On 8 Oct 1998, DANMAS@aol.com wrote:
> If it is started once in the morning, and ran
> most of the day, even one volt over the battery voltage will still recharge
> the battery during the day's travel, as long as excessive cranking is not
> required to start the car.
Entirely correct, and that is the state of affairs. Most batteries are
just charged half-full, if that.
> Normally, I get
> around 5.
That is exactly the consequence of running with a battery that is just
partially charged, if that.
> It may also be true that the generators maintained voltage better over their
> load range, but it is also unfortunate that the load capability of the
> generators was very limited.
My experience is from a -65 Jaguar, and as long as the engine is above
idle, it can take all loads and still charge reasonably correctly. But
the car has no heated rear window, no long-range lamps, and no injection
pump, so the demands are not that big. And the alternator is a huge beast
- my point is that the regulator works better - and has the added feature
of being adjustable.
> I assume this is a circuit of your own design?
Correct.
But at least here in Norway such things are also available commercially.
I think they cost a whopping $150-200 (which I'm sure is due to low
volumes, not component cost). They are much used for marine use (where
full charging is very important), but they are definitely also used for
cars.
I forgot to mention that such a circuit should measure the temperature
at the battery - a cold battery needs a higher charging voltage.
> I've always
> wondered why the designers chose to use a regulator circuit that emulated the
> "on-off" action of the relay type, rather than designing an analog system
Due to the long time constant of the inductance of the field coil, on-off
is fine as long as you do it often enough.
You're correct that doing it on-off so slowly that there is a reaction in
the actual charging voltage (like a relay does) is not particularily good.
In my design I use a differential P/I error amp measuring battery voltage,
and feed that voltage to a pulse width modulator of suitable frequency
that again drives the MOS output transistor and free-run diode. Thus
chopping the output just like in a switching power supply.
You could have used a fully analog (linear) output, too, but then you
would roast away some 5-10 watts for no good reason (and would need a
cooling fin too).
> It seems to me that the technology was there, and the
> integrated circuit manufacturing capability was sufficiently advanced for
> this.
The technology has been there, and has been inexpensive for the last,
say, 20 years. But even in mass production it would still cost a dollar
or two extra, and if people seem to manage without, then why bother?
> According to the manuals, the 18 ACR was rated at 45 amps. Do you have an idea
> how much yours wound up producing?
Only as a guesstimate. Firstly, I don't have a good ampere-meter above 10
(this could of course be a good excuse to get one ;-). Secondly, I'm
perfectly happy as long as it keeps fourteen and a half volts whatever the
load I care to throw at it ;-)
> Did the Schottky diodes
> fit without a lot of modification?
They are entirely different physically - being in small plastic TO-220
transistor houses. I fabricated a small glass fiber board with a bracket
which holds the six devices each with a small individual cooling fin. I
could only fit some time tiny fins, and was a bit worried since at full
output there would be a total dissipation of almost 20 watts (compared to
around 4 times that for the silicon diodes, BTW). No problem whatsoever,
it turned out, the cooling from the fan is excellent.
For the field circuit, I used 3 small diodes.
I somehow doubt that you'll see Schottky diodes in commercial
alternators, though. The ones I used probably cost more than the entire GM
unit ;-)
> 50 watts is not a lot, but every little bit
> helps. They would be of benefit even without the new regulator.
I would definitely address the regulator first, that is where the real
problem lies.
> I
> could see a very good market for modified Lucas units as you mentioned.
I really doubt if would make economical sense. Especially since the
number of madmen that really need it (driving their Triumphs all through
the winter) is quite limited ;-)
> Is the external regulator you describe small enough to be hidden out
> of site for these folks?
Using surface mount technology, it could certainly be made to fit within
the alternator - no problem whatsoever. You'd have to add additional
wiring, though: One sense wire for each battery terminal, and perhaps also
a temperature sensor. The wire that carries charge current to the battery
also would benefit from an up-rate.
And, as I discovered after a few months: The standard diodes are not up
to the job. Thats on an 18 ACR, though, they might be OK on a 15 ACR.
> I can see where you could get rich off this!
Oh, I think I would probably be better off spending that time maintaining
my Triumph! (Which reminds me: the OD sorely needs a new uni-directional
clutch - thats one of the drawbacks of engine tuning ;-)
> There's no doubt that a diode will increase the voltage output, but what will
> it do to the maximum current capacity?
Not as such. But it will significantly increase the current at which
there is still a proper charging voltage - which is really what you are
looking for.
A generator being able to output loads of current at 12.0 volts is of no
use to anyone. It is the current capacity at 14.5 volts which is
significant.
The diode thing is a standard trick-of-the-trade here, at least. Use any
standard silicon rectifier diode with, say, 3 amp capacity. And do check
that the voltage at idle isn't too high.
> If the applied load is not excessive, than the
> standard alternator will be adequate.
My point is that the standard alternator has room for improvement, and
that what you observe as an inadequate alternator output really may be a
problem with the regulator. Due to at least four reasons:
1. The regulator does not have enough gain to give you the current the
alternator is capable of, at the voltage you need it.
2. It does not compensate for the significant losses in the wiring and
ground return to the battery.
3. It is set very low to begin with (presumably to allow for
production tolerances).
4. It does not correct well for temperature.
If you like, you can do the following test to check this:
Remove the alternator rear cover.
Put an accurate voltmeter over the battery terminals.
Set the car to idle at high RPM.
Turn on enough electrical loads for the charging voltage to drop
significantly.
Observe the drop in charging voltage.
Bypass the regulator by connecting a wire from the field coil
negative side directly to ground.
Observe the battery voltage now. (Be prepared to remove the
bypass quickly if the voltage gets too high. Using a variable 10
ohm 10 watt resistor would have been ideal).
Perhaps repeat the test at a higher/lower load.
If the above test show that if the charging voltage increases
significantly when you bypass the regulator, then the root of the problem
lies in the regulator - not commanding the alternator to supply the output
it is capable of.
Egil
--
Email: egil@kvaleberg.no Voice: +47 22523641, 92022780 Fax: +47 22525899
Snail: Egil Kvaleberg, Husebybakken 14A, 0379 Oslo, Norway
URL: http://www.kvaleberg.com/ PGP: finger:egil@kvaleberg.com
|