yet_another_na

Hey all, I have been avoiding this post hoping I would be able to diagnose & solve the issue using google-fu, but I have failed. I have been facing a misfire under boost issue and have gone through many of the regular items but am lost on what to check/try next. Here’s a massive list of details below, which hopefully will alleviate some of the noobness.

The car:
early 1995 miata- stock block/2560r/kraken setup/FF640s/FF LS coils/SM clutch/MSPNPpro which was used & abused for ~2 years 20k miles before rod #4 finally let go. Found a cheap MSM with a built motor and stole the driveline. Car currently stands with the following parts:

04’ MSM built motor- ~1.9l Weisco pistons, manley rods, ACL bearings, ARP head/rod/mains, Boundary VVT 2 shim, Fluidampr w/ SM 36-2 trigger wheel
Kraken hotside
GT2860rs w/ 8psi spring & standard MAC EBC valve
Precision 350hp IC
Skunk 2 intake & TB
SM Coolant reroute
FF 640 injectors set to fully seq.
Aeromotive FPR with stock fuel rail, dampener removed for adapter
Walbro 320 fuel pump
FF LS coil on plug kit wired for batch fire
BKR7e’s gapped to 0.020”
94-95 MSPNPpro

The problem:
Car loses power once reaching ~150-170 kpa, with 165kpa being the usual number it occurs. A noise similar to a soft limiter happens, almost like a machine gun brrtbrrtbrrtbrrt. RPM dips up and down, and the AFR leans out from ~11/12 to 13/14. No sync losses are present in the log and it can be run all the way to redline at partial throttle with no issues.

The timeline:
Driveline was swapped in on 11/21. Only wiring changes made were connecting the stock CAM & Crank sensor from the MSM engine into the CAS harness per the VVT swap guide and changing the plug for the transmission harness. Car ran well for ~2 weeks with no obvious issues, just regular break in tuning with many pulls on wastegate & some on EBC. 11/30, came back to the car after being parked for ~8 hours and immediately noticed a heavy break up/misfire under boost. Checked compression and it was ~180psi across all four. Since then, the issue has persisted with no major changes in behavior despite the work done below.

Items replaced/checked without any change to problem:
Ignition/Timing-
Fresh BKR7e gapped to 0.020”
Flow force coil harness repined into 2x 3pin deutchs (the two 3 pins going to the stock plugs)
Borrowed set of known good BKR7e
Borrowed set of known good flow force LS coils & wires
switched between rising edge & falling edge
Retimed & verified all marks were correct
Verified timing belt was tightened at proper spot
Checked valve shim lashes & swapped some around to bring into OEM spec (was only 1-2 thou off)
Checked power, gnd, & signal for CAM/Crank sensors
Checked power, gnd for coils
Verified ECU grounds (2-3 ohms)
Verified each coil trigger wire (2-3 ohms)
Added extra grounds & cleaned the multiple in engine bay (back of head, near throttle body, under brake booster, in trunk, on PPF, battery itself)
New crank sensor gapped to 1 credit card thickness multiple times
Verified base timing with fixed 10 degrees on timing light

Fuel-
Swapped OEM pump (which was most likely tired) for known good walbro 320
Replaced fuel filter
Adjusted base pressure down to 30, up to 60, and back to 43.5
Cycled through multiple tanks of 93
Installed fuel pressure sensor which is included in datalogs (but looks good)
Checked power to injectors & tested for clicking

Tune-
verified against known good tune on the same built motor, efr, 95’ miata I have borrowed the above parts from
Verified timing table against tubro kitty basemap
Disabled overboost, ego, launch, oil pressure safety
+/- fuel across the board (both small and this-is-stupid-you-will-blow-your-motor +/-40 VE changes)
Tried different dwell settings from FF found on MT
Tried fixed value (no table) dwell at 2.5, 3, 3.2, and 3.5
Tried fully seq. vs batch fueling
Unplugged options port connector
Checked almost every setting I can think of

The weird part:
I had checked most everything but the ECU itself. I opened up mine and noticed there was corrosion inside, mostly on the I/O and some on the ignition drivers. This has been there since it was first installed in the car ~1.5 years ago, however I cleaned it off just to make sure. It did not change the issue, but was suspicious of ECU damage. I borrowed a friend’s DIYPNP2, and lo & behold, the car does not break up/misfire under boost and pulls like normal. The exact same tune was loaded, minus the MAP sensor settings (as they are different). Convinced but wanting to be sure, I borrowed another friend’s MSPNPpro, same FW, year, etc with a recent clean bill of health from DIY, and it MISFIRES. WTF? I have swapped the ecus around multiple times to confirm and have gotten consistent results. With the DIYPNP2, no issues. With either MSPNPpro, misfire. While I doubt it’s an ecu issue, I don’t know what would be different between the two which would manifest this way.
MS2:

MSPNPpro:


I’m hoping the powers above me have some advice on other items to check/test, as I’m completely out of ideas apart from buying new injectors, swapping to a ING1a setup, and coping with running a MS2 at this point. I’ve attached two screenshots above of very similar pulls displaying the difference between the MS2 & the MS3, datalogs for both, and a composite log taken while misfiring on the MSPNPpro. Some items to note-
-the DIYPNP2 options port is different, so oil pressure, temp, and AFR are not logged.
-the DIYPNP2 may have a bad capacitor(?) on the 5v circuit due to someone plugging it into the incorrect harness after buying it on ebay and then returning it, so, uh, that’s most likely why the TPS & CLT signals bounce and aren’t accurate.
-In the MSPNPpro logs, Oil Temp is actually fuel pressure. I am limited by the # of analog inputs and did not change the name from what the input used to be before testing.
-The MSPNPpro logs do not match up with 100% with the tune provided. Dwell and VE table were adjusted both +/- throughout the log, but the issue remains consistent. More logs matching up directly can be provided as well!

patsmx5

iTrader: (16)

Graphs are fuzzy on my side, but my guess is too much fuel. Looks like PW jumps up from ~10 to ~12 ms between the graphs. The behavior matches a rich misfire.

