Misfires on cylinders 1-4 with SD tune... out of ideas.

[tta1456]

Have a weird one that I've been trying to diagnose for a couple of days and still can't figure out.

Engine is a stock 2003 5.3 with a 2001 12200411 ECM with an HPtuners speed density tune in an offroad rig that a friend just purchased. Harness is a factory harness that was modified to be a retrofit installation. Transmission is a 4L80E with a standalone controller (why they didn't use the ECM to control the transmission, I don't know...)

Noticed when I test drove the thing it was down on power, sounded a bit funny, accelerated slightly rough, and had a bad raw fuel smell. Free revved it from idle with my hand on the engine, and it's obviously misfiring by both feel and sound.

First thing I noticed is both exhaust manifold flange gaskets were leaking... replaced both of those and no more leaks.

Plugged in my Tech2 and looked for codes... no codes. Nothing counting on the Tech2 misfire counts screen either.

Checked all harness to engine grounds as a first step, all connected and clean.

Thought it might be a bad coil since I just dealt with that issue on a friend's 2002 truck that had 2 dead coils. Went around and disconnected the coil wires one at a time while the engine was running. Idle did not change when pulling 1, 2, 3, and 4, but idle did noticeably drop when pulling 5, 6, 7, or 8. Double checked with a timing light-- it didn't flash for 1, 2, 3, or 4. Cylinders 5, 6, 7, and 8 triggered the timing light just fine.

First step was to swap coils and plug wires... I moved the coils on both banks so the ones that weren't firing were now in the spots of the ones that were firing, and vice versa. Still no spark on 1,2,3,4, but spark on 5,6,7,8 according to the timing light. I've seen a whole bank be dead before for a wiring issue, but only the front 2 being dead on each bank? This is odd...

Checked all 8 plug wires and plugs... all appeared near new with 0.040" gap, no cracked porcelains or carbon tracking. Checked compression while I was there, all 8 were running about 160psi. The engine isn't making any noises that would indicate a lifter going bad although I haven't pulled the valve covers and cranked it to check.

Next up was checking for power and ground at all 8 coil connectors... all were good.

Next I disconnected the connector at the ECM and checked all 8 signal wires and all 8 signal ground reference wires between the coil connector and ECM connectors... all good continuity and pinned correctly with no shorts to ground.

Just to see, I pulled the starter and crank sensor and looked at the reluctor wheel... it looked fine. Note-- I did not try and move the reluctor wheel with a screwdriver to see if it was loose.

Checked cam sensor counts with my Tech2 while idling... it appeared to count normally.

Used my oscilloscope and probed between the signal and signal ground reference wires at all 8 coil connectors while the engine was running. All 8 coil connectors, including cylinders 1-4 that the timing light didn't trigger on, show a nice 5V square wave coming from the ECM... that's odd. I expected to see no ECM signal on 1, 2, 3, 4 since those coils don't appear to be firing according to the timing light. Since they all had power and ground I was leaning towards bad coil drivers in the ECM.

At this point I'm questioning the timing light, so decided to do a "direct" test of the coils instead of using the timing light and dug out one of the spark testers that take the place of the spark plug... turns out cylinders 1, 2, 3, and 4 *are* sparking but not triggering the timing light. However, those cylinders are still misfiring. If you pull the coil or injector wires for cylinders 1, 2, 3, 4 while the engine is at idle the rpm or tone doesn't change. If you pull 5, 6, 7, or 8 the RPM drops substantially.

Did a crank relearn, no changes.

Pulled the vacuum line for the fuel pressure regulator, it's not ruptured and dumping fuel into the intake. Fuel pressure is good.

Did an injector balance test with the Tech2, all 8 injectors flowed nearly identical, so it doesn't appear to be bad injector drivers in the ECM or stuck/clogged injectors... although I suppose the front 4 injectors could be lazy at idle/short pulsewidths and not firing consistently at idle and causing lean misfires. I may try swapping the injectors for 1-4 with 5-8 tomorrow and see if the misfires follow the injectors.

A thermal camera shows the exhaust manifolds near the ports for cylinders 1 thru 4 are much cooler than 5 thru 8, so that backs up the behavior noted above when coil or injector connectors are pulled at idle... 1-4 appear to be either not firing at all or only occasionally firing.

