It's not often I come here with a serious puzzle, but...

kb58 wrote:

HiTempguy wrote:

kb58 wrote: No, checked and varified by me and the ECU techs. Redline fuel kill is at 8000.

Does the car kill spark or fuel at 8krpm? Maybe try raising the limiter a bit... just a thought. I seriously doubt that is the case though.

Either one or both but I have it killing fuel only.

Interesting. The AEM infinity appears to have a LOT of different variables that could be causing a limiter to be enabled. Specifically the "fuel cut start window". There are also a lot of TqReduceFuel tables and the like that one value could throw things for a loop. I'm sure you've looked through everything though, but sometimes a second look can't hurt...

Just trying to help, sorry if I am muddying the waters, but this is quite an interesting topic.

Right, in the controllers I've played with, some have a setting for a pre-hard limit window so you don't WHAM into the limiter. The fun part is when things are not undocumented, but documented incorrectly.

The guy who writes the documentation isn't always the guy who designed it. Or a lot of times the documentation is written off of the original working spec rather than what actually is shipped. Or sometimes ( ahem ) there's a cool new feature and they don't quite test for what happens if you aren't using that feature, and data/maps for that feature, that should be irrelevant because you have it turned off, end up being a part of the algorithm anyway.

Yes there's a story behind that one...

It's not because it's fun!

I don't want to run my car to 8000 rpm every time I drive it, it's just that I want know what's causing this stinker of a problem. I think root cause is sitting right in front of me, quietly chuckling to itself...

I can just give up and set redline to 7500 rpm, but I'll always know it's there, quietly mocking me... it's the not knowing that bugs me.

Knurled wrote: Right, in the controllers I've played with, some have a setting for a pre-hard limit window so you don't WHAM into the limiter. The fun part is when things are not undocumented, but documented *incorrectly*. The guy who writes the documentation isn't always the guy who designed it. Or a lot of times the documentation is written off of the original working spec rather than what actually is shipped. Or sometimes ( *ahem* ) there's a cool new feature and they don't quite test for what happens if you aren't using that feature, and data/maps for that feature, that should be irrelevant because you have it turned off, end up being a part of the algorithm anyway. Yes there's a story behind that one...

I've wondered that myself, whether it's a bug. I'll have to move the rpm limit maybe 500 rpm higher and see what happens. Since it's supposedly built to 9300, I should be safe, but I don't want to test that!

In reply to kb58:

If it's a rev limit bug, what's causing the MAP to drop? Spark is there, fuel is there- if those are cut off, the MAP loss would not lead everything. Who's forcing the throttle to shut? Isn't that mechanical?

Nothing is forcing it, that's me lifting off the gas after I feel the engine go weak. Note that map drops but not rpm, so I don't think rpm is playing a direct part in this.

I think "going weak" is a clue. The engine retains enough power to maintain speed, but loses the extra power needed to accelerate.

In reply to kb58: So in the end, it could be as simple as cam timing. Running out of breathing.

That was my first thought as well. Not enough airflow through the engine to sustain enough exhaust force to spin the turbo. But if other people are running similar setups without issue then maybe not. (Similar meaning same turbo/cams/boost, not similar as in well one is a 2 liter and the other is a 2.4, and the cams are the same but one is dialed in 2 degrees advanced and the other is 4 degrees retarded, and the turbos are identical except one is a $179 eBay Special and the other is a genuine Garrett...)

"Going soft" is also almsot exactly what weak spark feels like at the top end. Which brings me back to wondering what plugs are in it and what gap, and if going to a colder heat range would help since too-hot heat range manifests on long pulls... probably 1st and 2nd zip up there too quickly for it to be a problem.

alfadriver wrote: In reply to kb58: So in the end, it could be as simple as cam timing. Running out of breathing.

Maybe, though I'd be very surprised if such a sudden drop was due to it, but who knows.

I don't think an actual breathing (ve) problem could happen that abruptly.

So if you set the limiter lower, and bang against it at wot, does map drop?

I'll put that test on the list.

Couldn't find my spark plug measurer/gapper so that killed this evening's planned test drive. Did remove the plugs to check color and they looks really nice, which I guess is a good thing. New ignition coils (OEM Denso coil-on-plug units) should be here this week.

If the k24 uses the same coils as the s2000, then that could be a weak link. In the case of the s2k, when a cool goes bad the car will still run without misfiring but there is a big power loss when giving it the beans.

And I kinda think that soft rev limiter setting could be a possibility. A friend had an issue with his turbo answer where it just had no power at times. He had the soft limiter set to 6,800 and the hard limiter at 7,000. Turns out the setting window for the soft window was for rpm BEFORE the hard limiter (doh!). Changed the soft limiter setting to 200rpm and that fixed it.

alfadriver wrote: In reply to kb58: If it's a rev limit bug, what's causing the MAP to drop? Spark is there, fuel is there- if those are cut off, the MAP loss would not lead everything. Who's forcing the throttle to shut? Isn't that mechanical?

