I don't know anything about the how and why side of things, but I'm pretty sure the results can be seen in things like this dyno.

https://scontent-ord1-1.xx.fbcdn.net...06&oe=57386FA0

That's a 100% factory 1.8L engine. It's not mine, but I have the same engine in my winter beater. As Matt said, it drives like a big v6. When I'm just putting around in traffic I rarely take it above 2000 rpm, there's just no reason to most of the time.

Here's a link to the thread where I found that dyno. http://www.golfmk7.com/forums/showthread.php?t=16359

The dyno plot is very informative, in a proof-of-concept kind of way.

I'd like to see a lot more data to get to how and why, not least of which being the stock fuel and timing tables. And perhaps a little peek at the GDI timing.

Also like to know what turbo it's running, and do a little compressor map plot there.

Looking at the power curve, it looks like a tiny little turbo that's hitting the choke line around 5k RPM.

The problem with making big boost down low with a big turbo is you're way left of the surge line in many traditional cases.

In a modern, less traditional case... With lots of overlap, in the form of leaving the exhaust cam alone and tossing in serious intake cam advance, and a modern turbo that runs an intake/exhaust pressure ratio greater than 1, I could see some great results happening, even with a huge turbo.

Running in direct injection only, hitting injection timing right around the end of overlap...

I could see that working.

At lambda 1.3, with vaporization absorbing lots of latent heat, I could even see running reasonable timing... Provided Lambda 1.3 is lean of peak EGT.

I know enough to smile and nod along with what you're saying, I understand it, but I don't have anything else to add. The 1.3 Lambda is scary, but if you and Matt are saying it wouldn't cause the EGT to increase too much, then I'll agree.

As for the turbo, it's an IS12, which, as you expected, is a tiny little turbo. Here's what that engine's dyno looks like with just a tune. This is APR's 93 octane numbers for at the wheels. Their numbers seem to be consistent with third party dyno's, but take it for what it's worth. I'm only adding this as you might be able to glean more about the turbo's size and flow rates from this.

http://www.goapr.com/includes/img/pr...s_s1_93_cw.png

Thanks for that. I expect the low-end of the torque graph looks like a perfect mirror of the boost curve. But this isn't crazy unconventional turbo tuning. It's just a tiny turbo. I haven't seen the compressor map, but I bet money that if you graphed it, that kneepoint in the torque curve, the middle of which is coincident with peak power, is right at the choke line on the compressor map.

Just to explain the thing about EGT's a little better, if you graph EGT with lambda on the x axis, and EGT on the Y-axis, it will approximate an upside-down parabola, with a peak just a little bit lean of stoichiometric. In a perfect world it would be max right at stoic, but there's a practical reason that it doesn't happen in the real world; The mixture is not perfectly homogeneous no matter what, and any unburned HC's in a given volume will absorb more heat than the same volume of air would.

1.3 Lambda is like 19.1:1. EGT's hit peak and start to drop quickly when you go leaner than about 14.9 ~ 15:1.

Up at 19:1, you are running way lean of peak EGT. All other things held equal, you'd observe EGT's similar to those found in the 10's.

There's associated problems, and I wouldn't do anything like this without EGT monitoring on all 4 cylinders.

Anyway, more general principle stuff, old turbos used to make more "backpressure" than boost (I still hate that term), and therefore overlap was a bad thing. New turbos are cool because in many cases, boost pressure is higher than the pressure in the turbo manifold between the exhaust valve and the turbine inlet. (the... ahem... backpressure, if you will...) This is cool because, in theory, some overlap won't rob any turbo spool, it can actually help to increase it. In theory. Until you add port injectors. Because with a port injected engine, the mixing of exhaust and intake air will vaporize fuel, which will absorb latent heat, which will cool your exhaust gas, and invoke the pesky laws of thermodynamics, and steal your boost. Fuel vapor is for burning, not for spinning turbos with.

Mixing it with plain old cool intake air (without any fuel in it), however, is close enough to an isentropic process that it doesn't affect boost. The overall temperature will decrease in inverse proportion to the total mass of air, and the energy in the system will remain constant. The turbine could care less how hot the exhaust is. The turbine only cares that PV=nRT.

Enter direct injection. Yay! Now our intake charge need not have any fuel in it until the exhaust valve is closed, and we're sure we aren't going to waste any fuel on anything besides burning oxygen.

The tuning possibilities are endless.

If this were Reddit I'd give you gold. Thanks for the easy to read explanation!

It's been 13 years since my thermal dynamics classes and prime mover classes, none of which went this in-depth. You wouldn't happen to know of a good resource for more on this?

So what's the purpose of running lean? Is it to increase the EGT to provide more pressure for the turbo? Is the downside of running lean increased cylinder temperature due to lack of evaporative cooling?

