My 1992 Olds 88 with 106K will randomly just stop running while driving and not re-start for a while. When it does, it will run a little rough and smell lean...especially when I try to rev it.

It appears to be a fuel pump or pressure regulator issue but there is one thing...

I recently changed out the evaporator core and one of the lines going to the vacuum tank was left unplugged and the end of it got burned off and looks to have melted the line shut. Anyone know what that vacuum tank and lines actually do? Could fixing it solve my problem, before I have to drop the tank or something? Could that be the cause of a fuel pump or regulator going bad?

Plugging the open hole in the vacuum tank with my finger doesn't appear to have any effect on how it runs, when it runs.

you might want to check the fuel pressure with a gauge before you start dropping the tank... checking the fuel pressure while it is not starting by just cranking it you will know if the fuel pump is providing enough pressure...

are you sure it is not running rich instead of lean? it sounds like an electrical problem with your ignition somewhere...

Yes you want to fix the vacuum line. It would be only pulling vacuum when the purge valve for the canister is operating.

Thank you both for your suggestions!

I will first fix the vacuum line, since it'* such a cheap and easy fix. I need to hit the salvage yard for a couple nuts and bolts anyway.

It definitely is a lean smell. It smells like my '84 Chevy truck, which runs lean due to the headers I put on it and the oxygenated (10% ethanol mandated) fuel I use and the lack of adjustable mixture screws to compensate. I also replaced the O2 sensor on my Olds 88 about 1K ago. I don't have a pressure gauge, but would use it if I did. I do hear the pump when I turn the key to run, even when it didn't want to catch.

A few shots of starting fluid shot into the TB would make it run for a split second, making me think the ignition is okay.

Update: The problem appears to be fixed. I spliced in the correct portion of hose from a salvage car. I drove it for about an hour on backroads to make sure it wouldn't stall while I'm driving to work (in Milwaukee at 3 a.m.) Added bonus: my a/c on max now comes out of the proper vents!

Thanks again for the advice! I'll post pics of my car soon.

Of course it waited to stall until I was off the freeway, but it did just that at about 3 a.m. yesterday morning in Milwaukee. My dad towed it to a repair shop and it finally did restart about a half hour later, so I did pull it into a parking spot under its own power.

So...
-it doesn't make a difference if the a/c is on or off
-only stalls (or doesn't restart) when the engine is warmed up
-May run and drive for an over an hour or 20 minutes... or 5 minutes (after the 1/2 hour or so wait after stalling a first time)
-I'll hear a sort of "ding" come from the instrument panel and then all the idiot lights come on when it stalls while driving

It looks like the guys at the shop are scratching their heads, since they haven't yet got it to stall and so far the fuel pressure appears to be good.

The only abnormalities I can think of now are:
-A washer/nut did fall behind the fuse/relay center under the hood (shorting?)
-On the under-dash relay/fuse area (passenger side) it appears as though one of the relays is dangling (but still secured to its wiring harness) ...found it while looking for fuel pump fuse
-Long term damage done by one of the vacuum hoses to the vacuum tank being disconnected with the end melted off?

-Chris

Intermittent stalling after the car is warmed up... hmmm. Sounds like a fuel problem to me, although I'm not an expert. Ignition problems tend to be all-of-a-sudden and won't-restart. But it is probably one or the other. I've had cars do this and it was a weak or clogged fuel pump. You have, essentially, three fuel filters. A fuel "sock" pressed onto the end of the fuel pump--the gas goes through this first before it gets into the fuel pump. A wire strainer on the fuel pump itself--this is built into the pump. And the fuel filter on the gas line between the gas tank and the engine. In one car (a Buick Regal) the gas in the tank was so dirty that it would clog the sock and the strainer. The fuel pump would get louder as it strained to pull gas through, but couldn't. Other car--the fuel pump was weak--but it had to warm up first, then would fail to pull enough fuel. Freeway driving won't necessarily show this flaw as you use less fuel and have less stress on the engine than around town or hilly side-road driving. I had another problem with the Buick--there was a connector back near the gas tank that connected the vehicle wiring (power) to the fuel pump. One of the connections had receded enough INSIDE the connecter that it would occasionally lose contact and the car would stall...
Anyway, that'* what I think. If it isn't fuel, then ignition. My suspicion would be the ignition module followed by the engine control module (computer).
Keith
1994 Pontiac Bonneville SSEi
currently suffering from transmission dislocation syndrome

Hi,
Just talked to my friend who is employed by one of the component suppliers for the auto companies and he was involved with the testing of the fuel pumps for these (and other) cars back when, and he seems to remember the testing would involve putting just dust in the tanks, which would be enough to clog that fuel "sock" pressed onto the end of the fuel pump-- the sock would collapse into itself and cavitate and not let the fuel get pumped...
And he mentioned specifically that this occurred when it was a "warmed up" condition, and using what would amount to the "summer blend" of gasoline...
(So rust or other type of sediment would surely clog this "sock" too...)

