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

... this is one that neither I nor anyone else has been able to solve. Posting on Honda forums, the ECU forum, and turbo forums hasn't lead to a solution, so here's your chance to shine.

Background: Homebuilt car, "Midlana" (www.midlana.com)

Turbo K24

9:1 CR

Twin-scroll GTX3576R w/0.82 AR turbine

4-port boost control valve

38 mm Tial watercooled wastegate (brand new)

150 kpa wastegate spring

RSX cams

AEM Infinity ECU

8000 redline

HP, around 400, about 240 kpa boost on E85 with this engine

The problem:

At around 7500 rpm in 3rd-6th, MAP drops off from target boost to ~140 kpa. Note that this is less than wastegate spring pressure. (The problem does not occur in second gear.)

Things I've noticed or tried:

1. As I've worked trying to solve this, I realized that the problem's been there a long time, but I thought it was other things.
2. I hardly ever run the car >7500 rpm simple because it's so light and with that much power, breaking the law is never more than about 2.5 seconds away. Because of this, I hardly ever noticed the problem before. When it did happen (on-track mostly), as said before, I thought it was other things.
3. The problem existed before changing the ECU to the current one, that "seems" to rule out an incorrect ECU setting.
4. The problem existed before going with this larger turbo (the previous turbo was a GT3071R).
5. Removing the muffler made little difference.
6. Switching off closed-loop boost control and running solely on the wastegate spring reduced the problem somewhat, but didn't eliminate it.

Things that I've discounted:

1. I don't think it's an oil supply problem because it's running a dry sump. (Honda VTEC engines are very oil pressure sensitive because they rely on oil pressure to drive the cams to position.)
2. I don't think it's a leak in the intake tract. The blow-off valve tested good, but more important, I can run any boost level up to 240 kpa all day long - just as long as I stay below ~7500 rpm.
3. I no longer think it's an ECU issue, mostly because the problem was present with the previous ECU. Also, since MAP drops off at "about" 7500 rpm, and there's no limits in the tune in that range, I don't think it's a programmed limit kicking in.
4. Exhaust back pressure hasn't been measured, but since the turbo is good for >25% more power than I'm running, removing the muffler didn't noticeably change the issue, and because the exhaust is a free-flowing 3" setup, it doesn't seem like a contributor.

Other things I plan to do:

1. Check the plugs and possibly go with -8 instead of -7 heat range.
2. Replace the on-plug ignition coils. The thinking is that maybe one of the plugs is becoming intermittent at high rpm and high boost.

At this point I think it's a "can't see the forest for the trees" sort of thing because I'm too close to the problem. I "could" just ignore the problem and stay below 7500 rpm, but the not knowing is getting to me. I "could" give up and take it to a tuner, but I'm stubborn and want to solve it myself... well, okay, maybe with some help. As a bonus for reading to the end, I'm offering a free copy of my book (how to build your own Midlana) as a reward for being the first to nail root cause.

Below is a logger showing the problem. Note that MAP drops off even though the throttle remains at 100%. Also note that engine rpm does not drop off, indicating it's probably not an ECU engine speed limit. In the shot below, you can see the ECU increasing the duty cycle to the boost control valve in an effort to maintain target boost, but it fails:

Link to high-res pic of logger

It's load related. My bet would be that the turbo is not keeping up/cool enough under high rpm/load. Or you have a boost leak from things flexing/hitting im those gears/rpm ranges.

Have you tried without the boost controller? What about pressure testing the cold side?

First of all, I really like that software. plotting the data looks to come out pretty easy- we use something quite similar.

Two things come to mind- are you using a bypass valve? How are you controlling that, and check it's calibration to make sure you don't have some override opening it at 7500 rpm- that will bleed off pressure really quick.

Second- upstream of the turbo- how flexible of a system is that? It's possible that the massive flow of air flow at those speeds is collapsing something and causing a restriction. Why not second? Not enough time- you probably have to shift so quickly to change gears that it may not be moving.

that's what I can think of.

If you can rule out an ignition issue, which is where I'd look first, the next thing is boost leak. Since you swapped turbos, and you haven't mentioned an external wastegate, we rule that out. Some wierd blowoff valve vibration related leak that happens only at high RPM? Doubtful. A hole in the plumbing that only opens at 7500? Let's go back to ignition. I see your spark timing line thinks it's doing everything right, and knock sensor signals all show zero, but wonder if ECU is pulling timing for some other reason? We chased this problem to a ground loop that somehow interfered with the knock sensor signal... something wrong on an adjacent pin IIRC. I kinda doubt it's as easy as changing coils. If you have a coil laying down, it's kinda hard not to notice the whole misfire thing, unless you're getting real lucky in the sense that it's still putting out spark, but weak, so bad burn and lower output and lower boost but no discernible misfire. Either way, bad ignition should show up on the AFR data, sudden rich condition at 7500. Which would make ECU pull fuel. I don't see an AFR trace. What's it do I wonder? The only other thing off top o' my head would be some vtec problem where (assuming you've assigned such duties to the ecu) the cam timing is going wrong for whatever reason but I know nuthin bout such things.

