New Install M20: Only 3 cylinders despite solid connections. Resolved

[gybe,]

Well, my "new" wiring harness continues to have issues with the starter ( I am concerned I may have fried the solenoid as all of my electricals check out but the key isn`t cranking it. Can crank straight up from the diagnostic port though, so I went ahead and fired it up anyways.

And....Still only running on three cylinders

So, next up; I will swap out the injectors on that harness (despite being able to trigger them), check valve clearance (as earlier suggested), and order a complete "new" harness for the injectors (despite every electrical check on that side of the harness checking out).

Seriously about to push this damn car into a lake...

[mikey.antonakakis]

You'll figure it out, there's only a finite number of possibilities! And thankfully not too many on an e30.
What are the resistances between the pins on the C191 and the injector pins? You may have a crappy connection somewhere that checks out okay on an ohmmeter but can't handle the full current to power the injectors... This is for both the constant 12V and the DME switched output. 12V for all injectors supposedly goes through a single splice so it's unlikely that's the culprit, but worth checking still. Also check silly things like making sure the C191 pins are making good contact and aren't damaged anywhere.
And if you're going to consider the injectors, swap the injectors between banks and see if the running cylinders follow the injectors or the harness.
The only other thing I can think of is the cylinder ID sensor, if you have one (ring around one of the spark plug wires). I'm not super familiar with the stock DME operation though, not sure if it would cause your issue.

[gybe,]

You'll figure it out, there's only a finite number of possibilities! And thankfully not too many on an e30.
What are the resistances between the pins on the C191 and the injector pins? You may have a crappy connection somewhere that checks out okay on an ohmmeter but can't handle the full current to power the injectors... This is for both the constant 12V and the DME switched output. 12V for all injectors supposedly goes through a single splice so it's unlikely that's the culprit, but worth checking still. Also check silly things like making sure the C191 pins are making good contact and aren't damaged anywhere.
And if you're going to consider the injectors, swap the injectors between banks and see if the running cylinders follow the injectors or the harness.
The only other thing I can think of is the cylinder ID sensor, if you have one (ring around one of the spark plug wires). I'm not super familiar with the stock DME operation though, not sure if it would cause your issue.

Resistance going all the way through the harness is sub 1 ohm on every pin out (that is injector seat all the way to pin 16 and 17 at the 55 pin plug). Also verified each injector bank is getting ~2.4A (so ~800mA per injector). Both C191's that I have are beautiful. Whoever had the car originally pulled the throttle body coolant lines early, so no leakage down into the connector. Cylinder ID sensor could potentially be a culprit, but its resistance is currently in spec and given what I can follow in the logic/electrical layout It doesn`t seem to be as critical in the injector firing...That being said, none of this really makes sense so I can`t rule it out!

I have created some simple remote wiring harnesses that I will use to monitor duty cycle while running as well as remote an injector outside the car to verify it is firing. It is being a bit of a pain in the ass to get this setup, so we shall see how sucessful I am with it At this stage, I am almost fully convinced this is mechanical/valve related, but doing some final checks before I open up the valve cover and have to waste another valve gasket.

[gybe,]

Has taken longer than I wanted to today as I had drained the battery doing all of my other tests, but I have confirmed with my remote harness that when the car is running each bank is firing, so that's cool. Valve cover will come off tomorrow and just got my compression testing kit back, so will be checking that as well. My suspicion at this stage is the intake valves as when I pull a plug for those cylinders I get no smell of gas, but we will see. Fingers crossed that valve lash will resolve this finally (but let's be honest, nothing about this has been that straightforward so I am not hopeful that I will actually have this resolved tomorrow)...

[ForcedFirebird]

Open the injector harness and test the leads there. You should have 7 wires in that harness. Two power leads, one for each bank of injectors. Six grounds that go directly to the ECU and get grounded to pulse. One wire goes to the blue temp sensor, the last goes to the brown temp sensor.

If the car is running on 3 cyl, I would suspect one of the two. One, a power lead is not going to one set of injectors. Easy to test, jump the two thick red wire junctions under the injector harness cover. Number two would be a bad injector driver in the ECU. Each injector driver handles three injectors - but we can rule this out since you changed ECU's.

[gybe,]

Open the injector harness and test the leads there. You should have 7 wires in that harness. Two power leads, one for each bank of injectors. Six grounds that go directly to the ECU and get grounded to pulse. One wire goes to the blue temp sensor, the last goes to the brown temp sensor.

If the car is running on 3 cyl, I would suspect one of the two. One, a power lead is not going to one set of injectors. Easy to test, jump the two thick red wire junctions under the injector harness cover. Number two would be a bad injector driver in the ECU. Each injector driver handles three injectors - but we can rule this out since you changed ECU's.

Yep, have done that. I have used two different engine looms and two different DMEs. In all cases I can trigger the injectors straight from the harness and have equal electrical signals on any test (supply voltage, current, and duty cycle for each bank).
As most of the diagnostics are focused on static or single use (that is, continuity of the leads at all connections in the harness, impedance of the individual injectors/injector banks, firing the banks from the harness through pin 16 and 17) today’s testing was focused on “in operation” observation. I created 6 remote harnesses that allowed me to measure duty cycle at each injector as well as have a remote injector wired to the bank that doesn’t appear to be getting fuel. By doing this I observed that each bank is indeed getting power with the car running, their duty cycles are matched, and that injectors wired into the side that “isn’t supplying fuel” are indeed firing.

