×

Sale ends todayGet 30% off any course (excluding packages)

Ends in --- --- ---

A really wierd spark issue i havnt seen before

General Tuning Discussion

Forum Posts

Courses

Blog

Tech Articles

Discuss all things tuning in this section. News, products, problems and results. 

= Resolved threads

Author
199 Views

Hello everyone.

As some may know iv been fighting with my GTO for a time now.

Although it works and runs it has a really wierd misfire.

Everything checks out and i have spark on all 6. Needless to say if there is a misfire my AFR will be off.

So im going over the car coil by coil when i notice that the entire coilpack fire is off.

Coil 1 supplies to cyl 1 and 4

Coil 2 supplies to cyl 2 and 5

Coil 3 supplies to cyl 3 and 6.

Right all that is good. Buuuut coil 2 and 3 fire at the same time about 50 deg before cyl 3 reaches tdc.

Accordind to the ECU diagram i have 12V positive to coils then the ECU will earth each coil for 5ms to create the spark ( i have tried 3.2ms 3.5ms 4.0ms 4.5 ms no diffrence)

I cant think its interference as all the harnesses are shielded amd screened.

And also coil 1 works as it should.

It has to be interference but from where? Everything is shielded and screened and checked and rechecked but no cigar.

Also the car idles like this... Sortof.

I have spark on all 6 fuel on all 6 compression on all 6 but for some reason i still get interference.

Im 100% sure this is the last issue then the car will work.

According to manufacturer the coils should be wired off 12V+ from relay directly to the primary coil and then it gets earthed with an internal driver on the ECU.

I suspect the problem here is one 12V wire powers 3 individual coils running semi sequential wasted spark... So when one coil fires it creates some kinda wierd EMF feedback from the other 3 coils causing them to fire though some form of black magic or somtething.

I was thinking to go COP. Coil wil always have 12V in then when it earths it should work. But that is another $600 that i just dont have then hotta add shipping costs and import tax and ill need to cut my manifold and weld it up so the coils can fit under the plenum and and and...

Iv changed:

ECU x 3,

coils x 6,

plug leads x 2,

Crank pickup x2 (magnetic),

60-2 trigger wheel x1

36-1 trigger wheel x1,

changed earth locations,

changed harness x2

changed wires,

changed fuses,

Changed relays

changed sparkplugs x 24,

cam sensor x1

does anyone have an idea what could be causing this? Iv never seen anything like it before.

Thanks in advance.

If you would remind us of exactly which "GTO" it is, and exactly what engine? I assume it's some sort of V6?

Can you post a schematic, or sketch, of the wiring setup?

Ahh right yes... im running a 91 Mitsubishi GTO /3000gt

it has a 6G72 DOHC NA V6 with AWD 4 speed auto gearbox.

system is Spitronics Orion2 i quite like it and never had this issue when installing it on other cars.

it has 3 x Delphi GN10274 coils in wasted spark configuration

Dwell time is 4ms primary coil resistance is 3.5 ohm and secondary coil is 17.5k ohm

Can you provide a csv datalog of it running?

sadly that is a negative. i dont know how to log with this ecu or if it even supports logging.. gotta do it all manually. what i did pickup is a blur on the oscilloscope. like if you measure the voltage on the coil instead of rising and falling it keeps a constant 8 to 10v level but i could be measuring it wrong.

what gets me is coil 1 works but coil 2 and 3 doesnt.

i re-routed the wires along another route and will have a look tomorrow if it made a difference.

i think also the fact that im about 99.99% done now, i could be overlooking obvious things due to excitement.

ill do a writeup of this car when im done cause holy hell this was a massive project

Ya I wanted to see the rpm signal because my gut feel is that it is trigger-related.

that was my first suspect but i checked that on oscilloscope and it seems proper.

i have a new pickup in and airgap is set to 0.9mm with a feeler gauge and had the pickup gear machined on a lathe

I ruled out pickup problem after i added the screen wire. the magnetic pickup has 3 wire. V+, ground and screen. no difference was made from this. also ECU would give me an error code of "missing TDC slot" if pickup was faulty.

i can only think of interference. there has to be some break in the harness. The ECU supplier sells recommended harnesses to work with their units and im using those so i dont think wiring is the issue. EMI is the only remaining culprit but that wont explain why coil 1 works fine and 2 and 3 dont. their wires run grouped together.

my next plan is drill a hole in the firewall for V+ wires and other signal and earth wires through another. but then ill have to take out the harness the ECU came with and make my own

As you say, it's really weird how it's just tied into those two coils and not all three.

