Transmission will not shift

If I didn't mention it already, other than on diesels, you'll never solve a running problem on a Chrysler product by replacing the fuel filter on the frame rail. I don't know if that applies to the newer ones with the filter in the tank, but I suspect they did that, in part, because replacement really isn't necessary.

There's two things that can be done to try to find this problem. One is to attach a fuel pressure gauge and clip it to the radio antenna so you can watch the pressure when the problem occurs. I did that on my '88 Grand Caravan for a year and a half when chasing a real intermittent problem. I've never done this on a vehicle with the pressure regulator inside the gas tank, so I don't know what they will do, but with the regulator on the fuel rail on the engine, the pressure goes up during acceleration when intake vacuum goes down. That's to keep the difference in the two forces acting on the fuel at the injector tips equal at all times, and is why there's a vacuum hose attached to the regulator.

The second thing requires a scanner with record capabilities. You press the "record" button when the problem occurs. A few seconds of data is recorded that can be played back slowly later while watching the sensor readings the computer acts on, like throttle position sensor and coolant temperature sensor, and the results of those actions, like oxygen sensor readings. Simple fault code readers won't be able to do that. You'll likely have to find a mechanic with a scanner.

The first suspect is low fluid level due to a leak. Forcing the car to move by raising engine speed will seriously overheat and destroy the slipping clutch plates. That can turn a minor problem into a very expensive transmission rebuild. Slipping even once like that can do more wear to those clutch packs in a few seconds than normally takes places in over 100,000 of normal driving.

Actually, the fault detection system is real effective and is much easier to work with than on some other brands. The first part of it has to do with sensor voltages going out of range. For example, the throttle position sensor is fed with 5.0 volts and ground. Due to mechanical stops, the signal voltage will always be between 0.5 and 4.5 volts. The only way it can go to 0.0 volts or 5.0 volts is if there's a break in a wire, connection, or inside the sensor, and that's what triggers the fault code. If the broken connection is intermittent, the Engine Computer will detect a break that lasts for just a few milliseconds, while all scanners and especially all digital voltmeters respond much too slowly to show those.

The clinker here is when a sensor sends a wrong signal voltage but it's within the acceptable range. That isn't likely to happen to a throttle position sensor because it's a simple mechanical device, but a MAP sensor has a lot of circuitry inside it so it is very possible for it to send, ... Lets say 2.2 volts, which is within the acceptable range of 0.5 to 4.5 volts, when it is supposed to be 2.4 volts. That 0.2 volt error will result in a major miscalculation on the required amount of fuel, but it won't set a code because 2.2 volts is an acceptable value.

The next part of fault detection involves operating conditions that would be almost impossible to determine yourself by watching the readings on the scanner. A perfect example would be, "running too lean too long". In that case the oxygen sensor is obviously working as it is supposed to, but the exhaust gas is not switching between rich and lean fast enough. Normal is about twice per second, but you may see it stay in one state for much longer on a scanner. You can't tell exactly when this type of code will be set. One instance typically won't do it. The computer decides when the exhaust gas is staying in one state too much, and that can take hours or days.

The third part of fault detection was added with all '96 and newer vehicles. Now misfires can be detected by the slowing of the rotational speed of the crankshaft, (which is what we feel when we feel that misfire). The computer knows which cylinder is responsible, but it takes a certain number of misfires within a given amount of time to set a code. Along with this, the fuel supply system is monitored for leaks in the vapor recovery system. This is why the Check Engine light will turn on and a code will be set if you leave the gas cap loose.

The nice thing with Chrysler's system is when the Check Engine light turns on, there WILL be a fault code. Some cars, GMs in particular, use what they call "pending codes". Those relate to things the computer is watching in preparation for setting a code, but the problem isn't quite bad enough yet. Sometimes those pending codes can give you an idea of where to look, if you can read them, but often there is no code, ... Just the phantom Check Engine light that turns off before you get to the shop, and you have no idea why it came on.

Don't know if that will help or not, but it might explain why there is no code being set. The fuel supply system is not monitored as far as pressure is concerned, so you won't get a code related to that, but you could get a code related to the RESULT of low or high pressure. Those would be running too lean or too rich.

I can't argue with any expert on driveability issues because that's not one of my specialty areas, but I can share how the system is supposed to work. The Engine Computer varies the fuel / air mixture from too rich to too lean about twice per second. During the lean times, the extra unburned oxygen gets stored in the catalyst. During the rich times, the unburned fuel vapors mix with the oxygen and burn off resulting in basically carbon dioxide and water vapor as the exhaust. When the catalytic converter is working properly, the oxygen sensor after it will detect a slightly lean condition for a fairly long time, as in a minute or more, then a slightly rich condition for a long time. The computer watches how long it takes for that switching rate to occur. The "downstream" O2 sensor switches very slowly and the "upstream" one switches very quickly. That's how the computer knows the converter is doing its job.