If I'm understanding this correctly then would it possible to run variable lambda? Run lambda>1 to spool up the turbo but then switch to lambda<1 on certain compression strokes to prevent heat soak in the cylinder... I may be way off on this but the topic is pretty fascinating.

Missed the mark a bit.

I'll give you 3 scenarios, based on the following conditions:

Assume we are trying to make lots of torque at low RPM with a small-displacement engine and a large turbo.

In order to achieve this objective, engine load (mass of air ingested per rev) must be very high. This means lots of boost.

This is a very high compression, direct injected engine.

We need lots of air on the exhaust side to spool the turbo, and the intake side has to have a very high flow, to support the very high pressure ratio, otherwise you get compressor stall.

1) Low speed, very high load, 14.7:1 AFR. EGT's are catastrophically high, detonation is severe, everything that doesn't simply break, instead just melts.

2) Same conditions, 10:1 AFR. We have relatively high combustion temps, but everything runs fine. Lost of fuel is wasted overboard in the form of unburned vapors. Except we need to flow lots of air. So, we run lots of camshaft overlap, thereby introducing lots of fresh oxygen to lots of incredibly hot fuel vapor and metal. the otherwise unburned hydrocarbons continue to burn as a post-combustion event, EGT's rise, everything melts. That, or we get awesome turbo spool and nothing melts... I dunno.

3) Same conditions, 19:1 AFR. EGT's are identical to when we ran 10:1, except when we introduce camshaft overlap, no post combustion occurs. EGT's actually fall, but they do so at inverse proportion to the increasing mass of air in the exhaust, because no work is extracted from the hot gas, only an isenthalpic temperature exchange. The turbo spools much better than a conventional old-fashioned engine (due to 2 reasons I explained above), we can run lots of timing and make lots of torque, and nothing melts.

Let me again explain why the turbo spools better than conventional turbos in conventional engines. Two reasons.

1) In modern turbos, the intake pressure ratio is higher than the exhaust pressure ratio; in other words, the compressor actually helps to spin the turbine, rather than revert back upstream. Kind of like a jet engine. (in a very simple sense) Jet engines are just turbos for cool people...

2) Statement 1 doesn't work in conventional port-injected engines, unfortunately. This is because, in port-injected engines, the intake charge is contaminated with fuel. The nice, hot exhaust gases do a great job of heating and turning that fuel mist to vapor. That, however, is work. We are extracting that latent heat and doing work with it. Unfortunately, the pesky laws of thermodynamics say that if you use the energy in a system to do work, like vaporizing fuel, you can't use it to do other work, like spinning a turbo. I'm sure you could spin a turbo with fuel vapor if you shot enough at it, but unfortunately, the change of state costs more energy than can be recovered by the turbo.

Like I said, fuel is for burning, not for spinning turbos with.

Well, I'm certainly no expert, and my thermo experience is mostly with steam plants, but reading up a bit about anti-lag systems seems like there are two main ways to achieve it, but both involve fuel detonation post engine. The wiki page has a decent layman's summary.

This page led me to Brewed Motorsports, who markets anti-lag systems for turbo cars. They say their throttle bypass tuning is good enough to not melt your block, so it seems they must not be dumping as much fuel in like Matt@Cosworth suggested OEMs are doing.

No.

The brewed stuff is just standard traditioal anti lag. But it helps explain something.

Matt was discussing non-traditional anti lag, not possible without direct injection.

Imagine the operating principle of SAS ALS if it was achievable with tuning only, and without secondary air bypass hardware.

That is what is being proposed.

Read my posts carefully.

Basically, with secondary air, you are running pig-rich and shooting fresh air into the exhaust to combust all of the excess unburned HC's when you are off-throttle.

Imagine the blowoff valve, instead of venting to atmosphere or recirculating, instead blows off into the exhaust manifold.

That's traditional rally anti-lag.

The traditional throttle bypass system gets around the unburned fuel saturation problem by retarding the timing so much that the combustion event is still occurring during the valve events, and any fuel-air mixture that is ingested during valve overlap simply burns. This is traditional "tuning only" anti lag, and it melts sh*t. It melts lots of sh*t.

Now, imagine the air ingested during valve overlap had no fuel in it, because you have a direct-injected engine, and you could choose to introduce fuel after your exhaust valve event had ended. You reap the benefits of secondary air injection without secondary air hardware, and since you don't need to back the timing off to induce post-combustion, you can run enough timing to make useful torque..

thats basically it

sounds like no-one is yet exploring this in the aftermarket world
shame
you would need an EGT measurement to make a good job of it but other than that I see no reason why an off the shelf turbo system on ECUtek couldn't be running this

so who's 1st?

I'd love to see what Cosworth could do with a turbo, a reworked FA20, and this kind of tuning philosophy. But my god I'd hate to see the price. I have a lot of respect for what you guys do, but my pockets aren't deep enough to show that support. :'|

Well my wife's 2015 WRX would love to have a Cosworth tune. :)