There is quite a bit of a "whistle" noise at idle that got me to think about a cracked intake or leaking intake gaskets causing lean misfires, but it may just be the IAC making noise back through the aluminum intake tube. IAC counts are about 120 at a 650 rpm idle. Decided to spray ether around each intake port where it meets the head hoping to find an intake gasket leak... didn't hear any change in RPM or smoothness.


I found it slightly odd the engine isn't throwing any misfire counts so I downloaded a stock 2001 5.3 tune from the tune repository and did a file comparison, and it appears the SD tune that's on this thing has all stock misfire parameters.

Thoughts? I'm a couple days into messing with this thing now and no closer to a solution, so I figured I would ask for input. If there's any specific logs that would help, let me know and I'll get them for you.


I have attached a copy of the tune that was on the ECM as the vehicle was purchased. I have never messed with an SD tune before, so any insight would be appreciated.

Below are things I did notice in the tune that have me scratching my head, although remember I've never messed with an SD tune before...

P0101, 102, and 103 are set to no mil light. Based on some reading here about SD tunes, shouldn't an SD tune have those 3 codes set to "mil on first error" so the ECM "fails" to SD mode immediately?

Fuel stoich AFR is set to 14.22. This is a stock engine with a stock cam... why not leave it at 14.68? Fuel mileage isn't a goal of this rig, but I don't see a good reason to run it richer than necessary.




Finally... this is a stock engine and chances are it's going to stay that way. Any reason not to add an LS3 card style MAF and revert back to a MAF tune? I understand why the outfit that supplied the engine and harness combo provides SD tunes-- an SD tune doesn't care about the air intake and they don't have to deal with tweaking MAF tables for the end user's specific air intake. But once the misfire issue is sorted out, any compelling reason to either stick with the SD tune or add a MAF and go back to a MAF tune?

Thanks in advance!

[Lakegoat]

Yes--P0101,102,103 shpuld be mil on first error. I would not change to a maf. You have to get it the right size tube and the correct length straight section to get it to run right. I have been Sd for years and don't notice any difference. On your tune, I would modify the PE. What is your wideband saying at idle--or running? Maybe it is running pig rich? Does it have O2 sensors? In SD, you have to get the VE table spot on. 14.2 is about right for todays pump gas for stoich.

[tta1456]

One of the thing on my list of things to do today is weld in a bung after the Y pipe for the wideband and see what it says.

Based on the appearance of the spark plugs I'm going with lean misfires as they look very clean and not sooted up / wet / rich... but I could be wrong.

The tune and vehicle does have primary O2 sensors. Secondaries have been deleted since there are no cats.

I did look at the STFT while it was idling, and they were around +5% to +7%... could be the tune, could be a vacuum leak. At 650rpm idle the MAP was reading about 55 kPa, which seems a bit low for a stock cam IMO, but since multiple cylinders are missing at idle that may explain the lower than expected vacuum. I'm also going to double check the MAF with a vacuum pump and gauge today.

Just for some more info, the engine/computer/harness was purchased as a package a few years ago by the previous owner of the vehicle who did the LS swap. The vendor dyno tunes and creates the SD tunes for the engines before they leave their shop. I called the engine vendor on Friday to see if they had any input, and they're as confused as I am. Their initial guess was a damaged harness, but as of yet I have not been able to find any physical damage to the harness or internal damage/shorts with my meter. They did offer to send over both the dyno sheet and video of this exact engine/harness/ECM combo running on the dyno to show it was running correctly when it left their shop.

[Jeff7577]

Use a sealed headlight or something similar to load test the power/ground circuit on 1-4. I don't think it is a coincidence that the 4 cylinders that are missing won't trigger a timing light pickup. How did they modify the harness? Each bank has its own ground connection stock but they could have wired it differently. Even a poor ground can show continuity with a test light or ohm meter but fail under a load.

[2xLS1]

Misfire diagnostics are disabled in SD mode on P01 PCMs.

