Engine intermittently stalls when braking

Any chance this is all related to idle speed being too low? I could go down the list of parts and rule out most of them. Right now it's disconnecting the battery that is causing or adding to the low idle speed problem. Doing so erases all the memory in the Engine Computer, including any diagnostic fault codes. Al the fuel trim data and sensor characteristics will be relearned automatically without you even noticing, except for "minimum throttle". Until that is relearned, the computer won't know when it must be in control of idle speed. The engine may not start and run unless you hold the accelerator pedal down 1/4", it will not give you the nice "idle flare-up" to 1500 rpm at start-up, and it will tend to stall at stop signs. To meet the conditions for the relearn to take place, drive at highway speed with the engine warmed up, then coast for at least seven seconds without touching the pedals.

Let me know what happens after the relearn.

You originally asked about a lock-up torque converter problem. The only common problem with those was with early GM front-wheel-drive cars. They would stick engaged and cause stalling when slowing down, similar to not pushing the clutch pedal with a manual transmission. The "fix" for those was to unplug the two-wire connector so the lock-up function never occurred.

In your case, being below 35 mph, and being in reverse means the converter hasn't locked up, so it can't be sticking on. There are other things that can happen inside the transmission that might cause symptoms similar to what you described, but there should be other problems then when you're driving.

Chrysler's Engine Computers of this era caused very little trouble.

Lock-up torque converters were invented by Chrysler, as were most of the innovations from the '50s through the '90s that truly benefited car owners. The first ones used in 1977 with big block engines were controlled by valves in the valve body of the transmission. No electronics involved. The newer ones, like yours, are switched on and off by the Engine Computer that turns on an electric solenoid that opens a valve to let fluid flow to the lock-up clutch. That will only occur once the engine is warmed up, above approximately 30 - 35 mph, and only in third or fourth gear. Also, the computer unlocks it when the throttle approaches fully-closed, in anticipation of coming to a stop, and at anything over roughly 3/4 throttle. That is because when not locked up, torque converters cause engine torque to be doubled compared to the clutch on a manual transmission. Double the torque at high throttle is what allows you to pull out and pass with blazing speed.

The last and best clue to verify the lock-up function is working properly is to hold the accelerator pedal and road speed perfectly steady while you're cruising at highway speed, then lightly tap the brake pedal with your left foot. If you have a tach in the instrument cluster, you'll see engine speed pick up 200 rpm for two or three seconds, then it will drop down again. This is also done in anticipation of slowing down to a stop.

Since tapping the brake pedal causes the converter to unlock, a circuit from the brake light switch is an input to the Engine Computer. That same circuit is used for the "cancel" signal for the cruise control. This is a second part of the brake light switch, and is not the part used for the brake lights. The reason for this is one switch turns on to turn on the brake lights. If those contacts become burned or arced, you won't have brake lights. While that is undesirable, it would be a safety problem with the cruise control. If that switch was defective, the cruise control would not kick out when you need to slow down when overtaking another car. To address that, the second part of the brake light switch is turned on and provides 12 volts to that terminal on the Engine Computer. When you press the brake pedal, that 12 volts drops to 0 volts. That is what tells the computer to cancel the cruise control. This way, if that set of switch contacts were arced or burned, or the wire was cut, the cruise control would never set, and there would be no safety concern of it not kicking out.

That switch and circuit that cancels the cruise control is the same circuit that tells the computer to unlock the torque converter. What this means for us is we can use the operation of the cruise control as a diagnostic tool when there's a problem with the lock-up converter. If the cruise control sets normally, and it kicks out when tapping the brake pedal, you know that switch and entire part of the circuit are working.

Leaving the computer, the cruise control servo solenoids and the torque converter clutch are on their own dedicated circuits. Also, the electronic switching inside the computer is done by different circuits, so if one system works properly but the other one doesn't, the place to start looking is at the computer and the circuits leaving it. The one circuit coming in from the brake light switch can't be the problem if one of those systems is working properly.

In the rare event the converter seems to be sticking when you come to a stop, that is highly unlikely to be caused by a shorted switching transistor inside the computer. Most commonly when a high-power transistor like that shorts, it stays shorted. That means it would try to energize the lock-up clutch as soon as you turned on the ignition switch. I've never heard of that happening, so I'm not sure what the symptom would sound like. For sure it would snub the engine off as soon as you tried to shift out of "park".

It would be more likely the lock-up clutch would stick mechanically due to something coming apart inside the converter, but that would surely result in a miserable vibration. If you think crossed or shorted wires are keeping the converter locked up, you may be able to verify that by unplugging the three-wire connector on the left side of the transmission.