It's also out of phase for any expected interference from the other coil' electrics, and changing dwell doesn't seem to change anything, but that last may be due to the small time difference from the dwell. Oh

Because of the phase difference, I was wondering if this could actually be 'triggered' by a problem with the injectors - are they on a completely different circuit, or are they wired together somewhere? I'd also give some thought to the gauge of wiring used, and the separation between the wiring for induction triggers, although that shouldn't be a problem?

I'd also check the ECU wiring, that the gauge is sufficient for the power and grounds.

These are very granny-eggs question, so don't take offense - but have you got it correctly set for 6 cylinders, and it isn't using a base map from an uneven fire engine, like some GM and VW variations - although neither would explain the coils firing together.

well i went into this knowing it might not work. i know of at least 2 other 3000GT owners who also tried going standalone and try as they may they couldnt get it to work. but im determined to get it to run right.

okay so this ECU has 2 harnesses. 1 harness (E22) is input harness. that one works as it should 0 issues. then you have an output harness (E33). this is the one that supplies the circuit for the injectors and coils as well as their power. i have attached the wire diagram for said harness.

So from the ECU all the harness wires travel together until they pass through the firewall. there the power wires split from the signal wires so you have 2 extensions. one is coils, injectors, General purpose output (GPO), and the other is coils power, fuel pump power, and injector power.. then from the signal wires the travel straight along the chassis about 40cm then the coil signal wires split from injectors and teh 2 GPOs. the spark plug leads pass close to the injector wires but they cross at almost perfect 90 degrees and are a good 10cm away from spark plug HT leads. injctor power is off relay 1 and coil and fuel pump power is off relay 2. although i only have 1 red wire to power coils. other 1 is fuel pump and im assuming 3rd wire is for GPOs. so i was thinking what if all 3 wires where for coils? the diagrams says otherwise though and if i do that i wont have 12V for fuel pump.

So i found this. i went through the harness with a fine comb and it seems the 12V wires ARE NOT screened or shielded as i previously thought. this is where im going to start looking for the issue. i got hold of some shielded wire and plan on pulling the supplier harness apart and doing it right.

Also i noticed coil 1 is firing TWICE per revolution of the crankshaft ie. it fires when Cyl 1 is at TDC then a second pulse follows but not at BDC, it fires somewhere else.

so form the battery it has a terminal that attached to Battery positive. then it runs through the firewall to 2 relays 1 powers coils (red) other powers injectors (orange). now where i hypothesize the issue to be is where the coil and injector ground go back through the firewall. its a tight fit and the wires are squished close together paralleled to eachother. the other input harness also passes through this entry point. the ECU controls the 85 and 86 pin of the mentioned relays which then sends power back through the firewall to power injectors and coils.

this is where i suspect the issue because you have current flowing one way then flowing back the other way it came from then when the coil fires current flows back to the ECU as it earths the coil.

my plan now is to isolate the solenoid and coil wires from the power wires. shield and screen them then reevaluate from there. this has to be an issue if not the main issue.

my other hypothesis is earthing issue, the crank magnetic pickup turns at a steady 400rpm under cranking conditions so i dont think its that.

also i do think it might be something stupid that i overlooked and the problem is with me and not the equipment. something silly. its easy to blame the equipment but the equipment only does what the installer tell sit to do.

that is what i have got for today. its hard getting parts and things now during this time of year as most places are closed. i will update more as things progress

