very intermittent misfire under load on dyno - plug gap?

[adampecush]

I was having my car tuned on the dyno this past weekend. Things generally went well, with the exception of the car having an intermittent misfire under load. When i say intermittent, I mean one to two actual misfires per run. This misfire seemed to be independent of boost pressure (was there from 10 to 17 psi, billet 6065 turbo) and didn't occur every run (did 20+ runs). I'll be replacing the plugs with new coppers gapped slightly tighter (0.026"). This sounds like spark blowout to me, but could it be something else? Could there be any other sources of the misfire? FYI, the engine has under 1000 miles, Running AEM V1 with a DH61 igniter, 2jz-ge coils and new plug wires.

Thanks,
Adam

 

[Nick M]

The dyno does not have a 4 or 5 gas analyzer? With it, you know without guessing. A wideband is short on useful information.

 

[RacerXJ220]

What dwell settings are you using?

I would suspect the coils, but not after taking the gap down to .22 or so.

Your turbo is making all that nice cylinder pressure

 

[adampecush]

No multi-gas analyzer, only o2. Only thing I noticed was that there was a very fast, and very obvious puff out the exhaust when the car would misfire.

For dwell settings, I haven't messed around with the settings at all. 100% across all battery voltages and all RPM. Dwell factor 10, max teeth = 2, min teeth = 0.5. Keep in mind that I have no idea what any of that means or controls.

all comments or suggestions welcome.

 

[RacerXJ220]

Try the dwell at 3.5ms

3.0 is as low as I would go. It's a good range to make sure all things are happy, especially since you are wasted spark.

 

[adampecush]

Is this a matter of logging dwell time and targeting to 3.5 ms by messing with the coil dwell factor?

 

[RacerXJ220]

I don't use AEM, and am not familiar with their software. I've always had it locked in with the Stinger 4424 I had on the car, before Pro-efi.

I would not go beyond 3.5ms with pretty much any Toyota coil. 1zz coils and similar, I would not go past 3.0.

 

[adampecush]

so is your dwell ~3.5 ms from idle all the way to redline or does it vary with RPM?

 

[RacerXJ220]

Yes, idle to redline. You can set it lower at idle, and ramp it up near your max RPM. If you experience more break up up top, lower the plug gaps until it gets happy.

This is assuming the coils are strong enough. I am not that familiar with 2JGE coil output.

 

[RacerXJ220]

Have you had any success eliminating the hiccups?

 

[adampecush]

haven't had a chance to try, life has been getting in the way.

With that said, it appears as though the "dwell factor = 10" really means a coil dwell of 1.0 ms. I'll start with changing this and see how things go. Thanks for your help.

 

[Nick M]

Do you have a dyno sheet with a wideband trace?

 

[adampecush]

Unfortunately, this doesn't tell much as the hp and torque lines had minor smoothing applied, and this was from one of the clean runs. Due to time constraints, I didn't grab any additional printouts.

I can tell you that each misfire resulted in a very obvious, very short (100 rpm) 5-10 psi dip in the power curve, and was accompanied by a puff out the exhaust. At most, there would be 2 to 3 misfires per run. We never did look at the o2 printout that closely though.

 

[adampecush]

got new plugs and new injectors. gapped the plugs to .028 and that seems to have cleared up the misfire (as much as I can tell driving the car on the road). Logged my coil dwell and it seems to taper off into the 2ms range as the revs climb - going to try clean this up and see if I can hit low 3s to redline.

 

[RacerXJ220]

The higher past 3.0, the faster more heat will build up in the coils. 3.5 would be what I recommend to be the limit. You should try to get away with it under 3.1.

This comes from testing with 1ZZ coils vs. LQ9 data off a thread on Supraforums, and personal JZ experience.

 

[adampecush]

That's the project for tomorrow.

Also, I think the winter project for the car is a crank trigger wheel unless I can find a cps that actually works. I'm sick of the idle misfire that seems to plague many of these cars.

 

[RacerXJ220]

Fstoy (RIP) sold me his magnetic trigger CPS 7M wheel, modified to accept a similar sensor. It has been in storage for years... I don't think I have a pic, but I will look when I get to my computer. He used it on his Haltech EK6 ecu.

Could give you ideas, make it more simple.

 

[RacerXJ220]

While I am not completely familiar with AEM, there may be a way to slightly retard the injector phasing (I think that's what it's called) to improve idle.

I have also read, some older firmware can cause a miss at different RPMs.

 

[adampecush]

Dwell sorted. Sure does make the car rev smoother throughout the range. I think my default settings had he dwell at around 2 ms, so there is a noticeable improvement. It is tough to tell from pulls on the road, but I haven't noticed he wot miss that was there on the dyno.

As far as the idle goes, I think I'm going to go with a crank-mounted sensor. There seems to be a lot of measurement variability due to the timing belt flexibility and cam gear backlash. I can see this when trying to time the car. Crank mounted wheel should give dead-on timing, just need to figure out how to configure the software.

As far as the injector phasing goes, I've dialled that in. I think the idle issue is completely hardware related. The cps is the only part of the ignition system that was carried over from my tccs days - the car had the idle miss back then, but I wrote it off assuming it was the broken ringlands that had brought down that motor. My current cps ohms in spec when cold, but is out of range when hot. Consequently, the misfiring only happens when hot.

If I was on a stock weight flywheel and crank pulley, I could live with the miss, but I'm lightweight on both ends, so the misfire causes a dip in rpms and subsequently a disturbance in timing and fuelling.

By the way, thanks for your input, I appreciate it. Also, been away for a while, sad to hear about Fstoy.

 

[RacerXJ220]

Anytime. Glad to help.