Unsolvable random misfire code.
 

2004 MSM that I had for sale in another thread until a missfire code could not be solved. The car had been modified with a FM downpipe, wideband, begi intake and FMIC and passed emissions testing fine. Added a megaquirt, injectors and boost control for lapping days which ran well. Reverted back to the identical modified configuration after two years and got the missfire code. Tired to fix it with the simple stuff and gave up, turning it over to the professionals and now we are at the place of scrapping it or going wire by wire in the whole harness to try and find the issue.

Below is what has been done by a pro-race shop and if any of the members that work there chime in you will know they are not hacks (like me). My question for the community is this: can you think of anything that was missed short of a wire by wire search that should be done before I do something rash with the car?

"We've put it all back together and at this point I don't know if there is anything else that we can look at.

The list of things we have checked/replaced to verify proper function:
-Boost/vacuum leak tested for and verified with a smoke test.
-Leakdown and compression were fine.
-MAP sensors
-MAF sensor
-Boost solenoid
-Plugs/Coils
-Cam and Cranks sensors
-ECU swap
-Mechanical and Electric timing
-Fuel Pressure Verified
-Injectors cleaned and flow tested
-Catalytic converter inspected for clogs
-Primary O2 sensor replaced which did fix the fuel trims because the old sensor was not responsive
-Throttle position sensor and Idle Air Control Valve
-Bypassed the jumper harness to the ECU and plugged it straight in.

All the parts mentioned above have been verified to be working. The car will drive 5 miles just fine, and then randomly it will come back. It could be that there is a wiring harness issue somewhere in the dash or in the engine bay. We've spent roughly 35 hours of diagnostic time on this but everything seems to check out. At this point I can get the car back to you if you'd like or we can start pulling the wiring harness apart and inspecting it for damage."

Thank you.

Harmonic damper? Wonder if the crank timing wheel gets squiffy.

Not sure what squiffy means, but we did check the wheel. No obvious signs of damage, but it was never replaced.

do you actually feel the missfire? What is the exact code you are getting?

P0300? I scratched my head a long time battling a random misfire on my 99.

Things I tried:
New secondhand coilpack
New NGK wires
New NGK plugs
New Crank Position Sensor
New Cam Position Sensor

What ended up being the ticket was the CPS.

From what I read, these seem to go and cause some gremlins but not necessarily flat out stop working.

On this note, if you're having a hard time actually feeling the misfire and your CPS hasn't been flat out replaced, I'd replace it. It's cheap, like 20$ cheap and for the headache I went through I'd say try it.

Thanks for that feedback. Is that CPS you are referring to the Crank of Cam?

Cam position, sorry for the lack of clarity!

On the msm if the boost solenoid can cause all kinds of problems. I periodically had to take mine out and spray it with wd40 because it would randomly get stuck.I would get some random issues from that.

We replaced both cam and crank sensors, no effect. We fixed a lot of little things along the way, at this point it wouldn't be a horrible idea to try the cam sensor again as easy as they are to replace.

We did clean, eliminate, and replace the boost valve.

And yes it's a P0300, after about 5-10 miles/minutes of driving. We've gone as long as ~20 minutes without the code. Never feel anything. Little boggy at lower RPMs but pulls great above ~4000. Typical MSM from what I've read.

Hmmm. I never distinctly felt a misfire either, but I figured out a rhythm for when the code would come up. I would drive on the highway no problem and get off an exit, drive at 25-35 city speeds and it would trip after a couple minutes.

It seems like this P0300 is coming up almost entirely randomly?

If you aren't able to feel the misfire at all, I'd lean towards some flukey wiring or sensor failure and that the car isn't actually misfiring. I am by no means versed with MSM's in particular though.

I would start scoping things until you find the problem. Sensors and outputs until you find what is wrong.

Squiffy - slightly drunk.

P0300 is basically the same as sync loss on megasquirt. The way misfires are detected - very simplified - the last 4 crank sensor passes happened 2.2ms, 2.1ms, 2.0ms, and 1.9ms apart (RPM is increasing). We can expect another pass in 1.8ms - when that doesn't happen a misfire is assumed. If the occurrence happens often enough to pass a threshold for a certain cylinder, a P0301-4 code is thrown for the cylinder. That usually means some badness there - injector or valves. You've already done most of the things I would have tried and can't feel/hear a misfire, but my next thing would be to pull the pulley, inspect the keyway, and take the timing wheel off to see if the mounting holes oblonged at all. In a similar line of inquiry - check that the timing belt idler pulley is tight - slack in the timing belt could throw off the intake cam signal.

