3SGTE 5k RPM break-up...out of ideas

[jaypat01]

So I’ve been struggling with a 5k rpm break up on my Gen 2 3sgte running a Link G4+ Thunder ecu. I had it on the dyno the other day and we couldn’t get the car past 5k rpm without it breaking up. This was a consistent problem and happened whether it was in boost or in vacuum. We were only running 12-13psi of boost at this point and still on pump gas 93. We hadn’t even got to the Ethanol mapping yet. First thing that came to mind was a trigger issue. I am currently running a factory 5s crank trigger (36-2) and matching VR sensor for the crank trigger and I am using the factory 3sgte distributor for my cam sync signal. What is odd is I never get a trigger error for either the crank or the cam signals and when I pull a trigger scope while it is breaking up, the trigger signal looks perfect and the sync pulse isn’t close to the missing tooth (I’ve seen an issue with belt stretch causing the sync pulse to fall on either side of the crank missing tooth edge on another car so I new to check for that).

We started tuning it on NGK V-power coppers gapped to .022 and had the issue and tried a set of NGK Iridiums gapped to the same measurement (yes we know how to gap iridiums without damaging them) and we had the exact same issue...nothing changed. 

I am running 1zz coils (ordered 4 new ones from Toyota before the tune and I even tried my older ones, but had the same issue). I’ve rewired the coils again and increased the 12v supply to them and have even logged the voltage of the 12v supply at the coil to see if the voltage dropped ever, but voltage stayed above 14v the whole time...even during the breakup. When we connected the timing light...you can see the strobe increasing frequency in a smooth manner until we hit 5k rpm and then it looks like the frequency cuts in half. This to me looks like the ecu is performing and ignition cut. So we turned off all limiters and I logged “% IGN cut” and “% Fuel Cut” and they stay zero’d during the entire log and through the breakup.
 

Fueling is close enough when this happens that it shouldn’t be causing the break up and EGTs show all cylinders running pretty identical and the spark plugs confirmed it.

My question is to Link, is it possible to have a trigger issue without getting an error? I’ve thought about sensor gap, but the oem VR sensor is bolted to the oil pump with a “collar” on the mounting hole that centers it and I don’t see how there is anyway of adjusting the gap.  
 
This issue is really starting to boggle my mind. It even has my tuner stumped and he is a really intelligent person. I’ve never met an electrical problem I couldn’t solve and as an electrical engineer, this problem has really humbled me. I know this is going to turn out to be something really stupid. If anyone can shed any light on my dilemma, I’m open to anything. 

I will add my coils are grounded on the cylinder head and I have tried two different ground points on the head and haven’t seen a change.

[0x33]

Probably worth attaching your tune and a log whilst the issue is occuring so we can rule out if its an issue with software or more hardware/electrical. 

[jaypat01]

Will do. I will try to get one today or tomorrow.

[Adamw]

And a triggerscope from around the rpm it happens 

[jaypat01]

That’s what I need...a trigger scope capture of it. My tuner did that capture on his laptop and we looked at it but he never saved any of them. I have a log from the other day but it’s quite lengthy. So I will go out today and take a log of just a pull where it breaks up and then I will perform a trigger scope capture. Wish this Thunder was a G4x so it would do that all at once  

[jaypat01]

Alright, so I think I might have found something.

I went out and logged a pull going past 5k rpm and it broke up a bit and then cleared up, but you will see I hit the MAP limit at about 6800rpm. But the break up did happen in the first pull around 3.8secs into the log. This was about 4800 rpm. Spoiler alert... fueling is about 5% lean from my target lambda value but I don't believe this would cause a misfire.

So what I think I found was a very small downward spike in the rising portion of the trigger 1 waveform that is above the arming voltage (4.5v) at about 7.5ms into the trigger scope log. You can see it happens twice above 0v ref, but the second one is below arming voltage. If I am understanding trigger arming correctly, does the trigger arm for the ecu to count the next tooth once the voltage rises past the arming voltage and then the ecu counts the next tooth once it sees the first decrease in voltage (falling edge)? So could this be my issue? Maybe I am over thinking it, but I would assume that the ecu ignores the correct falling edge after that transient downward spike because it never dips below the arming threshold until the next tooth passes....so this is why I don't get a trigger one error. You will have to look at the trigger to see what I'm talking about. Basically, I'm thinking that its possible for the ecu to miscalculate engine position and speed for a short duration and not give a trigger error with what happened in that trigger scope.

 

 

Trigger Scope Log 2021-06-7 12;08;13 pm.llg Log 2021-06-7 12;05;38 pm.llg MR2_06-5-21(I-0,E-R7).pclr

[0x33]

I won't say that you don't have a triggering issue, but I can't think of a time where I had trigger issues that didn't cause a sync error to be present, or have some crazy RPM spike for only a few millseconds.  I'm sure Adam can give a better input regarding trigger issues without a trigger error. 