So, all in, I have three cylinders that are being supplied fuel and spark (again, have verified at each cylinder), but are not running and the timing lines up perfectly. If I pull the plugs from those three cylinders, there is no smell of fuel. I do still need to swap the injectors around (have been putting that off as it seemed electrical/bank related and pulling them is such an annoying thing with the intake still on) but I either have three bad injectors (from a freshly rebuilt and flow matched set), or I have mechanical interference that is not allowing fuel into the combustion chamber.

Thoughts? I am certainly all ears as this thing is driving me crazy and am out of ideas beyond these last two!

[ForcedFirebird]

Find a friend with a noid injector tester. Make sure you are getting PWM at the injectors missing.

[mikey.antonakakis]

This is definitely getting quite interesting. Those tests certainly seem to show a good harness and good signal to the injectors... Hmm.

My original suggestion to check the valves was based on my own recent experience - was dealing with misfires and seeming like I was running on 4 or 5 cylinders, and in addition to an un-seated spark plug connector/boot, I was entirely missing one of my eccentric nuts with a couple more not even finger tight, so had about 1/4" valve clearance on three valves between two cylinders! Got the plug wire on, and the valve clearances reset, and runs well now. (Although I still have to figure out where that nut went... Probably dropped right to the bottom of the sump - going to pull the pan in a few days).

So yeah, without actually being there, at this point it sounds like you are either dealing with internal mechanical issues (valves, holes in pistons or something terrible) or somehow bad injectors.

[mikey.antonakakis]

Find a friend with a noid injector tester. Make sure you are getting PWM at the injectors missing.

Also this, if you can. Oscilloscope would also work I think.

[gybe,]

Find a friend with a noid injector tester. Make sure you are getting PWM at the injectors missing.

YEp, that's part of what I did yesterday. I can measure PWM and duty cycle with my meter. Definitely getting signal at all injector connections. With the car off and firing from the harness I had verified each injector was indeed firing (using screwdriver stethoscope method), I just can`t actually hear it or feel it once the car is doing it's best impersonation of a diesel burble, hence all the hoops and remote harness nonsense.

While I am certainly frustrated at this stage, I have learned a ton about the car and have also now acquired a healthy set of spares to keep me running at the track. Here is hoping today brings final closure. I will also update the original post at some point to compile everything that has been done.

[gybe,]

She is finally running right!
The final verdict? Was a surprise to me, but it is the actual injectors themselves. That is, out of 6 new "flow matched" bosch injectors, I managed to get three that the solenoid energizes but is not opening the valve. I then managed to put those 3, unknowingly, on the same injector harness. So, every diagnostic test says they were "working" but the valve itself was not actually functioning.

Valve clearance was a bit off across the board as well, some of the intake valves were as low as 0.002", so it was good to go ahead and get all that straightened out too.


Moral of the story: Just because you can hear/feel an injector functioning does not mean that the valve is allowing flow. Even with "new" injectors.

[tomstin]

Congratulations! New, or recondition, part failures can be the hardest to diagnose. Glad you got it sorted.

[mikey.antonakakis]

She is finally running right!
The final verdict? Was a surprise to me, but it is the actual injectors themselves. That is, out of 6 new "flow matched" bosch injectors, I managed to get three that the solenoid energizes but is not opening the valve. I then managed to put those 3, unknowingly, on the same injector harness. So, every diagnostic test says they were "working" but the valve itself was not actually functioning.

Valve clearance was a bit off across the board as well, some of the intake valves were as low as 0.002", so it was good to go ahead and get all that straightened out too.


Moral of the story: Just because you can hear/feel an injector functioning does not mean that the valve is allowing flow. Even with "new" injectors.

That's nuts, but glad you figured it out!
Also re: valve clearance, if it isn't already obvious - if you've got rounded out tops to your valve stems, with a flat feeler gauge the clearances will read lower than actual. My valvetrain is pretty new, just a few thousand miles, but with the aforementioned loose eccentrics I dealt with, some of the valves rounded out already. Even though by feel the clearance was definitely greater than spec, it measured lower with feeler gauges - so for those, I set them with a little smaller gauges between cam lobe and rocker (due to rocker ratio). Let's just say it made my engine MUCH quieter 😂

[ForcedFirebird]

IT happens more than you would think. The first time I used an injector cleaner and didn't lube the injectors enough before setting them in a cabinet for a month or so. Took them out and they were all frozen, was able to only break two or three loose.

[TobyB]

So, even if everything says that the injector is functioning, if you haven`t seen it actually flow fluid, you can`t be certain.

I have a very 'farmered- up' flow tester that I use on everything that goes back into an engine. I had FF's problem with some cleaned ones once
(the cleaner gums up pretty quickly) and diagnosed it that way. Now I just test everything. Takes half an hour, rules out a lot of different things at one go.

Glad you found it.

t