Replaced the pcm on my 2003 Dodge Caravan exactly two years ago. We ordered it direct from Chrysler and paid $1,000 for the part and installation. Here I am two years later with a stalling van. The mechanic checked it out and said everything pointed to the pcm. So I ordered a refurbished pcm, flashed to my vin, from GoECM.com. Mechanic put it in and it did nothing. Sent it back, and they sent me another. This time the van started, and even drove about 10 miles before stalling. I have sent it back again, along with my core, and I’m waiting on the THIRD refurbished pcm. I have no hope that this one will work, but the company said they won’t give me my money back. The only thing I can do is keep sending these back. Should I have ordered from the dealer? Are these companies that sell refurbished computers legit?
You never know what you get over the internet.
When it comes to PCM’s I deal with local auto recyclers. They pull it off the shelf, make sure the bar code matches, and then it’s just a plug and play affair.
Tester
So how confident are you that a faulty PCM is the correct diagnosis? Swapping this many PCMs out could make that diagnosis suspect.
Maybe this stalling and/or no start is caused by a faulty ASD relay, fuel pump, etc, etc.
I agree with @OK4450 It is RARE for a PCM to fail on a car durring its life, forget about twice… A PCM is not a ware item, and yes sometimes they do fry like a home computer, but most of the time they are built so stout its not an issue. To be honest I have NEVER, EVER, NOT ONCE replaced a PCM on any of my cars. Am I lucky??? The lotto does not seem to think so 
I don’t disagree that the PCM is bad, but there is another issue CAUSING the PCMs to burn-out over and over. Either an over-voltage situation, or a crossed or shorted circuit causing these failures. This will not be solved just by replacing the PCM yet again. The underlying cause for the burn-outs has to be found and corrected.
Incidentally, this could be an expensive and time-consuming task. I do not envy you.
I am kind of reluctant to believe the pcm is bad, but there are not a lot of diagnostic tools out there for troubleshooting all the pcm circuits. You or your mechanic should start by inspecting all the connectors on the wiring harness for a pin that is bent or backed off. Then look for broken wires behind the connectors. Most wiring problems are in the connector or immediately behind them.
Next is to figure out exactly what circuits are not working and then follow the troubleshooting procedures for the wiring checks that are listed in the factory service manual. You need to have a mechanic that understands how to do this, not all can.
Totally different problem, but here in Mexico several years ago, my doctor friend was complaining about his Chevrolet. He said they did some work on it, and it overheats. The mechanic insisted the PCM was bad, so he bought another one. That went on several times and still overheated.
I asked if the radiator had ever been serviced. He said no. I told him no radiator was going to give service for that many years.
He took it to a radiator shop, the man ran a rod down through it, and the overheating stopped.
It seems once a mechanic decides there is a PCM problem, they can’t always let it go.
I guess this is the downside of the computerization of cars. Computers – when they work – work great. When they don’t work, it is very difficult – without having access to the engineering team who designed them in the first place – it is very difficult to diagnose and fix a bad one.
You’ve need a faulty computer circuit board replaced with the correct hardware version, along with the correct software version. It’s a nightmare.
If I had this problem I’d probably let the dealer’s mechanics solve it for me. They are trained on this exact problem. And to pay for it, live on macaroni and cheese for a couple months.
It could be that there’s a problem with your charging circuit.
Perhaps your alternator generates spikes that far exceed the acceptable supply voltage of the PCM. Those spikes are often very high in frequency so for a very brief interval - micro seconds or less - so the component may see several hundred volts when it isn’t really supposed to see more than a normal charging voltage of maybe 15V or so.
It doesn’t hurt these components immediately but does age them very quickly, causing premature failure.
You can’t see the spikes with a multimeter but could be measured with an oscilloscope by measuring what sort of junk is riding on top of the battery voltage. .Do you know anyone that has one and knows how to use it?
Obviously, you’d need to have the car running so need an operating PCM but it may be worth looking at, when you get your replacement PCM in.