Also, always a safe bet to drop the gap on the plugs. Try .015".

yet_another_na

Was informed the second MSPNPpro log is unviewable, guess the export to mlg from megalogviewer killed it (had to trim the log because it was 4mb over the 20mb upload limit). However, here's another screenshot from it with the same PW as the MS2 for a better comparison, and hopefully not fuzzy. I'll get some fresh logs with both 0.015 gap, and with slightly less fuel across the board as well, thanks pat!

The same MS2 photo:


MSPNPpro with 10PW:

redursidae

I would try setting your dwell to standard 3ms, and set your injector timing table to 360. Take a log with these changes, although you may need to regap as patsmx5 said. It could be me, but your MS3 MSQ has a lot of warnings when I open it. Maybe start with a fresh tune and move your settings over?

yet_another_na

I've tried the dwell at 3ms fixed, which was the broken MSPNPpro_2 log from the original post. A fresh slate on the tune is probably a good idea, I've had the same tune for a few years now and it's gone through many saves/opens. I can find a way to upload that log if you'd like to confirm, but it behaved the same. HOWEVER dropping the gap to 0.015 moved the misfire up by 40kpa to ~200kpa on average. It no longer misfires on wastegate (because it opens at 170kpa) but will when EBC is enabled. I included three logs below, all with the gap down at 0.015. There is one log with the same settings, which yields a ~190kpa misfire, then a log with -10 on the VE without a misfire (however, I realize the kpa did not increase beyond 190kpa so it's possible I didn't give enough beans), and finally one with -10 VE and a 360'ed injector timing table. The first pull did not misfire under high boost with the 360 timing table, however the other two pulls in the log did. I did do a single -20 VE, however it misfired the same and I did not record a log for that one.

0.015 gap, same settings


0.015 gap, -10 VE


0.015 gap, -10 VE, 360'ed injector table- no misfire high boost


0.015 gap, -10 VE, 360'ed injector table- YES misfire high boost


Given this evidence, my best guess would be weak spark combined with crappy/rich tune = misfire. I did not see any change when going from 0.035 to 0.020 and checking that gap multiple times, but I guess the spark was weak enough to need 0.015. Someone suggested I clean the ignition switch as it's the 12V supply for the coils, as well as trying a new main relay. I'm temped to say screw it and go with a fresh setup of the ING1a coils & new wires just to rule that out, even having tested a known good set of LS coils and wires.

patsmx5

iTrader: (16)

Weak ignition is a common problem on these cars. Knowing it responds to a drop in gap gets you looking at the right parts.

My have found your problem "FF LS coil on plug kit wired for batch fire"

What dwell are you running? What dwell's have you tried? A lot of coils designed to run 1 cylinder don't like batch fire.

Just run this coil setup for all the spark. 200mJ, 81k volts, 10mm wires.

yet_another_na

I definitely agree with that. I've read some of the other threads about the FF coils, esp. in batch and it wouldn't surprise me. I've used the following so far for dwell values, with no noticeable differences:
-Fixed values (no table ms2 style) of 2.0, 2.5, 3.0, 3.2, 3.5
-FF recommended values
-https://www.miataturbo.net/diy-turbo-discussion-14/lsx-coil-thread-82744/page10/#post1475657
-^ the above is what is currently loaded onto the tune

I greatly appreciate all of the advice and double checking of my dumbassery
That coil setup is calling my name (I assume those are the ING1a coils). I *may*, depending on how lazy I am, run two signal lines and attempt sequential on the FF coils first, but I'll most likely end up ordering the bits and have ALLOFIT spark

patsmx5

iTrader: (16)

Yeap, they are MSD IGN1A with a custom bracket I make. I haven't pushed these super hard yet, but I will be running 40-50 psi on methanol.

redursidae

I agree that it seems like a weak spark issue, but for my own curiosity, I wonder why the MS2 isn't misfiring. The only two major differences I see between the tunes are the injector firing angles and the dwell times. Any thoughts on that @patsmx5?

patsmx5

iTrader: (16)

Could be a few things. One, is ignition accuracy. The MS2 being less accurate, potentially could be delivering less dwell (or more), or could be firing the coil more advanced than commanded. The more advance you run, the lower the cylinder pressure at the time of ignition, thus less voltage is required to ionize and jump the gap. AFR could have been different which can definitely change things regarding lighting off the mixture.

SpartanSV

You are well within the capabilities of stock coils. Install NGK iridium plugs, stock coils, and put the big boy ignition system money into other things. No one needs aftermarket coils with a 2560r.

yet_another_na

No more misfires! DIY'ed a bracket and harness for some ING1As, and now am able to run ~230 kpa without any misfires or other weirdness. It's not you, it's me, or something like that. Difference between the MS2 & MS3 is still a bit of a curiosity and threw me for a loop, but what Pat mentioned makes sense. It's also likely it was still misfiring on the MS2 but to a lesser degree. I agree it might be a bit overkill for my needs, but it is a built motor running a 2860rs now, and the price difference between the ING1As & other COP options was minimal. Well worth the peace of mind to (hopefully) do it once and do it right. I appreciate all of the responses and feedback! TLDR- it's always ignition