You can also tell if the converter is locked up by observing the tach when you accelerate slowly, as in from 55 to 60 mph. If the converter is locked up, engine speed will increase exactly the same as road speed increases. If it's not locked up, engine speed will jump up right away, then it will take some time for road speed to catch up. This is easiest to see if you do it before the engine has warmed up. In my case, when leaving my house on a hot summer day, I can drive about two miles on the highway before lock-up occurs. On a cold winter day, it takes over six miles before the coolant temperature is high enough for lock-up to occur. When that lock-up finally occurs, it is not unheard of for some drivers to mistake that as up-shifting to a higher gear. The delay is designed in because when the transmission is cold, the fluid doesn't flow well to the cooler tank in the radiator. Higher engine speed equals higher transmission front pump speed, and that helps push the fluid through the cooler tubes. Once everything is warmed up, the fluid will flow much easier, so the lower front pump speed is not a concern. If lock-up occurred right away, the relatively low 5 - 10 pounds of fluid pressure might not be sufficient to push the thick fluid out of the cooler lines. The fear is the hot fluid inside the transmission won't get a chance to go through the cooler, which would result in overheating the fluid.

If all else fails, you can use a scanner to see when the computer is requesting lock-up. I have a Chrysler DRB2 and a DRB3 for all of my vehicles. They show the "torque converter clutch request" as "on" or "off". You can watch it switch to "off" when you tap the brake pedal, close the throttle, or slow down below the minimum speed. The idea is to watch for the lock-up request and the lock-up you feel not agreeing. I've never run into that. The problem that's usually being diagnosed is a failure of the converter to stay locked up. The clues to that are found under the cruise control diagnostic menu. It will show the state of the brake light switch, throttle position, road speed, and coolant temperature. You have to figure out which of those isn't correct. Most commonly it's caused by a brake light switch that is out-of-adjustment, or it's so close, that the pedal vibrates when driving over bumpy roads. That's the biggest cause of the cruise control kicking out intermittently. The torque converter will unlock and re-lock intermittently at the same time. The typical complaint is engine speed is surging, and people think the transmission is slipping.

You're confusing two totally different systems. As you mentioned, what you're doing with the overdrive switch is giving you misleading observations. That switch prevents the transmission from shifting into fourth gear, mainly for pulling trailers or a heavy load. That entire system is not in the picture yet until you hit around 30 - 40 mph and the engine is warmed up.

The torque converter lock-up clutch locks the engine's crankshaft solidly to the transmission's input shaft just like occurs with a manual transmission. That makes for better fuel mileage and it lowers engine speed by about 200 rpm because it eliminates the inherent loss of transmitted power due to the fluid coupling inside the torque converter. Torque converters also provide double the torque over a manual transmission and clutch, which is why they momentarily unlock the torque converter clutch at high throttle / hard acceleration. Unlocked gives you more power for acceleration; locked gives you more fuel mileage and lower engine noise.

You don't have any control over the lock-up clutch like you do with the overdrive switch. The only thing you can do to force the torque converter to unlock is to tap the brake pedal or press the accelerator more than 3/4 of the way to the floor.

Observe if engine idle speed is normal, and watch if engine speed jumps up to around 1500 rpm for a few seconds when you start it. If neither of those occur, "minimum throttle" hasn't been relearned and we'll know what is causing the other problems. If you do get that nice "idle flare-up" to 1500 rpm, the next step would be to connect a scanner to see what is happening when the stalling occurs. Most scanners have a "record" feature that allows you to take a snapshot of specific sensor readings when the problem occurs, then you can play that back later to see what happened or what changed.

The torque converter clutch also is not in the picture yet until the engine is warmed up, you're in third or fourth gear, and above the minimum speed, typically around 35 - 40 mph. When you have a stalling problem when shifting into one direction, either forward or reverse, at the age of your truck, a better suspect is a wiring harness with the insulation rubbed off some of the wires due to sliding back and forth on the body sheet metal. In particular, look for the large harness that runs along the top of the left inner fender and under the battery tray. You may be able to make the problem occur by tugging on that harness while the engine is idling. Also look for poor contact between mating terminals in the connectors in the harnesses that go from the body to the engine. Often you can make a problem show up by flexing those connectors.

Most computers today do in fact have non-volatile memories, but that wasn't the case before 2002 when GM developed Engine Computers in some of their trucks that had to be programmed to the specific vehicle. Since at least the 1950s, Chrysler has been the world leader in innovations that benefit their customers. GM has been one of the world's leading innovators in things that benefit GM after their customers buy the vehicles.

For '92 models, when you disconnect the battery, the Engine Computer loses all of its learned memory, but it reverts back to what was pre-programmed in at the factory for the engine size it was built for. Those default values are close enough to make the engine run right, so you won't even notice the relearning that starts as soon as you start driving again. This applies mainly to two things. The first is sensor characteristics. No two sensors are ever exactly the same, so there are strategies the computer uses to learn how to interpret the signals it receives from those sensors. The computer starts out with typical values programmed in, then, over time, it adjusts to the slight differences it sees from each sensor. As one example, it knows if the engine has been off for at least six hours, the coolant temperature sensor and the intake air temperature sensor had better be reporting the same temperature.

Other brands of vehicles also use a mass air flow sensor as their main fuel metering calculation contributor. Only Chrysler has never needed a mass air flow sensor. Their main reading comes from the MAP sensor.