Attached Files

Do you have the Advance or Ultimate version of the Spitronics Oriion2 ECU? According the chart below the Standard and Intermediate version of that products only have a single coil driver, where as the others have 4 or 6 respectively.

https://spitronics.com/orion2/orion2-ecu-product/#1617970009089-c76769f7-3ee9

Good point there, David. My initial thoughts, in no particular order -

OK, Enzo, I think you may be getting rather diverted there. Electrically, inductive interference, AKA cross-fire or cross-talk, is mostly a problem when wires with high (kV) voltages, high currents, and/or long runs lie close together - with electronics it can be more of a problem because some circuits are sensitive to milli, micro, or even lower voltage fluctuations. With coils they're normally fired by opening the ground side and collapsing the field that way, but if there's a drop on the positive side voltage, like from a sudden current draw, that may be causing a partial collpse, sufficient to cause it to fire.

With the voltages and currents in the coils' primaries and the injectors' I don't see that being a problem. If you ran the high tension (spark plug) leads together there may be some sort of cross-firing but, while this would be two cylinders together, it would require one to be firing from an actual timed spark, so they'd be 120 or 240 degress from the #1 cylinder's timing.

Hmmm, have you checked the injector timings, to see if they're also showing signs of irregularities, as if they've also got a problem it may help chasing down a common issue, whereas if they're unaffected you could probably dismiss the ECU, to some extent?

Before starting to rip into the rest of the wiring, I would test with some reasonably high current rated jumper/flying leads connecting the coils' 12V to a good feed (starter cable?) and see if that makes any difference, then do the injectors 12V feed, then both using two leads. If you see no change, that 'should' eliminate that side, but if you do, then investigate further. Better yet, would be if you could run separate 12V jumpers to each, to further isolate them.

As I said earlier, I would suspect the problem to stem from the injector side's voltage, or rather current, fluctuations but it could be from the ignition side. In this case I suspect you have unshielded (remember, ground ONE end of shields), or inadequately shielded, issues with the crank sensor wiring running close to the coil or injector wiring and it's being picked up by the ECU which is resetting the crank timing - with the HT leads they don't even need to be that close. Still doesn't directly explain the two coils firing.

Ahh someone who knows the system. im running a Spitronics Orion2 Advanced. Ultimate is for VVT applications. i got hold of some proper shielded wire and gona put that in. the diagram below is the hypothesis for what i think is happening. So if i fix the wires that should fix the problem.

before i do that ill first start with the injectors, i havnt not paid much attention to that circuit. ill check those circuits and make sure im not giving attention to the wrong section. i plan to split the injector circuit on a whole other branch from the coils. then after that ill check the crank pickup again. the idea is to start eliminating sections. like i said this ECU works perfect on other cars, i like the brand. within an hour in another car, its all installed, driving and im tuning. but with MY car.. noooo lol... anyway so i think im going to shield and screen as much as i can just to eliminate possibilities. i think the problem is with the car not the equipment. some hokus pokus Mitsubishi magic somewhere no one know about.

i have to admit the screen wires do confuse me. i dont get how it works really, so maybe i wired that wrong. My crank pickup has a magnetic pickup but it has 3 pins. 1 power 2 ground and 3 screen. The harness that came with the ECU kit does have a screen wire so i connected the pin 3 of the pickup to the screen of the harness. its just some metal on the outside of the shielded core.

so i have today to mess around with it a bit and see what comes up.

thanks again everyone for the assistance, this has been a real frustration although the challenge is quite satisfying. its really appreciated

Attached Files

Could this be rhe problem? I was busy goimg through it all and checkinh terminals when i found this.

These are the 15A and 5A connectors that supply power to relays for coils and injectors

This is the supplied harness. Now im no expert but im pretty sure a wire has to be connected at both ends for things to work

Attached Files

Also i pulled apart the factory harness of the coilpacks. And i found this (see image). The 12V of the coils are connected in series instead of parallele. So from this i suspect interferance could also occur. According to most sources each coil should have their own seperate circuit.