[tta1456]

Use a sealed headlight or something similar to load test the power/ground circuit on 1-4. I don't think it is a coincidence that the 4 cylinders that are missing won't trigger a timing light pickup. How did they modify the harness? Each bank has its own ground connection stock but they could have wired it differently. Even a poor ground can show continuity with a test light or ohm meter but fail under a load.

That's also on my list to check today. A wire with broken strands or a poor crimp will certainly show continuity with a meter but not pass enough current to operate what's connected to it.

I agree it's probably not a coincidence that the timing light is triggering on the 4 good cylinders but not triggering on the 4 misfiring cylinders even though I'm seeing spark on the 4 misfiring coils with the spark tester.

The harness coil grounds are still there at the back of the valley cover & heads. I confirmed with the shop that built the harness that all grounds are present and connected.

Misfire diagnostics are disabled in SD mode on P01 PCMs.

Disappointing as I'm down a diagnostic tool with the SD tune, but great info to know. Thanks!

[Jeff7577]

Do you have an adjustable spark tester like this? https://www.lislecorp.com/specialty-...g-spark-tester

You can do a quick test to see if those 4 won't produce the same voltage.

[2xLS1]

Is it SD with a stock OS or a custom OS?

[Jeff7577]

Was this engine out of a 1500? G104 is at the back of the drivers side head and G102 is on the drivers side block under the exhaust manifold, which is also where the negative battery cable attaches.

[tta1456]

Do you have an adjustable spark tester like this? https://www.lislecorp.com/specialty-...g-spark-tester

You can do a quick test to see if those 4 won't produce the same voltage.

Yes, my tester has an adjustable gap.

Is it SD with a stock OS or a custom OS?

Not 100% sure... I attached a copy of the tune to the first post.

[2xLS1]

I can't look at it. What is the OS number?

[tta1456]

Do you have an adjustable spark tester like this? https://www.lislecorp.com/specialty-...g-spark-tester

You can do a quick test to see if those 4 won't produce the same voltage.

OK, just tried opening the tester to 30KV (~0.650" gap) and tried it on cylinders 1 and 3. No problems with either coil throwing consistent spark across that gap. And with the tester connected and the gap opened that large, the timing light triggered consistently. So the timing light isn't triggering when they're connected to the spark plug... hmmm. Bad plugs and only in the front 4 holes? Seems odd. Engine has factory manifolds and great access so unless whoever put them in was a gorilla they shouldn't be cracked. Plugs looked extremely good and near-new when I pulled them out 2 days ago and I didn't see any cracks, but that doesn't mean it's not 4 bad plugs. Both plug 1 and 3 show 5 kohm resistance (factory delco iridium plugs.) Something else amiss in the cylinder? Compression on 1-4 was in family with 5-8.

I can't look at it. What is the OS number?

OS # is 12208322

The shop that put together the engine/harness/ECM/tune package says this was one of their earlier engines done about 7.5 years ago.

[2xLS1]

That is a stock 01 GM OS. There are advantages to running a HPT 1 bar SD COS if you are going to leave it in SD such as having both low and high octane timing tables available. I'm not sure on this but misfire detection may be available using the SD COS. Maybe someone else could confirm.

[tta1456]

OK, I feel like an idiot now for dismissing the "little things" as a possible cause.

After I saw the timing light working with the spark tester in place but not with the plug wire connected to the plugs in the cylinder head, I grabbed an old, somewhat fouled NGK TR55 plug out of the garbage, cleaned it off and put it in #1 in place of the new looking Delco iridium plug in there.

Guess what... timing light fires for #1 now.

Plug in injector #1... RPM picks up and smooths out... was it really this simple?

Decide to try the same in #2... tossed in another old, used, somewhat fouled TR55 plug in place of the Delco iridium. Same result... Timing light starts working and RPM picks up when the injector is plugged back in.

Wasted 3 days chasing the "big stuff" when it was probably 4 bad plugs this entire time. I guess I won't let a "good looking" plug deceive me again.

I thought I had moved plugs at the same time I moved the coils but I must have put them back in the same cylinders... Oops.

Started to swap plugs in cylinders 3 and 4 also for the used/fouled TR55s as a test to see if it would fire on all 8, and the terminal on the #3 wire decided to stay attached to the plug but pulled out of the crimp on the wire... guess I'll get new plugs AND wires today at lunch.