The second thing is "fuel trim" numbers. Every Engine Computer leaves the factory programmed with a set of "lock-up tables" it uses to calculate fuel needs. For every combination of MAP sensor readings, (engine load), intake air temperature, coolant temperature, engine speed, throttle position, and throttle direction and rate of change, there is a value in that table corresponding to what the engineers have calculated is the perfect amount of fuel needed at that instant. Those values would make any engine run well, but they can be improved upon to reduce emissions. Once the coolant has reached a specific temperature, the system switches into "closed loop", meaning it adds the readings from the front oxygen sensor(s) to the equation. All the other sensors tell the computer the amount of fuel calculated as needed. Based on the O2 sensor readings, that calculation is fine-tuned, or tweaked slightly. That is done continually as you drive. Those numbers are called the "short-term fuel-trim", (STFT) numbers, and they can be read or watched on a scanner.

When the computer sees it is making the same adjustments each time you drive, it gradually shifts those short-term numbers to the "long-term fuel trim", (LTFT) numbers. From then on, it's those LTFT numbers it uses as the starting point on the next drive cycle. For example, if under a certain set of conditions the computer sees it has to add fuel, indicated by a positive fuel trim number, it will increase the long-term number slightly positive. With that as the new learned reference, it won't have to make such big adjustments to the short-term numbers.

All of this learning up to this point takes place without you even being aware of it, unless you watch the fuel trim numbers on a scanner on a test-drive. The one notable exception is "minimum throttle". The computer needs to learn when it must be in control of idle speed. More specifically, it needs to know when your foot is off the accelerator pedal.

Most of the sensors that provide a voltage signal are fed with 5.0 volts and ground, then, for explanation purposes, the acceptable range of signal voltage is 0.5 to 4.5 volts. Anything outside that range is what gets detected as a defective condition and sets diagnostic fault codes. In actual practice, if you were to install five different throttle position sensors, you'd find no two of them show the same voltage at idle. One might be 0.47 volts; one might be.72 volts, and the others could be somewhere in between. If your old sensor showed.53 volts at idle, for example, then you installed a new sensor that showed.64 volts, the computer would see that as you have your foot on the accelerator pedal and you're holding it down just a little. The computer would stop trying to hold idle speed steady.

The computer needs a strategy to know your foot is off the accelerator pedal, then it takes the reading from the TPS and puts that in memory. From then on, any time it sees that same value, it knows it is time for it to be in control of idle speed. Part of that strategy is it needs to see very high manifold vacuum. That occurs when an engine is slowing down from a higher speed, with closed throttle. You could achieve that high vacuum by snapping the throttle from under the hood, but that doesn't necessarily mean you let it snap all the way back to closed throttle. That's why it wants to see that high vacuum for at least seven seconds. The only way to achieve that is to drive the vehicle, then coast with your foot off the accelerator pedal. At the same time the computer will only memorize that signal voltage if it remains perfectly steady. If it flickers or varies just a little, that's an indication you still have your foot on the pedal.

There are mechanical stops inside the throttle position sensor, so once minimum throttle is learned, the only way the signal voltage at idle can ever go lower is if a different sensor is installed. At that point, as soon as you turn on the ignition switch, the computer will see the new lower voltage, and it will put that in memory. No driving and coasting is needed this time. It's when you put in a new sensor that shows a higher voltage at closed throttle that you'll need to do the coasting procedure so the computer will learn that value. For most people, coasting for seven seconds is typical of half of a highway off-ramp, so they don't even realize they're doing the procedure.

When the battery is replaced or even just disconnected for some other service, that learned sensor characteristics and fuel trim data is lost. A conscientious mechanic will perform a test-drive to do this relearn, or at least inform the car owner of the need to do so.

I shared this great and wondrous information to help you understand how the computers work, but this probably isn't going to help with the problem you're having. From everything you described, this sounds very much like a problem I had with a '78 model. It took two transmission rebuilds two weeks apart before the specialist figured out the lock-up clutch was coming apart and the debris from the fiber plate was circulating with the fluid and tearing up the clutch seals. The final gasp was when the transmission would only go into reverse, and I backed up over ten miles until the engine overheated. When the problem first started to show up, it felt like the teeth in the rear differential were broken and skipping. The driveshaft looked like it was jumping when a helper shifted into gear. I never heard of anyone else having the same problem, and since then, other than the common problem with GM front-wheel-drive cars from the late '80s to mid '90s, lock-up clutches have had a pretty boring history with very few failures.

There was a service bulletin many years ago that had something to do with a check valve in the over-drive clutch drum. I thought it caused over-drive to not work at all, or it stuck on when down-shifting. I'm going to do some research to see if I can find that bulletin.

The bulletin I just mentioned is not listed for a Dakota, but I did find this one. Unfortunately they didn't list the symptoms this problem causes. There are some other service bulletins for your transmission, but those that might have applied referred to problems caused by incorrect parts installation, which we know isn't the problem you're having. One has to do with the over-running clutch being installed backward, and the other has to do with installing straight and waved snap rings into a clutch drum in the wrong order. That one resulted in no over-drive.

CARADIODOC is one of our best! Please let us know what happens.