Misfires at peak load - No AFR change?

[utc_pyro]

Ok, this one has me pulling my hair out.

 

At peak load, the car is jerking/shuddering intermediately. This is sometimes accompanied pops and puffs of black smoke out the exhaust, but NO changes in wide-band O2 reading like one would expect for a misfire/afterfire. It does show up on the roughness monitors though, and seems isolated to cylinder 1 and 3.

 

Problems started after installing a Grimmspeed intake and going through MAF and load comp tuning. More specifically during day two of WOT MAF tuning it did it when I had overcorrected by 7% and it was pig rich. Sence then it's gradually started to happen more often. As of tonight when I went looking for it, it's doing it every wot pull above 1st gear.

 

Relevant mods:

Fresh-ish Shortblock (10K miles)

Rebuilt heads

BNR EVO3 16G Turbo

ID1050X Injectors

Radium fuel rail and pulse dampers (three)

TGV Deleats

Grimmspeed TMIC

Grimmspeed Intake

Tomei ELH

AEM Wideband Failsafe Gauge

 

So what would cause a missfire/afterfiare on two cylinders only at high load, but not show up on the wideband?

romraiderlog_20180621_211257.csv

romraiderlog_20180621_210724.csv

[mwiener2]

What spark plugs and what gap?

[cryo]

same question as above as well as what was the afr when it went "pig rich" and how long was that pull ie full pull/partial pull etc. also how old were the plugs prior to this tuning session?

 

how many miles on the car and are the coils original? connectors in good condition?

 

It sounds like a misfire even though the wideband isnt going lean. It may only be on one cylinder but picking up two as its a calculation based on cam and crank position as to which cylinder is firing/misfiring.

 

Dave

[utc_pyro]

What spark plugs and what gap?

 

To my knoladge they are stock ifr8b’s with OEM gap.

 

as well as what was the afr when it went "pig rich" and how long was that pull ie full pull/partial pull etc.

 

10.4, 90% of a 3rd gear pull before aborting, fueling target was ~11. I corrected the MAF scaling based on that one and subsequent pulls came out within 1%. Upon further revision there was an up to 3% day to day variation at high RPM, but each day was fairly consistent with it’self.

 

That was just the first time it happened, it then started happening progressively more often in long third gear pulls about a week later. Those logs are from when it started doing it in shorter 2nd gear pulls.

 

also how old were the plugs prior to this tuning session?

 

10K miles, they were changed with the rebuild.

 

how many miles on the car and are the coils original? connectors in good condition?

 

155K on chassis/electrical system. Coils are original, and per the shop that rebuilt the motor connectors are OK other than one latch tab being broke.

 

It sounds like a misfire even though the wideband isnt going lean. It may only be on one cylinder but picking up two as its a calculation based on cam and crank position as to which cylinder is firing/misfiring.

 

Ok, that changes thing! Didn’t realize the ECU could confuse things, and thus was trying to figure out what would cause two to fail at once. Does this typically show up leading or lagging the problem cylinder?

Edited June 23, 2018 by utc_pyro

[utc_pyro]

Went and changed the passenger side plugs and coils this morning.

 

Plugs were LFR6AIX-11's, which I think are NA plugs:spin:. Replaced these with SILFR6B8's, the OEM side gapped ones from '10+ LGT's. Not sure how to "read" the plugs but my guess is this says "running rich":

 

The connector on coil #3 was disintegrating, and the coil had corrosion on one of it's contacts. I replaced that connector, and both coils with Cosmos Racing ones.

 

 

So thanks mwiener2 and cryo. It looks like it was exactly what both of you said.

Edited June 23, 2018 by utc_pyro
Removed theoretical bullshit about why it acted like this.

[rhino6303]

Good catch!

 

I've been running the cosmo coils for over 20k with no issues but I don't live in the salt belt.

[mwiener2]

LFR6AIX-11 is the aftermarket direct replacement for the factory plug SILFR6B. The -11 means it is gaped at 0.044" out of the box. This is too much gap.

The plugs in your pictures look fairly normal as far as color and wear.

 

 

 

Correct spark plug you should be using is LFR7AIX. Should be gaped to about 0.028". If you use E85, gap down to 0.024".

 

 

I've also never had good luck with aftermarket coil packs. New OEM coil packs are the best way to go.

Edited June 23, 2018 by mwiener2

[utc_pyro]

Aren’t LFR7AIX the one step colder plugs? Also at what point does one need to actually gap down to .7mm (0.028”) from the stock .8mm(.031”)?

 

Due to part avalibility getting harder on these cars my tuning is a little odd. I’m targeting a flat 3.0g/rev, about 16psig. I though you only needed to gap down when you run higher load/boost as you’re compensating for the fact the spark gap one can jump so far per unit gas density give the fixed spark voltage. So the maximum spark gap would be determined by the air load per rev and timing (les timing = higher dynamic pressure at time of spark = shorted maximum spark).

 

Not tossing the old coils, going to clean up the contacts and put them on the shelf Incase issues come up with the cosmos ones I had on hand.

Edited June 23, 2018 by utc_pyro

[mwiener2]

One step colder to prevent knock. You're over stock power, you go one step colder.

 

Stock plug gap is for emissions and fuel economy. 0.028" better supports power. You only need to go tighter if you are on E85 running high boost. You're actually not going to see much difference between 0.030 and 0.028 on gasoline.

 

 

What part availability problem? WRX's use the same motor through 2013

[utc_pyro]

Thanks for the info!

 

What part availability problem? WRX's use the same motor through 2013

 

Not so worried about the motor as the 5EAT. ClimberD is no longer making reinforcing parts for them, and low mileage junkyard units are drying up.

[solidxsnake]

One thing I'll mention: I assume you're reading the wideband through the ECU (TGV sensor input)? If that's the case, how are your logging rates? It's possible too that you're not seeing the rich dip because it happens quickly between samples in your log.

 

 

Looks like you found the problem anyway, but just wanted to add something for anyone reading.

[utc_pyro]

custom canbus to serial converter . It gets a new AEMNet value every 10ms and spits the data out at 19200 baud.

 

Even in logs with a 17/18hz poling rate I didn't see more than ~3% fluctuations.