I'll report back once I get the new plugs and wires in but I expect this thing will be running perfectly...

Note to self: don't dismiss the simple stuff!

[Lakegoat]

Don't feel bad. I hauled a modified '14 silverado to the dealer because it wouldn't start. Fouled plugs because it idled too long on a bad tune. Plugs looked good visually----400 bucks later

[tta1456]

Update...

New Delco plugs and Delco wires totally fixed the issue. Fuel trims are now about -7 to -8% at idle, engine idles and accelerates extremely smooth, no more raw fuel smell, timing light works on all 8 plugs, thermal camera of the manifolds looks nice and even. Idle vacuum is way better at about 36 kPa now that it's running on all 8 and the IAC can close down much further to maintain idle RPM, compared to how open it was trying to maintain idle RPM on 4 cylinders.

Didn't take it for a test drive tonight, but will tomorrow. I don't expect any issues.

Just for fun I took the 4 bad plugs, hooked them to one of the coils and grounded the body, and tested them to see what was going on. They didn't spark *anywhere*. No spark jumping down the plug wire boot to the hex on the body, no spark jumping from the electrode right down to the insulator (as they sometimes do if excessive use of anti seize contaminates the insulator.) Just no spark visible whatsoever. I have never seen plugs fail like this before, it's really odd. All 8 plugs (the 4 good and 4 defective) all measure 5 kohm resistance, but maybe there's a defect inside the plug body that's left just enough gap so the multimeter doesn't show a short, but it's a small enough gap that the coil voltage has no problems jumping it and it's also a small enough jump that coil current is so low it won't trigger the timing light.

If I had to guess, it was a box of 4 plugs that were all defective that just happened to get installed in cylinders 1, 2, 3, and 4.


Now that the misfire issue is solved, back to the tune... see any changes that would be worth making besides setting the 3 MAF DTC codes to "mil on first error"?

Lakegoat, you mentioned modifying the PE, so I just looked at the PE tables. I'm assuming that you're seeing that the current tune with the stoich AFR of 14.22 combined with the current PE EQ ratios is probably way too rich? At 4800 the tune is commading a 1.336 PE, which with the current stoich AFR of 14.22 is commanding a 10.64 AFR. That's rich even for boost IMO. Even at 5200+ past peak torque the tune is commanding 1.301 PE which is 10.93 AFR. Even if I changed the stoich AFR to 14.7, the PE would still be 11.0 at 4800 and 11.3 at 5200+. I'll need to log with my wideband and see exactly what it's doing and how good the current VE table is tracking to actual commanded AFR. Around 12.7ish at WOT sounds much more reasonable to me, maybe a tiny bit richer at peak torque. Comparing the current tune with a 2001 5.3 factory tune I pulled off the tune repository it appears the PE EQ ratio tables are identical.

Did some reading on the custom 1 bar SD OS... looks like you gain an extra VE multiplier table, dual spark maps work again (high and low octane), and RTT. Doesn't mention if the misfire counter will work with the new OS, but since they don't mention it... probably not. It seems the P01 ECM's automatically disable misfire detection when MAF codes are set. Be nice if it did work again though.

Thanks for all the input.

[Lakegoat]

Here is what I would do to the PE

[tta1456]

Your suggested PE table is right about what I was expecting to see and was thinking of changing it to... I just need to weld a bung in the exhaust for the wideband for datalogging.

Is your suggested PE of 1.15 keeping with the current 14.22 stoich ratio, or is 1.15 suggestion if I were to change it back to the stock of 14.68?

At 14.22 a 1.15 PE is 12.37:1

At 14.68 a 1.15 PE is 12.76:1

[Lakegoat]

Keep it at 14.2. I did that to keep it safe. After you get it tuned and see that everything is OK, you can always lean it out a little to 12.5

[tta1456]

Understood.

Doing some more reading, I see the reason for the 14.22 stoich AFR is that most fuel today is E10. Makes perfect sense and is very obvious now, I just hadn't thought about it before.

Also 100% agreed to start on the rich side then work leaner... It usually gets expensive taking the reverse approach.