This was interesting to me as I am having these symptoms currently
(88K miles; fuel pump replaced 10K miles ago (Delco brand); 75°F outside temp)

I ran highway for 1hr then city for 45min and then it stalled... started after 15min; happened again (after car had set for 2hrs in sun) after driving city for 15min (in 15 secs did a rolling restart... lucky)
after 5min more stalled again, no start...after had cooled down (1 hr) starts fine...

Car sat overnight, starts fine now...

On all these events, after stall the electrical all appears to be OK (car cranks very quickly but doesn't "catch", just like there is no fuel)(sometimes will try to "catch" and fire/start, but then doesn't)

I should also mention that during the hwy run I had discovered the cruise control wasn't working, and switched it a few times to see if it would work - heard a strange little clunk from underhood area when I was doing this (cruise still doesn't work)... 1st stall happened 45min after this event...
The later stalls also happened within minutes of me trying switching the cruise control again... I'm fairly sure this is cruise/stall is unrelated but felt it should be mentioned)

Hope I can figure this out (need help and insight from you guys)

Thanks for any help and replies!

These cars are famous for crank sensors, ICM'*, and fuel pumps, if they don't start right away after a stall....another thing to try, is disconnect the MAF sensor, and see if it will start.....

When it won't start, that is when it is easiest to diagnose....

Have a can of carb cleaner handy......spray carb cleaner into the throttle body, and if the car starts and runs as long as you spray, then it'* fuel delivery....either injectors(no ref signal from ICM or bad ECM) or fuel pump......noid light and fuel gauge would determine this....also, you can try a "tap" test on the ECM....the red, white and blue ECM'* of the early 90'* were suspect....

If it doesn't start with carb cleaner, then you check for spark.......if no spark, then it could be the ICM or the crank sensor....more often than not, it'* the crank sensor......an old trick was to tap the crank sensor PEDESTAL with a long extension while the car is running.....if the car stalls or stumbles, that is your problem.......THIS IS A LITTLE TRICKY TO DO......don't try it if you don't have the confidence doing it, because you have a rotating serp belt near the pedestal......you have to hit the pedestal, NOT THE SENSOR!.....have the car shut off.....find the pedestal, and practice tapping it LIGHTLY with the end of the long extension.......now when you are confident enough, to not have that extension hit the moving serp belt, try it with the engine running......it'* really not that hard to do......found many, bad crank sensors with this procedure.......but you have to do it with the engine and sensor warm....

By the way, have you checked for codes? You don't need a code reader for this car, it will flash codes, by grounding A to B in the ALDL connector, with key on, engine off....

Hi, thanks for reply!
What I have since found doing tests in the driveway...
1. Idled the car till it got hot (plus was 85°F outside) and car stalled and shut down...
2. Tried immediate restart: result was same quick crank/no start after (almost like cranking without spark plugs installed - really fast)
3. Sprayed water on crankcase sensor location as well as underside of ICM (sprayed both for about 10 secs, light diffuse stream of water on both), then used my air compressor to blow all the severe moisture off and away...
4. Car started easily and correctly!

Tried this sequence two more times at home in driveway, only these times I isolated the CKPS by only dumping about a pint of water on it and not the ICM...
Car started as above!

Thought I would try it live, so grabbed 1 1/2 gallons of water, and drove to three locations and it didn't stall when driving, however at stops one and three when I had shut it off, went in and did my business, and come out to the car to start it, it wouldn't start... ("heat quench" phenomenon is responsible?)
But, a pint of water on the CKPS... it starts right away!

So, am I correct that I have isolated it to the CKPS with some certainty? (it sure acts like it)

Was later told by a shade-tree mechanic that after 90k miles and 24 years the magnets surely have gotten weak with age and the heat expands them apart just enough to be borderline on the magnetic field it needs to generate...sounds logical

Just wanting to be sure this method is actually analyzing the CKPS getting cooled off enough to work again...

Thanks for all your help and expertise!