3800 swap stumbling under load

[Quaraxkad]

I posted this elsewhere a few days ago, but the w-body and 3800 relevant forums all seem to be slowing down so much I figured I'd post it here too...

I've got an L67 swapped '92 Cutlass. Donor engine, transmission, and PCM are all dropped out of an 04 Impala SS. All stock at this point, no mods except a custom downpipe and intake for the swap. It's recently started exhibiting a new problem that I can't diagnose. Under load, the engine seems to "stumble". I'm sure there's proper terminology for it, but I don't know what it is so I'm calling it stumble! I did find a video on youtube of a car with a very similar problem (not a 3800, just a seemingly identical symptom). In my case, when it starts to stumble if I slowly let off the gas, at a certain point it upshifts and then runs smoothly again.

Additional info for people who really like reading:

My first idea was the fuel pump. It's only got a few thousand miles on it, but it was the cheapest one the local parts store had which I only bought because the stock pump wasn't pumping at all after sitting untouched 8 years in the garage pre-3800 swap. Testing fuel pressure at idle or while revving in Neutral probably won't tell me anything, because it's smooth throughout the entire RPM range in Neutral. And I certainly don't want to drop the tank and replace it "just in case", because there's no access panel behind the back seats which means I'll have to drop the tank again and that's no simple job on this car, especially when I can't even confirm it's the cause.

Searching around on Google, the most common responses to people that sound like they have a similar issue is a clogged or damaged catalytic converter. So just today I took out the cat and rigged up a resonator that I just happened to have around in the garage, left the rest of the exhaust open on the off change there was a clog somewhere else in the pipes or mufflers. The cat looked to be in good shape, I can still see through the honeycomb. I drove down the street and had exactly the same problem, so that at least rules out excessive exhaust backpressure I guess!

I do have hpTuners, and have two logs of drives where the problem occured. Though I didn't realize it until after, the road speed wasn't being logged, but nearly every other parameter is. I added almost all the parameters listed in the software and went for a trip to the gas station and back, just to see which values hpTuners was able to read and which were just duplicates or useless. So there's a lot more here than is reasonably needed. It does confirm 0 misfires though.

 

Here's a screenshot of a 30s span where the problem occured. The green line in the middle graph is the TPS, the red line on the top graph is RPM. If road speed had been logged here it would show [I]very[/I] slow/no acceleration during this point. You can see here even with additional throttle, RPM stays relatively unchanged, and when I fully let off it upshifts. Does the MAF or MAP sensor line look correct (second chart, purple and blue lines)? I would have expected both to rise with a more linear relationship to the TPS. [EDIT: Of course not, dummy! There's no added airflow because the RPM remained the same! Disregard that question.] I've only just begun learning about tuning, so interpereting these logs as a diagnostic tool is beyond my abilities. If it's as simple as a bad MAF I'm going to be so disappointed in myself for not replacing it already...

And because your first question will probably be "CEL?": Yes, there are codes stored. Most of them just need to be disabled in the tune because they don't apply in this car. When the problem persisted, I reflashed the PCM back to the original code and tune to make sure I didn't mess something up, and naturally all the previously disabled CEL's came back.

• P0036 - O2 heater circuit (The Impala O2's are grounded through the PCM for PWM control of the heater circuit. My wiring harness from a '97 Lumina 3800 grounds them to the chassis instead. I [I]should[/I] reroute the O2 grounds to the PCM but I may be fine just disabling the codes in the tune.)
• P0130 - O2 circuit (Neither sensor appears to be working as seen the logs reporting default values, I have yet to figure out why. I'm hoping it's just because of the P0036 code being set it's ignoring them. Both sensors have been replaced with new Bosch units, but I definitely need to also recheck the connector pinouts because I've found conflicting information and watch to tripple check that it matches up with my engine harness)
• P0449+P0452 - EVAP codes (All related components [I]are[/I] installed and I believe wired properly, this will take further diagnosis)
• P0463 - Fuel level sensor (Cutlass doesn't have one, need to disabled code in tune)
• P0520 - Oil pressure sensor (This one I'm seeing for the first time so I definitely need to look into it)
• P0530 - AC pressure sensor (No AC components have been installed yet except the compressor, so obviously the AC doesn't work!)
• U1000 - Class 2 comm failure (Cutlass doesn't have all the extra modules the Impala has, need to disable code in tune)
   

I think I'm out of things to add.. So... Any ideas? Any more information I can provide? Any tests I should do?

