Cannot fix Single cylinder misfire under load.

Ok, so i just re attached my lead to the trigger wire directly at the number 3 coil. I am now getting a different look. I also swapped out the only aftermarket coil from 5 to 3 and that seems to be where the lower amperage spike is now.

Here is the new screenshot.

 

So, what do I need to do next? Josh had mentioned testing the injectors and others have mentioned testing voltage drops. If you guys could walk me through how to perform these tests, or point me to a site that explains it I'd appreciate it. I will also try to drive this again and make it misfire to see what the scope reads considering my previous attempt was wrong.

I 'suppose' in the other capture you did not connect to the coil but to an injector.There are a couple of reasons to believe this:

1 - how would still current flow through the coil as indicated by the blue trace without being triggered?
2 - the shape and time of the sync signal (red trace) looks nowhere normal for a coil trigger.

Where the signal onthe red trace disapperars is probably when the kijector was shut down.

Though the difference in coil current is not normal I would not focus on that for now. It maybe caused by different coil designs, simple as this. 9 amps is still enough to fire a coil under heavy load in my book.

Try to capture both injectors of the missing and of a good cylinder when the misfire occurs, one cylinder at a time.

You also may want to take crank sensor capture under different load and rpm conditions to check for any possible deviations that may cause the ecm to see a misfire, be it real or not.

I 'suppose' in the other capture you did not connect to the coil but to an injector.There are a couple of reasons to believe this:

1 - how would still current flow through the coil as indicated by the blue trace without being triggered?
2 - the shape and time of the sync signal (red trace) looks nowhere normal for a coil trigger.

Where the signal onthe red trace disapperars is probably when the kijector was shut down.

Though the difference in coil current is not normal I would not focus on that for now. It maybe caused by different coil designs, simple as this. 9 amps is still enough to fire a coil under heavy load in my book.

Try to capture both injectors of the missing and of a good cylinder when the misfire occurs, one cylinder at a time.

You also may want to take crank sensor capture under different load and rpm conditions to check for any possible deviations that may cause the ecm to see a misfire, be it real or not.

I 'suppose' in the other capture you did not connect to the coil but to an injector.There are a couple of reasons to believe this:

1 - how would still current flow through the coil as indicated by the blue trace without being triggered?
2 - the shape and time of the sync signal (red trace) looks nowhere normal for a coil trigger.

Where the signal onthe red trace disapperars is probably when the kijector was shut down.

Though the difference in coil current is not normal I would not focus on that for now. It maybe caused by different coil designs, simple as this. 9 amps is still enough to fire a coil under heavy load in my book.

Try to capture both injectors of the missing and of a good cylinder when the misfire occurs, one cylinder at a time.

You also may want to take crank sensor capture under different load and rpm conditions to check for any possible deviations that may cause the ecm to see a misfire, be it real or not.

Ok, will do. I’ll report back when I have a chance to do that. Hopefully this evening. 

Apologies for late report. Its a bit tough to get out and mess with this thing having a toddler.

So I hooked up my 2 channel scope to the port injectors on 1 and 3 to get a comparison and then made the truck misfire like before. It looks like the port injectors turn off at around 2300 rpm, so I will need to try the direct injectors next. All I could tell from the capture was that both of the injectors turned off at 23-2500 rpm and after the misfire happened at around 3500 rpm, only number 1 injector came back on when the engine rpm came down. 

This is where both turned off at around 2300 rpm or so.

 

And this is after the misfire and as the port injectors turn back on.

 

That makes sense. #3 was shut down after the misfire to protect the catalyst.

I've read that port injectors are used when the engine is cold and at low RPM, so that would make sense that they would shut down at 2,300 RPM.

Get one capture connected to #1 port injector on one channel and #1 gdi inyector on the other channel. This should show the transitions between both injection system on acceleration and deceleration while working normally.

Then get another capture, this time connected to #1 and #3 GDI injectors. .

This may allow to relate the misfire on #3 with 'somerthing' concerning the direct injection..

Did you ever play with the GDI injectors, swapping them around or looking for pressure changes in the rail while activating them?

Often GDI injectors' power feeds are grouped in pairs, so probably no fault there.

Ok, I'll do the tests you requested as soon as possible.

When I installed the engine back in March or so, the misfire was present. I installed a rebuilt set of direct injectors then and the misfire has stayed on cylinder 3. I suppose there is a chance that the new set had a bad injector and I just happened to install it on the same cylinder that already had the issue, but I doubt it. 

I have not tried to do any work on the GDIs. I do not have a scan tool that can do that activate them individually or check the pressure in that rail. I've only been able to check fuel pressure in the port rail. I did however already replace the high pressure pump last year. From what I've read and watched videos on, the GDIs get different voltages on both the negative and positive feeds from the ECM, so I may have to use both channels to view each direct injector. I'm hoping I can get a signal off of one of those by themselves so I can see both direct and port injectors like you requested.

Ok, so I went out and hooked up the amp clamp to one of the wires that feed the direct injectors and the other channel I hooked to the number 3 port injector. With this unfortunately, the trigger goes away when that injector turns off but I wanted to see what that injector was doing. I may try again using the number 3 coil for reference. 

