to
throttle back the field when the battery is charged
(and/or the load is light). (2) Load current to get load
sharing across four alternators.

Basically the development of this circuit would be used
for each alternator, some parts are redundant (sharing
the local power regulators, for example) .

Certainly it would be smart to control the throttle of
the engine from the demand in the electrical load.

Putting the cart before the horse...is 3 (or 4) a potential place to splice in a microcontroller?

Sure! What I have in mind is using the ADC converter(s) in
the uC side to capture operating voltages (that represent
the load voltage, load current, and load sharing across
the four alternators). The uC can also drive the analog
'demand' pot for a totally closed control loop.

I read your post on RPM data and couldn't grasp what
was happening. Not an issue, and most likely needs
a bit more work with instrumentation. The engine speed
from the flywheel or spark plug is obviously the better
route, and alternator speed should be available from
the alternator ripple. (With multiple alternators the
pick off has to be before the outputs are summed).
Even if all were at the same speed, there will be phase
differences and phase (or speed) jitter.

Comments Welcome!

[kheston]

Peter,

Welcome back from upgrade hades. I've been there countless times. While in the thick of it, I try to remind myself that if it was easy, I'd be out of a job.

I read your post on RPM data and couldn't grasp what
was happening.

Neither could I. I was expecting to get some sort of ripple reading from the alternator armature output, but perhaps my $89-Fry's special frequency counter isn't sensitive enough. The other readings were a poor stand-in, I know. I wonder if I used another load without batteries in-line whether I'd have a more profound ripple to measure? Not sure. Otherwise, I've been Googling for a photodiode-reflective-tape-or-hall-effect-based contraption that I could plug in to a serial port that might provide an accurate reading, but am not having any luck.

I did some thinking about what I'd like to know and control and came up with this:http://delanobay.com/DcGenSet/generator ... system.pdf
It's a bit of a Taj Mahal and there are things I could do without. But if I had all of it, there'd be plenty of data and control to keep the batteries healthy and for collecting efficiency metrics. We'll see how close to it I (with lots of help) can get.

[dyarker]

RPM data??

Three phases summed inside each alternator, four alternators randomly (as far as phase) summed. I wouldn't expect any useful freq readings.

Try a few turns of wire on soft nail near one of the air vents on an alternator. It might pick up a single phase signal from the field winding going around inside. If it works (and not too much noise from brushes), you can make a "pretty" version for permanent installation.

Or glue two (for balance) permanent magnets to engine pulley, and use coil or hall effect device to count pulses.

Cheers,

[Bigglez]

Greetings Dale,

RPM data??
Three phases summed inside each alternator, four alternators randomly (as far as phase) summed. I wouldn't expect any useful freq readings.

Correct. The plan is to isolate the AC ripple from each generator
(using ORing diodes at the outputs to feed the load) and divide
the count by three (to get one phase) and again by 'x' to get
RPM, where X is the number of pairs of poles.

Kurt didn't have the diodes and his frequency meter may not
like the DC component.

Try a few turns of wire on soft nail near one of the air vents on an alternator. It might pick up a single phase signal from the field winding going around inside. If it works (and not too much noise from brushes), you can make a "pretty" version for permanent installation.

Or glue two (for balance) permanent magnets to engine pulley, and use coil or hall effect device to count pulses.

Also good suggestions. I would like to avoid modding the
factory alternators (to make maintenance easier). Sampling
one phase winding would be sweet...

The engine speed could be optical or magnetic (inductive),
Kurt is looking at options.

Comments Welcome!