Edited April 12, 2022 by Quaraxkad

[55trucker]

I've not had the opportunity to use HPTuners.....here's what I see from just regular scanner use.

If the engine temp is a 208 why is the o2 stuck at .45v?  the program is giving off a reading that would suggest the fuel system is not closed loop, you're open loop at 200 degrees??

your long term on bank 1 is at 0?, there's only a single o2 sensor for both banks so there will not be a 2nd bank reading

your timing is at 12 degrees, your TPS is at 62%,

the injector pulse width is identical for both banks, the engine is not closed loop, the only way that both banks would be identical is if the fuel system is running off the default tables.

...lol, and according the the vehicle vss you're not moving at all

[Quaraxkad]

It is indeed in open loop, and will not go into closed loop, presumably because of the O2 CELs. And if if it's staying in closed loop, the fuel trims will remain 0.

The VSS wasn't being logged, that was entirely my mistake. Somehow I just scrolled past it while adding logged parameters to the list. I have since trimmed down the list to only the stuff that's relevant and works. Future logs will be more coherent, but I figured this one was good enough for this example.

[55trucker]

Might want to rethink your harness & stay with what the PCM desires to run properly.

[Quaraxkad]

I also had a spare 3800 throttlebody on the shelf, I took the MAF out of it and swapped it in, no change there. Can't say for sure *that* MAF was any good though. Also running without the MAF plugged in, same result.

And because I felt stupid skipping an easy test that I thought wouldn't reveal anything (and didn't want it to confirm a bad pump), I tested the fuel pressure. Priming gets the pressure up to 50psi, which drops to 46 when the pump stops. With the engine running, it's stable at 40psi, and 48psi with the vacuum hose removed from the regulator. Remains stable at 40psi under high rpm in neutral. Drops to 38 when turning off the engine then climbs back to 45 and does not leak down. The 04 Impala service manual states pressure after priming should be 53-59. So I'm 7-13psi low on priming. Running pressure should be no more than a 3-10psi drop, so worst case scenario it should be at least 43-49 (best case being 50-56) when running, so I'm at least 3-9psi low, and at most 10-16psi low! This seems like a problem! Even without the regulator attached for full pressure I'm still below the minimum spec for the Impala. I was really hoping it wasn't the fuel pump, I just refilled the tank and 15 gallons is a heavy tank to drop!

So it sounds like I'm due for a new pump... I'll probably go with a higher flow aftermarket now so I'll be set for future mods.

 

Unrelated side note, I've now got the resonator hooked up properly in place of the cat, and it sounds SO smooth. I can finally hear the supercharger whine. The exhaust was previously not hooked up properly due to two different pipe sizes from the downpipe to the cat inlet that I never properly attached so there was a small exhaust leak that made a much bigger difference in sound than I thought it would.

 

46 minutes ago, 55trucker said:

Might want to rethink your harness & stay with what the PCM desires to run properly.

I do absolutely plan on running new ground lines from the O2s to the PCM, but I think there's more important things to tackle first, assuming of course that's got nothing to do with this problem.

[Quaraxkad]

Well... problem at least partially solved! Instead of jumping straight to a fuel pump replacement, I decided to change the filter. It's not any older than the pump, but it's cheaper and a super easy job. So I unhooked the filter inlet. And found this:

Clean, clear fuel dripped out of both lines and the filter outlet, but all that gunk came out of the filter inlet. Yuck. My intention was to blow in both filters to see if I noticed a significant difference in resistance, but once I popped off the quick disconnect from the old filter I was stupidly unprepared to catch the leaking fuel and just quickly popped the new Wix filter on the line. I re-checked the fuel pressure. NO change. All results were *exactly* the same as they were before. Went for a test drive. NO STUMBLING! So that seems to solve the immediate problem. But I do still have low fuel pressure, and possibly a fuel tank full of gunk that needs to be washed out. I'm wondering if there's a tank or just fuel pump unit from another car that has a fuel tank pressure sensor that I could wire up to my PCM. The Impala expects one, Cutlass doesn't have one.

