Sparsely intermittent stalling under acceleration

[longbedder]

This may get a little long-winded, but here goes...'94 K2500 Burb w/ 454/4L80E

I've made a couple other posts related to my problem, but that was before I found the root cause. I finally figured out that my fuel pump is occasionally losing power when I accelerate. It's only happened maybe 8 times in the past year, so it's very hard to replicate.

It only happens when I'm cruising and punch it - then it backfires and stalls, and either catches again after 4-5 seconds or I have to pull over and crank to get it going again. Not cool on the DC beltway.

It happened for the first time since I got an ALDL cable and a new battery (hence no previously stored codes), and I was able to get Code 54 out of it. This indicates a loss of electrical power to the sense lead of the fuel pump. Potential causes are as follows (in my mind - correct me if I'm wrong):
1) Intermittently bad fuel pump
2) Intermittently bad fuel pump relay
3) Intermittently bad oil pressure sending unit/switch (my truck has both in one unit - three wires)
4) Intermittent wires somewhere
5) Shaky ground somewhere
6) Bad ECM

My basic quesion now is this: should the fuel pump relay be energized at all times when the key is on, or does it de-energize after the oil pressure switch takes over? My thinking is that it would be very unlikely that both the pressure switch and the relay would drop out at the same time. If the relay does normally de-energize after a bit, a faulty oil pressure switch alone could cause my problem.

Any thoughts/experience?

 

[dyeager535]

I'm 99% sure that the ECM keeps the fuel pump relay energized while the vehicle is running, but the oil pressure switch is doing the same thing, so it doesn't matter as long as one works.

With that said, I'd certainly not spend time on the oil pressure switch, relay or ECM, other than verifying that your oil pressure switch works. You should be able to disconnect the fuel pump relay while it's running without the engine dying, if the pressure switch is working.

Only way to probably try to find your problem is to check continuity, and wiggle all the wires you can reach while the tester is on. With it being so intermittent, unless it's an obvious problem like a chafed wire not in a loom, you are probably going to have a very hard time finding it. You could drop the tank of course, and see what the wiring looks like inside failing to find anything external, but I'm guessing that will be a fair amount of work. Then again, with the problem being this serious, it may be worth it.

Why do you need to pull over? Can you not put it immediately into neutral and crank it? Just making sure there is not another problem here that doesn't allow that...

 

[longbedder]

Why do you need to pull over? Can you not put it immediately into neutral and crank it? Just making sure there is not another problem here that doesn't allow that...

That's what I've done successfully the last couple times. When the problem first started, I just pulled over right away thinking the worst. It takes a bit of cranking (5-6 seconds), and then runs rough for 3-4 seconds while the fuel fills the lines again. Dependent on the road and traffic situation (some of the worst in the country here), pulling over is sometimes still the best option.

I agree that this is a beotch to diagnose. I've looked it up on thirdgen.org, and Code 54 is rather infamous (especially when intermittent).

I am tempted to build a simple $20 aftermarket circuit as shown below, using a power lead straight from the battery and an aftermarket (generic use) relay. The output juice from the aftermarket relay would go into the factory relay connector via a spade terminal. If the problem then happened again, it would be narrowed to 1) the pump, 2) the hot lead going to the tank, or 3) the pump ground. If I drove it for 4 months and nothing happened, it would be either the ECM, pressure switch, OEM relay, or the harness wiring.

 

[dyeager535]

Well, when the fuel pump relay dies, it DOES take oil pressure time to come up enough to close the switch before fuel starts flowing. It's either that, or your pump is going dry. (which could be a leaky something in-tank, but I would expect that to be more consistent, unless it only happens when fuel gets real low or something, and you don't typically let the gas get that low)

Hopefully someone can comment on the fuel pressure regulator, but in any case, your lines shouldn't go dry that I can imagine. The pump can't push air, and the injectors don't suck fuel, so if the pump loses power, the engine should stop running immediately, and there should still be fuel in the lines. It starts up quickly first thing in the morning though, assuming you let the fuel pump prime before trying to start?

 

[longbedder]

Yup - starts up quick in the morning, and I get the nominal two-second pump noise when I turn the key...although with a problem as infrequent as this I'm not surprised.

 

[dyeager535]

I'd have to think that for all the effort, you might as well just replace the factory wiring, running it the same way, since you'll have to be in it anyways.

I would still probably tend to focus on some of the hard parts. Are you sure the fuel pump is losing power, or are you losing ignition by any chance?

