Re: [DML] Auto trans governor NO-NO-NOs

["Mark Hershey" <dmcinfo@xxxx>]

Rats! Two mistakes in as many days. Guess I'm really slipping. Sorry again,
Ralf......I'm not an engineer. Try to be helpful and then go off and mislead
folks into potentially dangerous situations like capacitor explosions.
Reckon I'd better reconsider offering up advice here.

As to the risks of governor repair, I guess I drew my conclusions from a
statistically invalid sample of one--mine. Ran the darn thing for about
three years with my caps in backwards. Too cheap to buy new ones so I just
turned 'em around and kept going another 86,000 miles. Guess one of these
days I ought to replace 'em just to avoid a breakdown somewhere and have to
limp home.

Good idea on the alternate types of caps--I couldn't find any other kind
back then that would fit in that tiny space. Probably can now, what with
everything getting smaller, so if y'all find any non-polarized cap that can
be squeezed into that space, should work just fine and last a long time and
you won't have to worry about installing 'em backwards :-)

Regarding the risk of heat, fire, and explosion...I was never able to get
more than about 15 milliamps out of that AC alternator (speed sensor) that
eventually feeds those caps. Considering that the signal gets fed through
the throttle position sensor and then through a full-wave bridge rectifier
and a series dropping resistor before being applied across the two caps in
question (with a series smoothing resistor between 'em) to smooth out the
last remaining pulses riding on that mostly speed-and-throttle dependant
0-20Volt or so DC signal, there ain' much left to produce heat. For the
record, that signal then gets applied to the input of both op-amp
comparators which compare it to a fixed Zener- reference voltage before, in
the classic Op-Amp Configured as a Comparator style, and at the appropriate
shift point, the comparator removes the bias from the base of the
appropriate TIP-42C pass transistor that then unsaturates, removing drive
current (unfiltered, direct from the 12 v supply, and without antispiking
diodes, can you believe it!!) from the associated shift solenoid, causing
the transmission to shift.

Whew! Sounds a bit Rube Goldbergish, eh?

Today, we'd just use a PIC microprocessor and it's internal A/D converters
to read vehicle speed and throttle position separately, calculate the shift
points, build in a little hysteresis loop in the program to make the upshift
and downshift points a couple three-or-five MPH apart (upshift/downshift
points are supposed to be different and we SURE don't want the classic
closed-loop servo control "chatter" at the shift points), and drive the
solenoids via a couple pass transistors from the PIC's parallel I/O ports.
Much simpler, truly a digital computer, could also have a programmable shift
point / performance curve via the PIC's on-board serial I/O port using your
laptop computer to set them wherever you want, and it would all fit in half
the space of the original board leaving plenty of room for whatever filter
caps and varistors would be needed to give a good, clean, safe DC supply.
And look entirely stock, to boot. Wouldn't be a bad idea at that-- and you
could modify the shift points when you drop in DMC's high-performance engine
to take advantage of it's added HP and desirable shift point changes. Been
thinking about building such a thing but I just haven't had the time. With
my record of late I probably couldn't get the darn thing to work anyway.

Anyway, I always figured that if those two caps ever shorted out from
natural failures or from installing them backwards it would just take the
speed/throttle position signal to ground through the two aforementioned
resistors. Those resistors would, if my Thevenin nodal current calculations
are correct, limit the maximum possible signal current (i.e., at full 114 or
so MPH vehicle speed in third gear) to about half a milliamp through one cap
and even less through the other. Wouldn't have guessed that a half a
milliamp could cause enough heat to make anything explode, but even if it
would, at sustained100+ MPH we might well deserve it. Shoot, just occurred
to me, I could even measure the short circuit current at that very point by
just placing a milliammeter in series with the throttle kickdown switch and
jam the accelerator pedal to the floor. If you have the car at full throttle
(regardless of speed), that's exactly what the kickdown microswitch is
doing-- it literally shorts out the vehicle speed and throttle position
signal completely at that very junction and causes the transmission to
downshift. In non-Engineering terms, the Governor no longer has a signal to
reference, so it just gives up, so to speak, and the car downshifts.

Guess that means that if either cap ever shorts out on it's own, the car
would limp home stuck in first gear. 'Course with that second resistor
between the two caps, depending on which cap shorted it might just limp back
shifting back and forth between first and second. Wouldn't want to drive it
that way for long, might make the whole transmission explode.

Now in all fairness, if those caps in question had anything to do with
filtering/smoothing the car's 12VDC supply voltage, in the manner that most
electrolytics are used in automobile electronics, and which a reader of the
years-long Governor Computer thread would very likely assume from all the
discussions about them failing from inductive kickback /counter EMF off of
jump starts and the like (I suspect you assumed this too, Ralf), then I'd
agree completely on the risk of shorts, heat and explosion from improperly
installed capacitors. I rather suspect that the very lack of such supply
filtering on the Governor's DC supply lines may well contribute the
Governors checkered performance reputation, but then on the other hand the
lack of such caps saves it from more consequential damage should they go and
decide to short out on their own, even without the misfortune of being
installed backwards. In my experience these guys tend to fail shorted when
used in DC supply filtering applications-- often entirely on their own
volition, without being provoked by things us humans do like installing the
car battery backwards, jump starting, overcranking, and other electrical
maladies. Those Siemens comparators seem remarkably tolerant of unfiltered
DC supplies. Spec sheets show a wide VCC / VDD range with excellent
differential balance characteristics, guess that's why. It's certainly why
I've never found a drop-in replacement for 'em, but luckily (and
unexpectedly) have never needed to since they don't seem to fail.

Protecting us all from the consequences of component failure is, I reckon,
one of the reasons they put fuses on radios and governor computers.

So, my apologies for omitting details like the governor's small-signal use
of these caps and indirectly causing undue alarm amongst fellow intrepid
Governor experimenters. And Ralf is completely correct that if you ever do
install one backwards you should discard it, not re-use it. Reverse voltage,
even at low currents, will usually damage them.

Farewell.

Mark Hershey
Sr.Director, Engineering
Advanced Product Planning and Development Group
Intecom, Inc.
Addison, TX.


----- Original Message -----
From: "Ralf Philipp" <doc.brown@xxxx>
To: <dmcnews@xxxxxxxxxxxxxxx>
Sent: Wednesday, February 07, 2001 11:46 AM
Subject: Re: [DML] Auto trans governor problems


> >
> > Possible you installed the caps backwards but the Tantalums would
probably
> > tolerate and work OK anyway, at least for awhile. Wouldn't likely be
> > intermittent. The black rubber seals were the positive end.
>
>
> No, no, no!!!
>
> If you install Tantalums backwards then they will definitely go bad. They
> WILL get hot, and they might even explode.
(snipped right about here, and so was this message :-)