Could the timing wheel be on backwards?

I like that as a possibility.


Maybe the distance or positioning of the Crank Position Sensor? Odds are the car wouldn't even start then, but potentially worth double checking.

Been done.

I don't think it would run if this was the case, and see above.

If the 2004 MSM has the connectors hanging on the front of the engine by the throttle body, open the connectors and inspect for corrosion.

They are super clean and I think all NBs have those same connectors?

Yeah, the car overall feels great...as great as it can feel with the stock, laggy-ass, ECU.

Clean all the grounds, wire wheel and scrubby pad, a loose/finnicky ground could make this happen. Also, battery connectors. I would get sync losses when my battery terminals were loose/corroded. Also try a different MSM ecu.

Yeah, they did the whole ECU/immobalizer/lock swap and the same issue.

Really, I think it is some finicky-ass wire some place and more than not likely related to the timing wheel.

I wonder how hard it would be to trace and replace the entire line for the cam and crank angle sensor.

Its buried in the harness pretty good.

Maybe check resistance between the ECU and the sensors. Might tell you if there is a partial break somewhere.

Do you still have a megasquirt? Can you take a log of it happening? Might be able to catch a sync loss.

I'm sure you dug through M-S about this (if not, you should) it seems to be a relatively common issue with various culprits. Have y'all checked the egr valve?

Don't just focus on the timing sensors, it very well may not even be related to that system at all.

I'm not versed enough with the Megaquirt to see if a log there can help seek out the issue. The only reason I knew there was an issue was changing back to the stock ECU.

So, it sounds like analyzing some data logs to see if anything comes through with enough resolution is worth a shot?

What you would do would be to log for 15 minutes and see if the sync loss counter went up. If it does you need to take composite logs to try and catch the sync loss. Then you can look at the cam and crank logs and see if something is amiss.

Cool. Guess I need to get some plates on there then.

I would do this, and if it does not show you anything I'd get a scope and start checking sensors and outputs until you find the issue. With a scope you should be able to diagnose the issue in a few hours. If the cam/crank is clean according to the MS log, check it with a scope and then scope the injectors and coils. If it's actually a misfire from loosing sync in some way it will drop an ignition event or injector firing, which is very easy to see on a scope. You can also scope every other sensor to look for any odd behavior.

So, the P0300 did not come back after driving it around and I thought I had finished the drive cycle so I went to emissions and while waiting in line got P0106 instead. I'll investigate this later for the manifold pressure sensor range. I'm still not convinced that the random missfire is solved, either.

Any updates on if the P0300 returned?

It will likely remain unsolved unless someone buys the car from me and figures it out. It will stay for sale until the spring at which point I will fully part-out the car and afterwords, drag the chassis out of the garage with my bare hands and toss it down the driveway.

Does the code not trip if you keep the RPMs say under 4k? I had this exact same thing happen to me last year when it was time for smog check. The only way I was able to pass was to drive around staying under 4k RPMs until all my readiness were ready.

I'm super surprised that the car turned ready with no rpms over 4k.

Even if that bypasses the p0300 to get tagged, I still have the issue, and other CELs have emerged that also need attention. So, I might try this but likely unbolting the whole this is easier at this point.

Just an update on this, it now throws a P2006, which is a VTCS system error, and then the P0300, which I believe is a lean misfire. I've found one thread where someone discusses the P2006 here:

https://forum.miata.net/vb/showthread.php?t=485541

With, of course, no resolution. I've visually inspected the system externally, it stays cold until over 3000 rpm when cold and then stays open above 140 degrees.

At this point it may need a lower intake manifold clean, or it could be something random like an EGR valve needs cleaning. I'm not hugely familiar with the system.

The car is still for sale however, so I thought I'd update this thread. Car still drives very well however.

Curly, can you scope the cam and crank sensors?

Yeah, that is what is so interesting. And "wel"l being relative to the stock ECU of course.

And....the car is sold.