So i think what is happening here is when could 1 fires (the coil furthest from the harness plug), it creates a voltage dip resulting in coil 2 and 3 firing synchronous.

Im not certain about this theory though but im going to rewire it nicely so that each coil has a direct 12V source just to elimimate possibilities.

I did some digging and i found a vid on youtube where Andre said there isnt really a need for shielded wires for coils amd injectors. So im not going to bother with that. On other cars with the same harness i dont have issues so i doubt this one will give issues. There is also now only 1 x 12v input wire going through the firewall adjacent to the harness.

Attached Files

Those connections are parallel. The wiring is just jumpered -- if you measure the resistance, between each ring terminal and the source wire, I'm sure you won't be able to measure the difference. That is an acceptable technique for wiring a multi-terminal power connection.

As David said, the coils are in parallel, and that's a common way of doing it. There may be a small fluctuation in the voltages as the coils operate, but I can't see it being a problem unless the wire gauge is significantly undersized, or there is some other wiring issue.

What could be a problem was the very poorly crimped terminals, though.

Just to clarify, what do you mean by "one" wire?

i just saw now my reply never posted.

i did actually measure it but the resistance was 0.1 ohm difference between coils. so from closest to furthest its 0.0 ohm 0.1 ohm and 0.2 ohm. but that is less than a 3% difference so i doubt it would be the issue. im just eliminating anything possible. but i do the think lose terminal on battery positive definitely has an influence somewhere. but i still dont see anything that isnt how it should be. So you start finding solutions to items that arent problems.

as for the 1 wire you are asking about i am assuming its the 12V? so instead of running a wire for the fuel pump, narrow band, ecu power etc. i run one into the driver well then divide it from there. so only 1 "live" wire is next to the harness instead of 5. but this is also probably not an issue either.

having gone through the harness again, i really cant seem to find any issue except the paralleled connection and the terminal. im betting on the terminal and being hopeful cause i cant find anything else.

once again thanks for the inputs and clarification. im really thankful for any feedback

I thought that might be the case, just make sure you have plenty of current capacity for th wire, it should work OK, as long as there's a minimal voltage drop in it.

I do have a habit of going a couple of sizes, or more, over what others may use where power is a factor - but then I don't work on 'race' cars where weight or packaging may be an issue

If you think you have a voltage supply problem. Use voltage drop as a method of finding it. live circuit loaded voltage drop is where while loads are on in this case while the problem is occurring you have a volt meter with one lead at the source and one lead at the consumer. if while the problem occurs You see voltage in excess of 0.3 you may have an issue in excess of 0.5 you definitely have an issue.... you move the two leads towards each other in the circuit to isolate the problem....

yep i actually removed that one wire completely. i decided that i should add another shielded wire along the side of the car that carries the power through to the dash powering the relays for the coils, injectors, fuel pump as well as the ECU, narrow band, input sensors, drivers and General outputs. i used a relay to switch on the ECU so i removed that and added a push button instead. fused obviously. It also has an earth wire so that i can earth everything from a single point on the inside in the car. if i need to earth something in the front or back of the car i can tap into that earth inside the shielded wire, ensuring that everything has the same earth point.

to clarify; my car has this weird thing where when you "start" it, it cuts power from the ACC and ON key positions while the engine is cranking. so i had to build in a separate circuit for power provision. but i really do not like the idea that my ECU has just 1 fuse between itself and the Battery terminal.

JaK you actually might have a point there... even though i did do many (o soooo many) voltage drop tests i never actually did it with the car running. the most i got was 0.2V battery was 12.9 and coil was 12.75. so i never thought much of it. Granted, i cant make the car run in a way that doesnt make me want to switch it off immediately yet. when the ECU switches on it primes the whole system for 5 seconds and the FP, coils and injectors all gave readings within spec.