I’m not saying the PCM ISN’T bad on your car, but like others have said, this component is usually the most reliable on any vehicle, and often becomes the “must be” component when a mechanic hasn’t properly diagnosed the actual problem or is scratching his head. There are a lot of other things that can cause the problem you’re seeing with your car, including the crankshaft sensor, which will not even cause the check engine light to come on in most cases when it’s bad.
I agree with @oblivion. They hardly ever break. What are the chances that three in a row are bad?
Either it has been misdiagnosed or something in that car is killing them.
@Tester and @RemcoW
The supplier of these PCMs Siemens/Continental Automotive had 28 different variations of PCM for the entire Chrysler platform when I quit working there in 2008. That PCM would run every vehicle they had from 4 to 8 cylinder in gas versions. I say this because a lot of variations were there. Tester has the correct plan by following the bar code to ensure he gets the correct one. Maybe the other 2/3 weren’t bad but from a reman it could very well be the wrong variant. I know this doesn’t help the OP now especially if the original is now gone and finding the “right” bar code becomes impossible.
Wow, crazy. You’d imagine they could have made that smarter somehow, maybe have that thing configure itself by knowing the VIN or perhaps reading some sort of external ROM that stays with the verhicle…
Typically don’t they just take one that will work and flash the firmware with the correct programming? Maybe not if it’s really old…
So bad wiring, faulty connectors, poor power, and bad grounds were the first go round on this car?
@RemcoW: Why would they bother having it configure itself? It might be possible, but it would add a lot of cost to the system. And it would mean that every PCM would have to have the programming for every vehicle stored, including possibly variations for low-emissions states. Since these rarely fail, it’s much easier to just install a compatible unit and update the programming, which can be downloaded in a minute from the manufacturer. At least it’s easy for a dealer… it doesn’t help the do-it-yourselfer or independent mechanics much…
@RemcoW, it is different operating ranges of sensors for different engines that throw all the auto config out the window. I think it may also have to do with possibly saving money on only installing the parts required for the particular platform instead of installing a lot of parts that are along for the ride. And last a FMEA (failure modes and effects analysis) may have shown that these optional loads could have also exhibited failures of the whole module if they failed even if not used. Engineering rationale I guess.
Sorry to add nothing to the OP issue.
@oblivion:
I know it isn’t very difficult or expensive to design in. All the car specific stuff could be stored on a chip that stays with the car. They have these devices called iButtons* that basically look like button batteries that you talk to by means of a single wire. They are made for rugged conditions and are programmable devices. You can store quite a bit of info on them. It isn’t even scary bleeding edge technology; I’ve been using them for at least 15 years.
So when you build the car, you burn whatever parameters like a feature list that may be specific to the car into this ibutton, slap it onto the car somewhere and give it some wires that hook into the ECM.
When one replaces the ECM, you power the car down - that’s a given.
The ECM, it loads those parameters, compares them to what it has stored and makes the decision to configure itself, if it needs to. That ECM could be ‘generic’ as far as a manufacturer is concerned - that depends on how well they ‘futurize’ their design.
I’ve done it several times - not for cars - but for for systems with similar issues when it comes to parts replacement. It adds very little to the bill of material, works well and allows you to change things out with impunity as the iButton is reliable and stays with the system.
*there are others out there as well besides the iButton.
The schematic shows the ASD relay is part of the Power Control Module and considering that a fair amount of electrical current flows through that module maybe the PCM is erratic due to heat breakdown.
When the same part is continually blamed and replaced then the diagnosis does become a bit suspect.
When our '02 Chrysler T&C van was fairly new we had a problem with the engine just quitting on us. Fortunately it would start up again. The shop at first though the trouble was with the injectors but after that proved to be the wrong fix they moved on to the PCM. That proved to be wrong also. They finally decided to replace the hardest thing which was the engine wiring harness and that cleared the trouble.
Since you have had so many, I too also doubt the trouble you are having is with the PCM itself. Perhaps it is with the power circuit to it or maybe the injectors circuit. Look for any clues to the trouble with any engine codes. Make sure the engine and chassis grounding is ok so electrical noise is kept to a minimum. You need a tech that is very sharp on electrical problems to be working on this; and that can be hard to come by at times.