Automatic Transmission Computer Theory

Date: Wed, 28 Jun 2000 22:58:07 -0500
From: "Mark Hershey" <dmcinfo(AT)thehersheys.com>
Subject: Auto Transmission Governor Secrets Revealed...(long post)

OK, maybe not everything we'd like to know. But as promised awhile back,
here's enough to explain some strange auto transmission symptoms and get started on the story in detail.

First, I'd like to thank Dave Santos and others who have contributed bits and pieces of knowledge, experiences, and spare parts to my What Makes the Governor Computer Tick and Why Most Eventually Fail project.

Over the years I have experienced multiple failures of the GC and have gained considerably more experience with them than I intended. Been able to repair my own, but experiences reported by other owners revealed a variety of failures, only some of which I saw on mine.

To determine what was going on I reverse engineered the mis-named "computer" and made my D into a test fixture for several possible circuit changes. I'm working on a comprehensive document but for now here's some of the results. I'll also report perodically on how well some changes work after they have more time and miles on them.

Here is some key info for the technically inclined among us. I'll try to get schematics and pictures up on a forthcoming Web page after I get 'em cleaned up.


Missing parts:
Good Engineering practice dictates some preventative measures that simply are not included in the original design. I'll describe a few examples, but first it may help to understand a bit about how the GC works.

The GC works by comparing a variable voltage derived from a small alternator attached to the transmission (and reduced by advancing the throttle position) to a constant 7-volt source derived from the battery with a voltage regulator device called a Zener diode. When the variable speed/throttle position voltage equals the 7 volt constant voltage, the first transmission control solenoid valve drops out and the transmission
shifts from 1st to 2nd. Similarly, when the variable voltage continues to climb to a specific point (about 9.5 volts), the second solenoid drops out and the transmission shifts to 3rd. All this depends on keeping the 7-volt reference at 7 volts no matter what-- and in a hostile electrical environment, that can be difficult. The cheap Zener diode circuit used in the GC is a typical design, and it appears adequate so far. However, Zeners can be forced out of their Zener regulation "knee" by large spikes and
transients. Though they rarely fail with permanent damage the circuit can behave erratically. Expected symptoms:if your transmission goes flakey (downshifts erraticlly, refuses to upshift) BUT you can immediately stop, shut off the ignition, restart the car, and everything is OK for days and weeks at a time this circuit is a likely culprit. I'm interested in feedback if that happens to any of you; if so I'll test a design with a better regulator. For now I'm leaving that as is.

The main "computer" devices are two integrated circuits technically referrred to as op amps configured as voltage comparators. These particular Siemens Corp. chips are long unavailable but rarely fail. The circuit they are used in is extremely sensitive; even condensation on the board will cause the circuit to function improperly. Holding your thumb on the board while driving around (don't try that one yourself)will even cause it to fail! This was the core of my research since the most common GC failure symptoms are caused by this circuit, that being the car randomly downshifts from third to second with no cause/effect pattern other than perhaps relative humidity. Reason for such instability: two things...First, there are two electrolytic capacitors ("caps") on the GC board that, when new, help filter out transient variations (electrical noise)from the car battery.,These caps nearly always fail, just time and transmission heat will kill 'em. Every single GC I've looked at (5 years or older)had bad caps. That in itself would be OK, except the comparators depend on a good, clean reference voltage and without these caps they are less likely to get it. The fix: replace the caps with more rugged parts (I used tantalum caps seven years ago, still OK)AND add .1 microfarad bypass caps across pins 1 and 4 (supply and ground) of each of the comparator chips. Yes, all you incredulous engineers out there, the Renault designers actually left out the IC bypass caps!Amazing...Anyway I added these last week so time will tell if they reduce instability long term. I can, at least, handle the board live on rainy days without downshifting :-)

Also for you engineers, they left out antispiking diodes in the solenoid driver circuit-- an invitation to intermittent / erratic operation. Add reverse biased 1n4007's (radio shack) across the transistor output leads.

The number one problem I saw on examination was poor solder joints! All seven units I have examined (two mine, one Dave's, and various Dallas friends over the years)had really bad solder job all over the board. Worst problem is where the 8-wire cable attached to the two boards and where the six-wire inter-board jumper cable solders on to both boards.Can't stress enough how removing the red conformal coating (I used lacquer thinner), generously fluxing the board, and resoldering everything can help. Certain it causes many of the intermittent problems.

Drive Transistors:
The two transistors that drive the solenoids are RCA units long out of production. Although they only fail ocassionally, they can be replaced with a more rugged TIP42 or TIP42C part from Motorola. TIP42Cs can still be found in some Radio Shack stores and are available from Radio Shack's online store. Symptoms:car starts out in second gear but shifts OK to third, or car starts out in 3rd and stays. Note that a blown A/T fuse can also cause it start and stay in 3rd. Generally these transistors don't fail
intermittently, so if your car sometimes shifts OK but gets screwy on occasion, this ain't it.

Multiple switch

The multiple or "combination" switch controls backup lights and 1st gear hold when you move the lever to "1". The switch rarely fails, and isn't repairable in any case. Just make sure the retainer screw that holds it to the transmission case is firmly in place. I'll offer more detail on this switch's complete function later.

Well, enough for now...I'll try to write up a more coherent description with pictures for download. Perhaps Mr. Zilla could be persuaded to offer TransZillas with various design improvements. I'd like to see someone replace the whole circuit board set with an A/D converter driving a PIC microprocessor and let you adjust shift performance curve via a Laptop serial port connection. Wouldn't be all that hard to do...

Bye!
\\ Mark Hershey
Vin 2790,now full-time automatic (so far)