Though from my past experiences i have learned that when you have THIS much fidgeting with equipment something is at the brink of breaking and i hope my engine holds up.

im using Mylar 0.22mm screened and shielded wire... could that not be the problem? i know its not the right stuff but its what i can get here and it worked fine on the other cars. i just think that for my car that gauge is very thin and although the supplier of the ECU recommends this wire i have my doubts

I often hear this as a response to the suggestion of live circuit loaded voltage drop tests "I checked the voltage at source and it was 12.8v and I checked the voltage at consumer and it was 12.5v" That's not what I'm talking about. Voltage can change with a key on engine off due to consumption or oscillation in voltage due to charging differences and whatnot what I'm talking about is actually checking the voltage drop across the circuit so putting one lead at the source positive and one lead at the consumer positive what you will see is the resistance in the circuit represented in voltage. This takes all the other potential errors out and if you can do this while the problem is occurring you can positively identify or eliminate this as a possible problem rather than wondering if you put too small of a wire you will know whether you have or not.... You can also perform ground drop tests with the circuit loaded as well so checking ECU grounds like back pinning a ECU ground at a sensor with the voltmeter and then putting the other lead to chassis ground for the battery negative to see what the drop is across that circuit....

Have a wonderful evening

Jak Stoll

i checked the voltage drops. 0.13V positive to positive 0.02V on the negative

Okay so Update:

No improvement. no error codes either

car: 1991 Mitsubishi GTO 6g72 V6 N/A

System Spitronics Orion2 Advanced 36-1 split sequential

issue: coils firing at incorrect intervals simultaneous

what i did so far:

replaced all harness wires with shielded and screened wires.

replaced relays

separated signal circuits (coils and injectors) to two different branches to eliminate interference.

relocated all earth wires to the same point

ran power wires through a different screened wire

repined the terminals connecting to battery voltage

repined injector connectors

repined coil connectors.

no change in system. interference is not present, wiring is correct, ECU is functional.

im going to park the car again until i can think of something else

Update:

im looking into ground loops. so far the only off thing i could find is if i remove the gearbox and engine ground (ie there are no ground wires attached to engine at all) i get a resistance from engine to ground point on chassis of 7.2 ohm. with wires attached its 1.5 ohm. the resistance of the wire by itself not connected at either end is 2.5 ohm

before the wires ran from chassis to battery to gearbox, then also a wire from chassis to engine (this was removed) and then a 3rd from chassis to exhaust (this wire was removed). so now i have 1 wire from battery negative to chassis then i have 1 wire from engine to chassis. i routed all ECU wires to the same point on the chassis

after doing this i do feel that the car stars much easier but the timing light is such a mess i cant determine where the engine is. now it looks like everything is firing when cylinder 3 and 6 are on TDC and every 5th revolution it fires on cylinder 2 and 5. they are all firing at TDC though.

from all this it points to some kind of earthing issue.

nothing interesting to report yet other than that

you guys are not gonna believe this.

finally nailed it.

it was a ground loop. but a special kind not caused by wiring.

turns out engine mount was busted and the bolt was resting against the chassis and shaved through and thus the engine grounded to chassis. Had a voltage drop on my ECU ground section of 0.1V.

I removed all ground wires from the engine and i stil had continuity to the ground point thats when it kind klicked. Resistance on that was 7.5 ohm from engine block to ground point with all earth wires removed from engine and trans. Its suposed to be betweem 1.5 and 2.5.

i did have to relocate the ground wires since this car is littered with ground wires.

also im running factory ECU as i dont want to struggle with my TCU not shifting gears or my suspension not working.

also for some reason the front head was not earthing to the rear head properly so was losing spark there

the ground on the engine was not enough so i had to keep the transmission ground cable the exhaust ground was severed but i left that one off as the sensors there have their own ground.

set the timing and she started right up. still have coil 1 not working but thats probably a foul plug.

my car looks like a spider web from all the wires so just gotta make it neat and set the AFR. then go for a drive

TLDR: there was a ground loop caused by engine mount against chassis, i grounded both heads to ground point, car started right up. engine has 4 ground points (rear head, front head, plenum, transmission).

Some of the things i changed might not have been necessary but these are the results i found and after i changed them the car worked.

im trying to upload a video of how nice it starts but i think the file is too big