1996 Dodge Caravan 3.0L 3spd. Van stalls every 1 out of 10 times when starting from a stop, hot or cold. Worse if you try to accelerate suddenly or under load. Will always restart immediately.

Also stalls when coasting (especially downhill), have my foot off the gas for a bit, then try to accelerate – the engine cuts out completely for about 3-4 seconds, then comes back to life. I can push the pedal right to the floor and nothing at all happens - I don't think the engine is even maintaining an idle in those 3-4 seconds.

Difficult to start, cold or hot – requires multiple attempts.

Idles well. Runs at highway speeds just fine, no missing, good acceleration. Multiple 3 hour trips at highway speeds without problem, even WOT – lots of power, doesn't miss a beat. No Codes.

Replaced:
Fuel filter – no change.
Air filter - no change
Throttle Position Sensor – new. No change.
MAP sensor – used. Small improvement.
PCV – new – no change

Inspected/cleaned:
Spark plugs - look OK
Idle Air Control valve – no change
Crank position sensor – no change
Cap & rotor - no change
PCM – batt disconnect – no change
EGR – blocked it off with a coin - helped a bit but threw a code (P0401) so unblocked it.
Fuel Injector cleaner - no change
Seafoam - no change

I couldn't check fuel pressure because the 3.0L doesn't have a test port, I'd need a line adapter (Miller 6539), which I don't have, but I'm kinda doubting if it's the fuel pump because of flawless highway performance and WOT.

The stall is pretty specific – accelerating from a stop or from coasting.

As the title says, I'm stumped.

 

Welcome to DF
Do you know if you have stored codes in the PCM? You may want to have the dealer hook up a co-pilot to record ant problems you are having.

 

Thanks Master Tech. No stored codes.

I've got an Elm 327 hooked up to my laptop - it's only got the free sotftware so I can't get a dynamic picture like graphing the sensors, but I do get a snapshot of the readings every 6-7 seconds. I'll try to take my son along to read the values as I stall it down a hill.

Sure feels like fuel when it stalls. Maybe its a flaky fuel pressure regulator?

 

You have a Vacuum gauge? compare the actual Vacuum to Scanner vacuum data.
I dont remember is 96 3.0l caravan has an Intake air temp sensor, but look into that too. But I believe it has an intake manifold temp sensor.
Also check coolant temp sensor data and compare to actual value, I've seen if data on these sensor is off, it could cause engine to flood with fuel.

I doubt is fuel pressure related, as you mentioned, it has flawless WOT performance, and these fuel pressure regualtor is not vacuum operated.

Also check if any of your acessory drive is locking up on you too.
You mentioned difficult to start, see if you get an crank signal and good spark when you try to start it. It is possible that one of the sensor is off and causing engine to be flooded with fuel.
Shoot some starting fluid in and see if it help starts as well.

 

I have a 96 Caravan also. I'm having the same issues. The code that I got was the camshaft sensor. I'll let you know what happens when I change it today.

 

Thanks for the suggestions. I hooked up the Elm scanner and drove the van around today - most of the time the van behaved itself (of course), but near the end it started acting up. Here is what I found before my laptop battery died ...

My LTFT is ~ -6
The temp seemed to rise fairly quickly - it switched over to closed loop about 45 seconds after startup. But then the temp hovered in the 150-160 range forever. By the end of the day it was breaking 180.
The one time I was able to catch it stalling with the scanner, it seemed to enter "Open Loop-Driving" and the fuel trims snapped back to 0. I think it entered Power Enhancement mode? The TPS values looked fine ~ 20%.
Each time it stalled and I restarted it, the LTFT got more negative: -6, -9, -12 ... not sure if that suggests it stalled from flooding, or it was just a symptom of stalling and nothing more.

So, this is what I'm thinking:
- The stalling only occurs when I go from a condition of low pressure (high vacuum) to higher pressure - like sudden starts or accelerating after coasting. I think this should rule out a lot of things like CPS and fuel pressure.
- the engine seems to be running a little rich normally - temp sensor, thermostat, or leaky injectors?
- looks like the PCM went into PE mode just before stalling - but I'm pretty sure the TPS wasn't responsible. Could a flaky MAP sensor do this?

I thought it was a fuel supply issue before, now I'm thinking it might be flooding when it stalls. Any way to confirm this?

Thxs

 

Try running the vehicle with Coolant temp sensor disconnected and see if it stalls. Yes, it will give CEL, but it rules it out.
45 second and it goes 160 degree? doesnt seem right. Grab the upper radiator hose with you hand and see if it feels like 160 degree.

If the car goes open loop suddenly, it could flood the engine and cause stall condition. Things that could cause PCM to go open loop would be O2 sensor and Coolant temp sensor.

You said u tried another MAP and same thing, right?
I doubt it would be CAM or Crank sensor, since u said it ALWAYS starts right up.

 

Have you looked at the speed sensor. If the pcm see's your speed above 120 mph the sensor will start triming the fuel.

 

Just wanted to update you guys:
I got a P0300 code (random misfire) last week – cleared it and it didn’t come back.
I removed the throttle body and cleaned it out – van idles a little better but still stalls as before.
I found a road with a nice downhill grade that I can make the van stall on fairly consistently. I hooked up my elm scanner and just kept driving down that road, stalling the van and watching one parameter at a time to see what changed. Here’s what I found:

TPS, MAP, Coolant, RPM and Speed sensor seem to work fine and send the expected signals before and during the stall. It stalls in both closed and open loop.

My O2 sensor gives a really low voltage (0.02V) starting about a second after the stall, so I think it’s in reaction to no combustion with only air coming thru. Consequently the STFT jumps up but nothing happens.

My timing advance (Cyl #1) was the most obvious wrong reading. Normally it will give a value between 20-35 degrees advance, depending on if I’m coasting or flooring it. But when the engine stalls, the advance drops to 8 degrees at about the same time and stays there (this happens to be the default value I get on my scanner when the key is in the on position but the engine isn’t running and my foot is on the gas).

I disconnected the coolant sensor as Steak59 suggested which threw the computer into an open loop (fault) with no STFT and a default LTFT. The van ran rougher and still stalled in the same way, with the advance dropping to 8 again during the stall.

So it seems to me that during the stall, the PCM is either shutting the engine down, or being told the engine is already shut down? Not sure which comes first. I would suspect the crank position sensor or a slipping tone ring, but I’m still getting an RPM reading. Also, for the most part this isn’t a random occurrence – it happens when going from high vacuum (idle or decel) to acceleration. And it restarts right away every time.

I also swapped the Auto Shutdown Relay with the Fuel pump relay, as they are identical, but it made no difference.

Could it still be the crank sensor or tone ring?
Maybe the Cam sensor or its tone ring?
Or could it be the PCM goes flaky only under certain conditions?
Is the Auto Shutdown Relay kicking in for some reason?

Still stumped.

 

Could be a bad flex plate. Its not uncommon to come across a cracked flex plate.