I can not see any differences in the current ramping between the cylinders, but there is definitely a hump every other firing. Once the misfire happens, that humps no longer exists. I don't know why, but thats the only thing I could see that was different before and after the event. I will add photos below. I did the same tests with my trigger on cylinder 1 port injector and the capture looks exactly the same.

 

 

 

Because of the low sample rate the meaning of the amp capture is limited. You may want to select a shorter time base and crank up the sample rate ..

Is there anything else on this circuit that provides power to the port injectors?

So it's the port injector that (gets) shut(s) down, good to know.Do the same test with the GDI injectors.

Misfire detection is based on crankshaft acceleration so a varying/defective ckp signal may cause an issue. I would look at the ckp signal under varying load/rpm conditions, again with the highest sample rate possible, alone and with the CMP signal.

Is there any misfire count for the companion cylinder, mane in Mode 6?

I'm not sure how to get a better sample rate? I have googled it and can't seem to find a straight forward answer. I can change to timeframe to 20ms, or should I go even shorter?

I believe the power wire to the port injectors does not power anything else. Just so we're on the same page though, the tests I did last night was the current probe on the direct injectors with only channel b on cylinder 3 port injector for triggering.

What I believe I've proved though is that after the misfire, the ecm turns both the port and direct injector off for cylinder 3, but the coil does not seem to be affected. 
During normal driving both sets of injectors are used up until around 2300 rpm or so and after, the port injectors turn off and that also seems to be the case under load as well. Generally speaking this misfire has happened only under load at higher rpm which makes me believe that the port injectors aren't part of the issue. Josh had mentioned earlier in this thread that he thinks its a spark issue, but I couldn't see any differences in before or after the misfire with the the current clamp on the power wire to the coils. Perhaps that is the lower sample rate? I will try to resolve that and test again. 

I have not tested the crank sensor yet, but will try that as well.

When looking at my scan tool, I am seeing a misfire count for cylinder 3, and maybe 1 or 2 misfires on an adjacent cylinder, but will have to check again to see which one. 

I’ve messed with sample rates and this seems to be the best rate I can get with the 2204a and picoscope 7

I think you should disconnect the direct injectors (perhaps unplug the fuse unless that cuts power to other things as well) and see if the misfire happens under load using the port injectors.

I can't remember - did you scope current on the coils under the fault condition?

I re-read your original post. The dealer said there was a 25% leak out the exhaust valves (or something like that), yet the machine shop found no mechanical problems at all? Did they replace the rocker arms for cyl #3?

You should scope the direct injectors and find the precise moment that #3 cuts out (intentional shutdown) and see if there's any anomaly in the pattern in the few seconds before that moment.

If after all this you can't find any problem, I would swap #3 direct injector with another one.

I’m messing with it now, so I’ll try that. I’m not sure if it’ll even run with those unplugged though. We’ll see. 

I did a current scope with the power feed wire on the coils during a misfire event and found nothing. I’ll have to see if I still have the screen shot though. 

 

The rockers were all in a bag when i got the heads back, so I doubt the rockers that were on #3 before I took the engine apart made it back to the same posts again. If  cant find any issues here with the electrical, I'll definitely pull the cover off and check it all out again. The rockers on this engine are the type that sit on the valve and lash adjuster and the roller in the middle sits on the cam. 

I don't mind swapping the direct injectors if you think that's needed. I'll need to get new intake gaskets though beforehand. They are the metal crush type and I don't think they are reuseable. Do you think I need to replace the direct injector seals when I swap them? They are the teflon type. 

I don't know about the seals but I assume they should be replaced.

JoshuaK wrote:

I don't know about the seals but I assume they should be replaced.

 

Hopefully the quote part works now so you guys know who I'm responding to.

I tried unplugging the direct injectors and the motor would stall at idle after it warmed up and wouldn't run at all even with my foot on the throttle, so that was a no go. 

I did another current ramp test on the power feed to the coils and made the truck misfire. I see nothing in waveform that suggests that the coil isn't firing. 

I tested the CPS to see if it was acting up during the event and again, I see nothing to suggest that. 

 

I then tried hooking up the IGF for a shot of its waveform and I couldn't get a signal. I checked as best I could online and in the wiring diagrams I've been using for this truck to see if toyota uses that extra coil wire for a ground or for IGF and it appears that I should be getting a squarewave out of that wire to signal that the coil did in fact fire. All I'm seeing is what appears to be a random trash signal that doesn't see to match anything. I'd really like to confirm this at some point if possible. I am currently watching ScannerDanner video "operation and testing coil over plug ignition (4 wire) toyota/lexus" so if my truck is the same as this one, I should be able to get a signal. I am using my main lead on the signal wire at the ecm for bank 1 and the ground is grounded to the frame of the truck (same ground I'm using for the other lead). This is the only waveform I got from this.

 

I will try again next chance I get to do this test again, with both banks.
 

I’m curious. If the ecm can’t get feedback from the coils, would it shut down the cylinder? I’m shooting in the dark of course, but it’s a valid question