Automotive Circuit Protection - Part 1

[garwillis@msn.com (Garfield Willis)]

Obligatory pardons for the dupes to those on both lists.

I intended getting back to this alot sooner, but got distracted. Wanted
to also address a number of errors/misconceptions found in the two
snippettes below.

One person wrote:

>1. Alternator wire breaks from alternator to battery and there is now no
>load on alternator.  Normal situation here is alternator fries diodes
>due to overvoltage.  Does High power transorb protect alternator until
>aircraft or boat or car reaches home port?

>2. Battery wire not properly connected to battery and battery becomes
>disconnected.  Alternator and entire electrical system experience high
>voltage.  Would not the fuel pump provide enough of a load to keep the
>alternator voltage low enough or safety?  Would Transorb be any better?

Another responded thusly:

>The alternator's control circuitry should protect against such an
>over-voltage situation anyway by manipulating the excitation
>current.  There's typically a Zener diode to sense over-voltage.
>Such a circuit can react quickly enough to protect the rest of the
>vehicle's electrics.

Then, same respondent further states:

>When the connection between battery and alternator fails, the
>alternator's unsmoothed ripple voltage is seen by electrical
>consumers. Note that most alternators supply at least 3 phases (6
>diodes) and the resulting ripple is only a couple of volts; enough
>to confuse ECUs, but hardly enough to cause damage to electrical
>consumers designed for automotive use.
>
>If you're looking at transsorbs to smooth the ripple, you should
>probably be looking elsewhere for something more useful.

Instead of treating what are mostly misconceptions above, I'll just run
down both cons & pros all together.

(1) "Load Dump" is a POSITIVE over-voltage transient out of the ALT, so
it's in the same DIRECTION as normal ALT output (rounds about 70V), and
therefore *doesn't* usually provoke a failure in the diode bridge.
Diodes normally fry from thermal cycle failures these days.

(2) Contrary to the second person's glib dismissal of LoadDump as merely
minutae ripple on the output of the bridge (which is complete nonsense),
it IS a very REAL threat that SAE and ISO have sped'd out extensively.
Such dismissal shows that the person isn't in a position to comment in
what often sounds like an authoritative tone, when they don't even know
WHAT Load Dump is, let alone that it even exists! Always better to ask
questions and learn something, than wax authoritative but clueless.

What Load Dump IS, is a deregulation-type failure in the ALT, where a
large enough change in load causes the ALT's regulator to be unable to
*quickly enough* cut back the ALT's field coil current (that's the
rotating part of the ALT, BTW). It's not a design flaw or failure of the
electronics (usually), but rather an intrisic weakness in the
method/means of regulation of ALT output; namely, controlling the ALT
output thru regulation of it's FIELD winding current. Basically, it's
like an inertia effect. You're controlling the ALT output via it's
smaller field current, but big inductors like the field winding can't be
turned off on a dime, they have electrical "inertia", because they have
stored magnetic energy in their windings. So if you have the ALT putting
out full tilt boogie amps and high rpms to boot (that's the spec'd worst
case scenario, BTW; max ALT output at max ALT rpm), and you try to
suddenly drop from 100A load to 10A say, that's a "load dump" (you're
dumping load from the ALT, not adding but *removing*) that will likely
cause the ALTs output to rise well above it's normal regulated setpoint,
as the ALT's regulator tries to shut down the field current to pull back
the output current.

You might think, well hell's bells, isn't that going to be happening any
time you turn off some major load? Ahh yup, BUT first off, you don't
normally have transients as big as above, PLUS the BATT is a tremendous
capacitor of juiceful proportions, and normally happily absorbs an such
transients. Whilst it's charging, the BATT also presents a pretty decent
load of it's own to the ALT. Take both of those away at once by
inadvertant disconnection of the battery and some other major loads, and
presto: "Load Dump".

This is enough of a major issue that the main Chip houses that design
the regulators, all have various "load dump" protection schemes they use
to promote their regulator chips overNabove their competitors. There's
even a debate as to whether you should try to suppress such transients
centrally at the possible source(s), or use a "distributed" scheme where
you sprinkle protection around at all the loads. Part of the reason for
this debate is that there are other transients/dangers in need of a
defense, so the issue is over how to most economically provide ALL the
protection that's needed, of which Load Dump dangers are just one piece.
[We'll get to those other dangers in Part Deux].

(3) The way we got into this was recommending transorbs as part of the
protection against these transients. But from the first poster above,
it's clear more needs to be said about the overall "plan". Transorbs are
NOT used to regulate power buses if the primary regulator goes outta
spec (like the ALT case above), they are there to protect the load
circuitry just long enough to call in for heavier reinforcements like
fuses or breakers if needed. Plus, they're there to absorb fast
transients that are too SHORT to ever actually trip a fuse/breaker. EMP
from a nearby electrical storm, or other kinds of EMI/RFI from other
electrical devices onboard, for example. So no, transorbs don't
temporarily "fix the problem" until you can get back to home port.

(4) Yes, there IS a design issue for "mission critical" systems, if you
don't want everything shutting down because of some transient/fault.
Enter in the "distributed vrs central" protection debate. Normally,
serious Load Dump is caused by some major screwup, either accidental
disconnect during repair procedures, or even wiring failure. But yeah,
if you're on the last lap of a roadrace with a huge purse at stake, you
don't want to shut down the fuel pumps if the ALT momentarily blips up
it's voltage to something that wouldn't harm the pumps, but MIGHT kill
the sensitive datalogging and telemetry equipment. Hopefully your ECU
designer's made IT adequately burly.

Just a make-believe what-if mind you. But HERE, you might design the
protection so if the transient is short, the transorb just swallows it,
and IF it does end up really long, yeah, maybe your transorb protection
ON THAT CIRCUIT or AT THAT LOAD pops the fuse/breaker just for those
devices, but you go on to finish the race. If OTOH, you put all the
protection at the ALT, then you've got to take a worst-case more/most
conservative approach, and that probly means you need to shut the whole
electrical system off when even a modest dump occurs. This might not be
best, if you're on that last lap, in the water offshore, or in the air.
So YMMV. Ain't "engineering" fun?

If you don't think these are real-world problems, consider the
following: One major automotive EE house has a ALT regulator design that
CROWBARS the ALT output in the event of a load dump. If you're
unfamiliar with the term, it's an electrical analog to putting a
"crowbar" across the ALT's output terminals. IOWs a nice big fat short.
And the nature of crowbar circuits is that usually they STAY shorted as
long as any current continues to flow once they're tripped. That's a
REAL design that's offered to the automotive OEMs. And it would mean
that in the case of a Load Dump transient, the ENTIRE electrical system
would most definitely and robustly be shut down.

Food for thot if your intention is to keep running and win. This will
come up again as the mini-series unfolds. :)

Gar


----------------------------------------------------------------------------
To unsubscribe from gmecm, send "unsubscribe gmecm" (without the quotes)
in the body of a message (not the subject) to majordomo@lists.diy-efi.org