Hi Greg,

On 26-Sep-03, Greg Mayman wrote to George White:

 GM>> First you sniff a bit of the AC input, amplify it and clip it and
 GM>> feed it back to the transistor bases so they turn on/off at zero
 GM>> crossing.

 GW>> You also, for a self contained unit, need to take some of the AC
 GW>> and create a sufficiently high rail relative to the emitter to
 GW>> enable you to drive the base hard on when there is very low AC
 GW>> from the source. The most difficult part!

 GM> Yes, amplify it up to a couple of volts. while the AC is still
 GM> just past zero crossing.

 GM> I have included the quote from my own message to show that we are
 GM> apparently saying the same thing 

We are, in general :-). However I've spent over 10 years in automotive
electronics design working for CAV (as it was know then), including
work on their 440 regulator (I still have some in the loft!). I also
have a degree that majored on semiconductor physics (though it is 30
years ago now, I don't think the basic physics has changed).

 GW>> Correct. The dedicated isolator units are just two isolation
 GW>> diodes. Leaving the alternator sensor connected to the primary
 GW>> (it should be the vehicle or existing) battery means that
 GW>> continues to be charged correctly and the slave (usually a
 GW>> caraven, trailer or separate RV) battery will get charged up as
 GW>> well. For systems without a separate alternator sense wire (and
 GW>> they do exist) things are more difficult, the batteries are never
 GW>> fully charged.

 GM> No, they _would_ reach near full charge eventually, just that
 GM> they'd take a bit longer.

That "near full" is probably 99% of full and most people would never
notice the difference. However, nit picking, I only said "the
batteries are never fully charged" for an alternator sensed system.

 GM> There are two factors here:

 GM> One is that the alternator regulators that sense the output of the
 GM> alternator directly are almost always set about higher than
 GM> optimum to allow for voltage drop in the cables to the battery.
 GM> Some will charge the battery to 14.5v.

When I worked on the CAV 440 regulator it had a choice of 3 settings,
from memory 12.5 (for NiCad systems), 13.5 (normal) and 14.5 (for
heavy use/cold climate).

 GM> The second factor is the behaviour of the diode. The voltage drop
 GM> across the diode is not a constant, it is proportional to the log
 GM> of the current. Also the voltage drop DECREASES as the temp goes
 GM> up. So at elevated temperatures and reduced current, the diode
 GM> drop goes down considerably.

Indeed true, however there is _always_ some volt drop, and of course
the lower the current, the lower the temperature rise within the
diode, so there is a balance...

 GM> This means that the alternator can still be putting current into
 GM> the battery right up to the optimum voltage, but at a lower charge
 GM> rate than it would have with the battery connected directly to the
 GM> alternator.

It all depends on the "optimum voltage", which depends on use
conditions and battery type.
I agree with you in general, however as there is _always_ _some_ volt
drop in the diode, for a schottky diode this is approx 0.1 Volts, so
compared with the direct wiring it is, for an alternator sensed
regulator, always _slightly_ less charged. If the alternator was set
to overcharge the battery (not unknown) then it would _improve_ the
charging regime and battery life...

 GW>> Most modern alternators don't have any facility for altering the
 GW>> terminal charge voltage.

 GM> Very true.

From the manufacturers point of view it stops people who don't know
better setting things up wrongly and then having problems when they
kill batteries through overcharging or their engines won't start in
cold weather because the batteries are undercharged and then blaming
the manufacturer for the problem (and sueing them in these litigious
times).

 GM>> There is a lot of stuff on the 'net, but only very rarely do you
 GM>> find data for outdated devices. I suspect the same would apply to
 GM>> CD-ROMs :-(

 GW>> But at least you can keep the _old_ CDROMS and still have the old
 GW>> data.

 GM> The problem is that most of the old data never got onto CD-ROMs in
 GM> the first place.

Unfortunately true. However I have a reasonable collection of old data
books to hand... :-)

 GW>> They were normally just diodes to protect the gate against
 GW>> breakdown. They clamped the gate voltage to 1 diode drop above or
 GW>> below the other terminals.

 GM> Yes, they could be clamping against reversal of the control
 GM> voltage.

 GM> OTOH if they are zeners, they could also clamp the forward
 GM> voltage. Most insulated gate FETs are rated to a maximum
 GM> gate-source voltage of 15v.

We're agreed then :-).

George

--- Terminate 5.00/Pro