I don't think your hot side is big enough. Likely need a 1.2AR

1.2 On a 2.4l? That seems MASSIVE! Expecially considering how large the turbo is already.

Hondas flow a ton of air at high RPMs. The previous turbo likely had a smaller or same size hot side - so same issue.

A 1.06 AR would likely work too, but I do think the overall issue is the hotside.

Dusterbd13 wrote: It's load related. My bet would be that the turbo is not keeping up/cool enough under high rpm/load. Or you have a boost leak from things flexing/hitting im those gears/rpm ranges.

Agree it's likely load-related. The turbo is a bit on the large side - "Honda kids" are making well over 600 hp with this one, which is why I don't think it's the limiting factor.

I don't see how this could be due to a boost leak. That is, I can run up to 240 kpa all day long - no leaks. However, if I set boost target to something much lower, say, 170 kpa, it dies at ~7500. How can there be a leak at 170 kpa but not 240 kpa.

I'm not arguing, just trying to filter out what it "probably" isn't.

chiodos wrote: Have you tried without the boost controller? What about pressure testing the cold side?

Yes, I noted that above, that running solely off the wastegate spring, the drop-off still occurs. What's strange is that boost drops off to less than spring pressure.

alfadriver wrote: First of all, I really like that software. plotting the data looks to come out pretty easy- we use something quite similar. Two things come to mind- are you using a bypass valve? How are you controlling that, and check it's calibration to make sure you don't have some override opening it at 7500 rpm- that will bleed off pressure really quick. Second- upstream of the turbo- how flexible of a system is that? It's possible that the massive flow of air flow at those speeds is collapsing something and causing a restriction. Why not second? Not enough time- you probably have to shift so quickly to change gears that it may not be moving. that's what I can think of.

Bypass valve, also known as a blow-off valve, yes, I have one. It's plumbed to the intake manifold, and will open when negative pressure is seen (like when lifting off the gas). I checked that it's functioning correctly and does not leak.

Upstream of the turbo inlet side is a 4" silicon coupler to the air filter housing assembly. I doubt that's collapsing.

The car has a very "long" gear set. That is, 8000 rpm in second is around 80 mph, so there's actually a fair bit of time spent there, but I get what you're saying.

Here's the air filter box - straight into the turbo

DaveEstey wrote: I don't think your hot side is big enough. Likely need a 1.2AR

Understood, but why would boost drop off like a light switch? If the turbine housing was too small, I can see it slowly dropping off, but not going near-vertical. As said above, since other people are making 600 hp or so, I don't tend to think that's it.

I realize that I can find the problem by parts substitution, but take the turbine housing for example, a new one is $550. That's a very expensive experiment if it proves to not be the cause. Also, the previous turbo was a smaller GT3071 with a similar 0.82 housing. I'd expect that - if the housing was the cause - to produce the problem at an even lower rpm, but it didn't.

That said, the clues are right in front of me, I just don't see the answer yet. At the same time, I'm trying to avoid ending up with $2000 worth of new spare parts sitting on the shelf.

Are you sure you're running 240kpa? That's a ton of boost for "only" 400hp

Jay_W wrote: If you can rule out an ignition issue, which is where I'd look first, the next thing is boost leak. Since you swapped turbos, and you haven't mentioned an external wastegate, we rule that out. Some wierd blowoff valve vibration related leak that happens only at high RPM? Doubtful. A hole in the plumbing that only opens at 7500? Let's go back to ignition. I see your spark timing line thinks it's doing everything right, and knock sensor signals all show zero, but wonder if ECU is pulling timing for some other reason? We chased this problem to a ground loop that somehow interfered with the knock sensor signal... something wrong on an adjacent pin IIRC. I kinda doubt it's as easy as changing coils. If you have a coil laying down, it's kinda hard not to notice the whole misfire thing, unless you're getting real lucky in the sense that it's still putting out spark, but weak, so bad burn and lower output and lower boost but no discernible misfire. Either way, bad ignition should show up on the AFR data, sudden rich condition at 7500. Which would make ECU pull fuel. I don't see an AFR trace. What's it do I wonder? The only other thing off top o' my head would be some vtec problem where (assuming you've assigned such duties to the ecu) the cam timing is going wrong for whatever reason but I know nuthin bout such things.

In order of your questions:

I previously had two external wastegates (it's a twin-scroll setup). Earlier during troubleshooting I got it into my head that perhaps the combination of MAP and exhaust back pressure was pushing the wastegate valves open. At the time I was running a 3-port boost controller, which cannot control the above condition, but a 4-port can actively hold the wastegate closed. Also, the existing wastegate diaphragms leaked slightly, so I wondered if my trackday antics had perhaps fried them.