With that said, this smells like a ignition breakup issue to me. The slight disturbance in the map signal around the areas where you have the issue shows the misfire, but I'd agree that is does not seem to be relating to your fuel mixture (although it is lean).

Your dwell times are what I would consider very high for 1ZZ coils. Averaging 3.6ms. I typically run between 2.0 - 2.4 ms between 13-14volts with stock spark plug gap ~0.030'' on high boost JZ motors fairly regularly. 

As you seem to be running E85, it's also worth checking that you don't have debris in your fuel system clogging an injector. Had a similar issue a while back when dirt clogged injector, which started off as a misfire under high boost, then low boost then misfire under most conditions. 

[Adamw]

Crank sensor is wired back to front.  Swap the +/- wires at the sensor plug.  And I agree with 0x33, way too much dwell for 1ZZ coils.  In my brothers race car they start to overheat and misfire if I run anymore than about 2.2ms.  

[jaypat01]

@0x33 So it is a flex fuel setup, but it is currently running E13 as we were tuning the pump gas map and never got to the high ethanol content map. That being said, I have heard of a lot of people running 600whp on a 3s with these coils on ethanol, so its hard for me to imagine they don't have to power to ignite it.

Additionally, my dwell times at 14v was 3.0ms and I bumped everything up to the settings someone else had been using for a while on their 3s with 1zz coils which made it 3.6ms at 14v. I did that right before this pull to see if it fixed it, but it didn't so I will probably drop it back down.

Also, the injectors are brand new (ID1700xds), but we pulled them off regardless and flow tested them at the shop where we were tuning it. All were matching.

5 minutes ago, Adamw said:

Crank sensor is wired back to front.  Swap the +/- wires at the sensor plug.  And I agree with 0x33, way too much dwell for 1ZZ coils.  In my brothers race car they start to overheat and misfire if I run anymore than about 2.2ms.  

Are you saying to run 2.2ms at 14v? What would you make the 13v setting?

So what am I misunderstanding here, when I look at the help section in the software it would seem that my polarity is correct by looking at the trigger scope and this. @Adamw can you help me understand where I am wrong in this so I know for the future?

[Adamw]

The image in the help file shows the waveform that a tooth produces, I think you might be looking at the waveform of the gap rather than the teeth either side of the gap.

The easiest way to confirm with a missing tooth wheel is look at the falling edges.  The ecu triggers when a falling edge passes through zero volts.  

I have highlighted where the falling edges cross zero in your scope as it is now (pink lines).  Notice you end up with an extra "tooth" right in the middle of the gap.  So rather than 36-2, the ecu will see something like 36-1+1-1

 

And here is your crank signal flipped upside down (same effect as swapping wires), now we have the correct "-2" tooth gap:

 

[jaypat01]

Well that makes sense, you learn something new everyday! Thanks @Adamw. I will update this post if it fixes it...I mean it certainly needs changed anyway.

[0x33]

Nice spot on the polarity! Suprised it ran like that with no errors. 

My general rule with the dwell time is to test the setup and try and find the least amount of dwell that I can run for a given condition. Less dwell = less heat & strain on the coil over a period of time. 

So for your example, just set 13/14v to 2.0ms and go test the car. If you have slight break-up then increase in small increments until you find the sweet spot. Sometimes, less is more. Some coils can do weird things with too much dwell and possibly damage themselves. 

Edit: Might be worth rechecking your base timing after swapping polarity on trigger. 

[jaypat01]

1 hour ago, 0x33 said:

Nice spot on the polarity! Suprised it ran like that with no errors. 

My general rule with the dwell time is to test the setup and try and find the least amount of dwell that I can run for a given condition. Less dwell = less heat & strain on the coil over a period of time. 

So for your example, just set 13/14v to 2.0ms and go test the car. If you have slight break-up then increase in small increments until you find the sweet spot. Sometimes, less is more. Some coils can do weird things with too much dwell and possibly damage themselves. 

Edit: Might be worth rechecking your base timing after swapping polarity on trigger. 

Yeah I was definitely going to double check my offset after swapping polarity. I would assume it would move a little bit. And I will try out your method with dwell. 

Thanks for the help. Hopefully I have good things to report tomorrow.

[jaypat01]

Alright @Adamw, where do I send beer? haha...that did it! It runs incredibly smooth now. Thank you so much for your help as always and thank you as well @0x33 for your input. It ran great without break up on 2.0ms dwell time at 14v so I'm gonna keep it there unless I have issues on E85.