Less talk...more logging and mechanical diagnosis sorry, I'm only saying what you were thinking

kb58 wrote: Couldn't find my spark plug measurer/gapper

May or maynot be your issue but if your talking about the 99 cent "gapper" you get at auto parts stores, toss that thing in the trash and use feeler gauges. Those things are WILDLY inaccurate, like not a couple thousandths but ten thousands off. Mine was 13 thousands out at .030

Raze wrote:

alfadriver wrote: In reply to kb58: If it's a rev limit bug, what's causing the MAP to drop? Spark is there, fuel is there- if those are cut off, the MAP loss would not lead everything. Who's forcing the throttle to shut? Isn't that mechanical?

Less talk...more logging and mechanical diagnosis sorry, I'm only saying what you were thinking

This thread stared 4 days ago now, and we still only have one data file to look at. Need more info!!

Blame it on the pesky day job...

I have a few things to try but need a block of time to do so. What doesn't help is that the only time I have to do street runs is 6-7pm, and there's still a lot of cars on the road. That can be dealt with by waiting for an open space (one of my self-assigned rules is not blasting by anyone). More worrisome is the potential for getting an exhibition of speed and/or speeding ticket (being over 50 yrs old usually helps though...).

This really needs to be done at a track or on a dyno. It's just that each time a promising it's-gotta-be-this-for-sure suggestion comes up, I justify doing one more run on the street, because, surely, that'll be the fix that solves it. The longer this drags out however, my enthusiasm for pushing my luck lessens.

(Putting it on the dyno is no guarantee of finding it. Dyno operators always warn people that the tune may act different on the street. Another unknown is their troubleshooting skills, maybe they'll know what to look for and maybe they won't. It's up to me to come up with a test plan and I found that when building cars, I'm about the only one I can count on to figure things out.)

Here's the current to-do list, doing a run after each change. In order of ease and likelihood:

1. Temporarily raise redline 500 rpm
2. Check plug gaps
3. Close gaps to 0.020 if greater
4. Replace ignition coils.
5. Run hose direct from turbo to wastegate, bypassing controller

Am I missing anything?

In reply to kb58:

I'm more suggesting that you record data on every run. It would be really great if we could manipulate the data to look at it.

BTW, I can... so if you have a .txt or .cvs file of a data run, I can look at it. Being that we are not sure if this is a strategy bug or a mechanical problem or something else- it's really important to record everything you can each time you test it. I just finished a test, and for the almost 40 min of the running test, I was recording something near 300 things- some at 10hz, some much faster. The sum of the data, all compact in a specific file is over 100Mb. It may not be even looked at. But we take it to make sure everything is ok.

Record every run, with as much data as you can.

alfadriver wrote: In reply to kb58: I'm more suggesting that you record data on every run. It would be really great if we could manipulate the data to look at it. BTW, I can... so if you have a .txt or .cvs file of a data run, I can look at it. Being that we are not sure if this is a strategy bug or a mechanical problem or something else- it's really important to record everything you can each time you test it. I just finished a test, and for the almost 40 min of the running test, I was recording something near 300 things- some at 10hz, some much faster. The sum of the data, all compact in a specific file is over 100Mb. It may not be even looked at. But we take it to make sure everything is ok. Record every run, with as much data as you can.

We do the same thing and all of our data gets posted to a location so everyone can access it for the various different ways to look at the data. Just because you don't need it now does not mean you won't need it later to find the beginning of a problem that creeps into being. Storage is cheap compared to time spent guessing at what you did not record.

Once I figured out the logger, I now have it always running anytime the car is moving.

I finally realized that the ECU knows exactly where the cams are - I forgot that it's always reading position because otherwise it wouldn't know what phase of the 4-stroke cycle it's on, doh! So all the cam variables were added to the datalog set. I'll also add the ECU error flags for the cams as well... maybe the sensor signals are getting noisy at high rpm and the ECU is backing off on cam timing? Hopefully I'm making this more complicated than it is and that it's something simple. However, logging variables is free, so whenever I come up with new things to monitor, they get added to the list.

What physical plugs? Brand, number, etc.

In reply to kb58:

Is there a file drop location you can share the files? As long as they are just text or comma separated files, I can use my tools to look at them, too.

Dropbox account!

I would also (if you can) add the data for all (most) software settings that trigger rpm cuts. The columns will just be 0's hopefully. Also, when doing your logs, please continue to keep your foot in it to the redline.

Just like Alfa, I do stuff like this all day (I work in R&D) and I happen to like sifting through data

Knurled wrote: What physical plugs? Brand, number, etc.

NGK BKR9EIX (NGK p/n 2669)

It's what all the cool kids use.

HiTempguy wrote: Dropbox account! I would also (if you can) add the data for all (most) software settings that trigger rpm cuts. The columns will just be 0's hopefully. Also, when doing your logs, please continue to keep your foot in it to the redline. Just like Alfa, I do stuff like this all day (I work in R&D) and I happen to like sifting through data

I'll have to see if I can export csv files from the AEM logger app.

Keep my foot in it until I hit redline.... noooo. Once the "thing" happens, the car stops accelerating and the rpm doesn't increase much after that. Given that I have no idea what it is, I don't want to force it to persist.

In reply to kb58:

As long as it happens before you lift- that should be fine. No reason to break anything beyond what's not working....