 

[longbedder]

I'd have to think that for all the effort, you might as well just replace the factory wiring

It would only take about 1 hour of work and $20 worth of wire, connectors, and a relay to get that circuit up and running. The factory wiring is all harnessed and not that easily replacable (in my mind, anyway).

Are you sure the fuel pump is losing power, or are you losing ignition by any chance?

The only data I have to go on from the ALDL is a code 54 (fuel pump relay fail). For the ignition to cause the problem and this code to show up, the following would all have to happen:
1) Ignition flakes out
2) Engine dies (oil pressure drops to zero)
3) Fuel pump relay is de-energized
4) Fuel pump turns off while ECM stays on, tripping code 54 in ECM memory

 

[bigblock72]

My truck does the exact same thing. It's actually a problem with the electronics, namely the computer can't recognize when there is boost. I've watched the BLMs with my scan tool when climbing a grade at part throttle and less than 1 lb of boost, the fuel pressure starts to rise causing a rich condition so the computer compensates by leaning (BLM around 100) and then I lean on it a little harder and the boost goes up to 4 lbs and boom; lean backfire. I've messed with mine trying to compensate for the problem but it only makes it worse. I think the only way to truly correct the issure is change the computer to one that uses a 2 bar MAP sensor such as a 7749, a Megasuirt or other aftermarket computers. Of course that opens up a whole new can of worms with transmission controls and stuff so I just decided to leave it be.

 

[dyeager535]

It would only take about 1 hour of work and $20 worth of wire, connectors, and a relay to get that circuit up and running. The factory wiring is all harnessed and not that easily replacable (in my mind, anyway).

*IMO* you are better off fixing the factory stuff. I'll leave it at that, I won't get into my reasoning. It's mainly personal prefernce.

The only data I have to go on from the ALDL is a code 54 (fuel pump relay fail). For the ignition to cause the problem and this code to show up, the following would all have to happen:
1) Ignition flakes out
2) Engine dies (oil pressure drops to zero)
3) Fuel pump relay is de-energized
4) Fuel pump turns off while ECM stays on, tripping code 54 in ECM memory

So lets work through that:

1)module hiccups (just making that up as the ignition cause...could be coil wire from fuse panel, etc)
2)engine dies because there is no spark
3)fuel pump relay is de-energized because the spark signal to the ECM is lost
4)fuel pump turns off because relay is de-energized, and the ECM is still on because you didn't turn the key.

Heck, it could be something bad on the injector circuit for all we know.

 

[longbedder]

you are better off fixing the factory stuff

I wholeheartedly agree. The setup I mentioned would be a long-term troubleshooting rig, so to speak. This is because the problem is very elusive, and it'll go a very long time without happening.

My truck does the exact same thing. It's actually a problem with the electronics, namely the computer can't recognize when there is boost. I've watched the BLMs with my scan tool when climbing a grade at part throttle and less than 1 lb of boost, the fuel pressure starts to rise causing a rich condition so the computer compensates by leaning (BLM around 100) and then I lean on it a little harder and the boost goes up to 4 lbs and boom; lean backfire. I've messed with mine trying to compensate for the problem but it only makes it worse. I think the only way to truly correct the issure is change the computer to one that uses a 2 bar MAP sensor such as a 7749, a Megasuirt or other aftermarket computers. Of course that opens up a whole new can of worms with transmission controls and stuff so I just decided to leave it be.

So, when it backfires, do you subsequently stall? Is your problem predictably repeatable? Mine seemed to come outta nowhere and is inconsistent. It also didn't happen at all when I was doing some heavy towing cross-country, through mountain passes, etc (right after buying the truck).

On a side note - where did you tap in for your boost guage?

I GOTTA get this thing to misbehave while I'm collecting data
...and a garage
...and time

 

[dyeager535]

With a problem like that you really want to fix it. Sometimes it's just easier waiting until something starts malfunctioning more often to figure it out, but in this case that's a real annoying problem even if it's just once in a great while. Especially since it just stops and starts. At least if it didn't re-start, you'd be able to spend the time checking what's not working right.

 

[bigblock72]

The fuel pump relay can cause a code 54 but won't keep the truck from running. The oil pressure sender also sends power to the pump when there is oil pressure. You can remove the fuel pump relay altogether and it will still run assuming the sender is still good.

My truck doesn't die as long as yours does, it will restart within 3-4 senconds. It isn't very consistent but it seems it does it do it under lingering load more often. Although it has done it just cruising down the road once or twice.

To connect a boost gauge:There is a pipe plug in the intake on the driver's side just below where the throttle body would have bolted, I think it's 1/4" pipe thread.