For all the above reasons I switched to a new watercooled wastegate... but surprisingly, the same problem occurs even when the ECU tries propping up boost.

Knock detect is currently disabled so it has no effect on this. (Ironically that's when I discovered this problem, while I was doing runs to establish knock noise levels in anticipation of enabling knock detect and timing retard).

Good point about a misfire showing up as a suddenly rich mixture, which doesn't appear to be happening. (Lambda is logged, it's the bright green trace in the bottom plot.)

Yeah... I don't know a lot about the VTEC timing. My concern is that if it is that, it could be a tough nut to crack, as while the cam position command can be logged, I highly doubt that actual cam position is provided. Looking at the command doesn't always tell us where the cam actually is if things are misbehaving. I'm going to have to read up on how that system works.

Well I have no doubt you'll find it eventually, and no doubt that you'll kick yerself fer not seein it earlier. That is the nature of krap like this, ain't it..

DaveEstey wrote: Are you sure you're running 240kpa? That's a ton of boost for "only" 400hp

Yes and no. Remember that kpa is absolute, so 101 kpa is ambient. That means that the engine's actually seeing "140 kpa boost", or about 20 psi.

Also (I left this part out for simplicity) I have no idea how much power the engine actually makes. It's been run on three dynos at different times, two DynaPack dynos and one was the more typical Mustang drum/inertia type. Running about 13 psi boost, the DynaPack reported "410-440 whp" on gas, yet the drum-type dyno reported about 320 whp. On E85, the DynaPack reported ~530 hp at ~20 lbs. So, power is somewhere between 400 to 500 hp, depending upon the dyno*, but it's to the point that - and I never thought I'd say this - I don't want any more. Traction in 3rd gear is "iffy" on the street, and that's good to well over 100 mph. It's mid-engine, so there's a heck of a lot of traction, but it's pretty much at the limit. I have a LOT of respect for this machine and it's probably why I'm still around. I pretty much drive like an old man except for rare instances when conditions are deemed safe (no one around, no passing, no passengers, etc.)

I should also note that I'd like to keep the turbo AR as small as possible. You can see how nice and broad the torque curve is and I'd like to keep it that way. Increasing the AR might fix the issue yet both shift the torque curve higher and make it more peaky. If someone told me that the AR is definitely the problem, Right now, the not knowing part is driving me nuts. Once that's solved THEN I can decide what to do about it.

Lastly, and this is interesting, in this plot (done during the dyno tune) boost does Not drop off (220 kpa solid purple line). I don't know if the anomaly doesn't showing up on the dyno (I'm pretty sure that it existed before the dyno run) or whether something has broken since then, or if it's something I did to the tune (though again, it did it previously with a different ECU).

• I find it ridiculous that in this day and age of six digits past the decimal, two different dynos can differ by 20%... sheez. Granted, they're intended for comparative tuning, but still.

K20 and S2000 close to the same

k20 on 14psi on a 30r is 400+ whp

the last S2000 i wrenched on turbo with a 30R @16-18psi hit 518whp

K24 should be even more. Geez

What are the narrowest pipe diameters in your intake pre and post the compressor.

I am also wondering if there I something in the ecu that is kicking I and doing somthing to valve timing. A safeguard of some sort.

Reach out to Steve Delacruz https://m.facebook.com/pages/Delacruz-Motorsports/291607480951268?fref=ts

DaveEstey wrote: K20 and S2000 close to the same k20 on 14psi on a 30r is 400+ whp the last S2000 i wrenched on turbo with a 30R @16-18psi hit 518whp K24 should be even more. Geez

Read my post above in response to DaveEstey, I just edited it to add more detail about hp produced on different dynos.

dean1484 wrote: What are the narrowest pipe diameters in your intake pre and post the compressor.

4" on the air filter side and 2.5" into and out of the intercooler.

dean1484 wrote: I am also wondering if there I something in the ecu that is kicking I and doing somthing to valve timing. A safeguard of some sort.

If I hadn't changed ECUs, I'd be heading down that path. Since the problem is happening with both the old and new ECU however, it seems unlikely. That said, I did go through every single table and limit to check and found nothing. I also sent the tune to the ECU manufacturer and they confirmed that there weren't any limits anywhere near 7500.

I think you nailed it with the wastegate being pushed open by exhaust back pressure. Is there a convenient place you can tap with a pressure gauge pre-turbine? I measured back pressure when I did my VGT Mercedes Diesel and it's shocking how much back pressure there can be. I bet the measurement will show it dropping at the same time as boost. Compare your peak reading to the area of the wastegate valve and spring force. Or you could just shim or swap the existing wastegate spring with a stiffer one and see if the problem goes away.