Based on the voltage readings from the upstream O2 sensor, the computer adjusts the mixture so the average is a perfect 14.7 / 1. That is the only oxygen sensor involved in fuel metering. Even when it has a bad enough problem that it doesn't develop the proper signal voltage, it has very little effect on engine performance. It affects emissions and can affect fuel mileage, but usually the engine will run fine. The rear sensor has no affect on engine performance at all, and no affect on fuel metering calculations. Its only function is to monitor the efficiency of the catalytic converter.

Two things can happen to the converter. It can become plugged, usually from overheating caused by some other defect that resulted in too much raw fuel going into the exhaust system, or the catalyst stops working by becoming coated with a contaminant so it can't store oxygen. You can tell the severity of a diagnostic fault code by how the Check Engine light acts, and the most severe is when it's flashing. That means too much raw fuel is going into the exhaust and will overheat and melt the catalyst. Typically it turns into a big rock that is not porous enough for the exhaust gas to pass through freely. That usually results in easy starting, but then very low power, almost non-existent acceleration, and an unusual hiss from the tail pipe instead of the normal "putt-putt" sound. The engine often idles fine, and if you can drive the car, the engine overheats easily. This type of problem is not intermittent. The test for it involves drilling a small hole in the exhaust pipe, then inserting a pressure gauge to measure the back pressure. A little is normal. I don't know what is considered excessive, but it seems to me I heard once that two pounds is too much.

When the catalyst stops working, the exhaust can still flow freely through it so there is usually no running problem. Any stumbles or hesitations are likely due to the computer causing changes to the fuel metering calculations in an attempt to clean up the emissions. As the efficiency of the converter goes down, it has less affect on the composition of the exhaust gas so what comes out starts to resemble what went in. When no change at all takes place in the exhaust gas within the converter, what goes in and what comes out will be exactly the same, so the downstream oxygen sensor will switch between rich and lean at exactly the same rate as the upstream sensor. It takes some time for the downstream sensor to go from switching once per minute or so to switching twice per second like the upstream sensor. There is a magic number programmed in that must be met to set the fault code for "catalytic converter efficiency". It doesn't have to reach the same twice per second of the upstream sensor. If I had to guess, I'd suggest the code might be set when the downstream sensor starts switching at maybe once every five seconds for a brief spurt, or perhaps once every ten seconds steadily. Those are just estimates to illustrate how the computer determines when to set a code.

There's a very high percentage of fault codes that relate solely to things that could increase emissions while having no negative effect on engine performance. To affect performance, the converter would have to be plugged. I'm not convinced the symptoms you described would be caused by the converter, but I certainly won't argue against it. With such an expensive part, I'd want to see some evidence on a scanner to support that argument before I spent that much money.

Its most likely a shift solenoid but it really needs to be tested by a transmission shop.

I don't believe stuck clutch plates are the issue. To get a fault code related to the clutch pack staying applied means it was set by the pressure sensor for that clutch pack or hydraulic circuit. Clutch plates physically burned and welded together is a mechanical thing that is not monitored, so that wouldn't set a code.

A stuck clutch pack would cause other symptoms too. If you look at the truth table in the service manual, it will list each clutch pack and in which gear(s) it is applied. I don't know which ones are applied when, but as an example, lets say the under-drive clutch is applied in second and third gears but not in fourth gear. If that clutch pack was sticking, the transmission may go into second gear when the hydraulics are telling it to go into fourth gear. If yours is going into all gears properly, I would suspect a different type of problem.

There could be a problem with a sticking pressure switch telling the computer there's hydraulic pressure applying the clutch pack when there should be no pressure. Mechanically the transmission is fine, but electrically it's giving incorrect information to the computer. There could also be a problem with the hydraulic circuits. If the pressure is bleeding off too slowly, a clutch pack may appear to be applied for a fraction of a second too long and will overlap the application of another clutch pack. Excessive wear could take place during that fraction of a second when two clutch packs are fighting each other by trying to rotate at different speeds.

I'm not a transmission specialist. I can only go by what I learned at some Chrysler schools in the '90s, but this might give you some ideas of what to look for. I've heard some people say they've had good luck with "Sea Foam" products, but I've never used it, or anything else in a transmission for that matter.

At this point I can't add anything of value except to say the computer will detect an open circuit to a solenoid. The only time you'd need to do a continuity check is when you get a fault code that indicates a circuit needs further diagnosis.

Remember too that throttle position and road speed are what determines the transmission's shift points. If the converter is plugged, you'll have to push further on the accelerator pedal, and that will change how the shifts feel and when they occur.

That should be second gear it is stuck in, and the check engine light should be on. Second gear is "limp mode" that allows you to drive slowly to a repair shop without needing a tow truck. Your mechanic will start by reading and recording the diagnostic fault codes.

Dandy. Happy to hear it's solved. I'll try to remember this in case I run into it again.