After further investigation, I checked my 92 Cutlass service manual. Fuel pressure specifications are 41-47 at prime, I got 46, within normal range. Engine idling pressure should again be a 3-10 drop, my result was a drop of 6psi, within normal range. So it looks like my stock-style fuel pump just may not up to the task of providing the pressure needed for the 3800. Nearly all information I read concering a L67 swap suggested that while swapping in a Walbro 255lph pump was recommended, they all also say your stock pump would be fine until you start modding. Seems I have stumbled across yet another piece of information about this swap that every. single. guide. got wrong.

[Bake82]

Glad it was an easy fix to get rid of the stumbling issue!

If you're not having any other issues now, technically the stock pump works if stock, but it's certainly not ideal!  I'd wait until you're out of fuel and drop the tank to replace with 255!  if you have never removed the pump before, I'd recommend getting a new OEM tank seal, just ot have on hand shall you need it.  the aftermarket ones are junk.  

[55trucker]

All 3800SC engines require 60+ lbs pressure to satisfy the fuel system when under full boost. Going to the Walbro is the right direction.

 

Is that 04 PCM looking for the active feedback pressure signal from the fuel system?

Edited April 13, 2022 by 55trucker

[pwmin]

I put an AEM 340 in my Subaru.  Higher flow and my tuner said they were better than Walbro.  They're about $15 more.  Nothing wrong with the Walbro, though.

[Quaraxkad]

27 minutes ago, pwmin said:

I put an AEM 340 in my Subaru.  Higher flow and my tuner said they were better than Walbro.  They're about $15 more.  Nothing wrong with the Walbro, though

People do say the Walbros have a tendency to whine loudly. I've also seen recommendations for the TRE340 fuel pump, roughly the same price as a Walbro 255 and quieter.

 

56 minutes ago, 55trucker said:

Is that 04 PCM looking for the active feedback pressure signal from the fuel system?

From what I've seen in the Impala service manual, the fuel tank pressure sensor is just for EVAP controls.

 

2 hours ago, Bake82 said:

if you have never removed the pump before, I'd recommend getting a new OEM tank seal, just ot have on hand shall you need it.  the aftermarket ones are junk

I have removed it already, and the replacement seal that came with the new pump was indeed junk! Will definitely get an OEM one next time.

[Quaraxkad]

Dropped the tank today. Found some pretty nasty stuff. The seal had definitely slipped out of place. The strainer was so clogged it basically imploded under the force of the pump. There's less than 4000 miles on this pump and strainer.

The gas in the tank looks awful cloudy!

The entire inner surface of the tank is covered in rust.

I'm replacing everything. New fuel filler neck, tank, sending unit, AEM pump, all hoses. I'll also flush out the fuel lines and maybe pull the fuel rails as well. Maybe bench-test the injectors to see if they're clogged up too.

And then to make fuel jobs easier in the future, I'm considering cutting an access panel behind the rear seat. I picked up the cover plate from another w-body at the junkyard years ago, it looks like a pretty good fit. Obviously the picture is just to check the size, that panel itself will be screwed in from inside!

Has anybody done that before? Is it a terrible idea? I don't know how I'll even cut it. I have a drill to get started, but don't really have any kind of cutting tools for metal.

[Schurkey]

On 5/13/2022 at 11:07 PM, Quaraxkad said:

And then to make fuel jobs easier in the future, I'm considering cutting an access panel behind the rear seat. I picked up the cover plate from another w-body at the junkyard years ago, it looks like a pretty good fit. Obviously the picture is just to check the size, that panel itself will be screwed in from inside!

Has anybody done that before? Is it a terrible idea? I don't know how I'll even cut it. I have a drill to get started, but don't really have any kind of cutting tools for metal.

Make VERY sure that the hole you cut gets reinforced, and sealed.

"I" would absolutely NOT cut an access hole.   The last thing I want is to be trapped in an upside-down car after a collision that ruptured the gas tank, with gasoline pouring into the passenger compartment through some kludged access hole that isn't liquid-tight.

Bad enough to be unconscious as you burn alive, but worse would be to be conscious, paralyzed, and having to watch the family burn, too.

[Quaraxkad]

After further thought I decided against cutting an access panel. The design of this sending unit has hard lines that wrap down around the tank and I think the hole would have to be huge to be of any use and I don't want that.