Report: BRZ STI Model Already in the Works, No Turbocharger

[tranzformer]

Originally Posted by serialk11r

Not really, a twin scroll turbo separates the cylinder pulses and is thus more efficient but at the end of the day the "total exhaust energy" delivered to the turbine for each given level of constant load (by this I am referring to the desired load given by accelerator pedal position that the ECU tries to interpret) is completely non-linear, and even a slight amount of lag makes it worse.

But response aside (which some people don't care about, and that's fine), a twin scroll turbo doesn't change the fact that the turbo can only be optimized for a narrow range. You need a variable geometry turbo to fix the high end vs. low end tradeoff.

The Hyundai Sonata 2.0T that I test drove was damn near perfect in power delivery. And that is for a grocery getter. I bet Subaru could do better. I am on my mobile device so can't post, but inside line compared the Sonata turbo vs NA engine (with dyno results). The turbo has grear power throughout.

[madfast]

Originally Posted by Dimman

Is it the different EGTs that limit these to being common on diesels but $$$ on gasoline cars?

yes, basically. i believe porsche/borg warner use proprietary materials and that's why they're the only ones who use variable vanes. and even then, there are stories of the vanes failing...

[tachi1247]

Originally Posted by tranzformer

The Hyundai Sonata 2.0T that I test drove was damn near perfect in power delivery. And that is for a grocery getter. I bet Subaru could do better. I am on my mobile device so can't post, but inside line compared the Sonata turbo vs NA engine (with dyno results). The turbo has grear power throughout.

+1. Bmw has been using the twin scroll turbo on the 335 for several years now and that car has almost no lag and makes great consistent power and torque across most of its rev range.

I think most aftermarket turbo setups are affected by the narrow range problem as they are usually targeted at a specific purpose and are limited by what the factory setup on the car is. On a factory turbo charged car, if something isn't the way they want it they can just move it unlike in the aftermarkt.

[Dimman]

Originally Posted by madfast

yes, basically. i believe porsche/borg warner use proprietary materials and that's why they're the only ones who use variable vanes. and even then, there are stories of the vanes failing...

This thing is a Garrett VNTxx. My co-worker's TDI threw a rod at only 18x,xxx kms and he said something about replacing his turbo with a bigger one.

I heard "VNT" and offered him $50 for the old one, wanting to see what a Garrett variable geometry system was/how it worked. Didn't know the damned thing was attached to the manifold...

My understanding is that Garrett's is only a single big vane and is the least efficient (lots of turbulence) of the variable geometry setups.

Second best I've heard is the Mitsubishi which is similar but two vanes that open sequentially and 'backwards' compared to Garrett's VNT. Supposed to be very efficient and very reliable. Is this on USDM Evos? It is apparently on certain JDM models.

Porsche's multi-vane is supposed to be the best performing.

The twin-scroll thing puzzles me as to why it is even a buzz-word. My understanding is that it is just a twin-entry divided turbine, the same as Toyota was using with the CT26 ~25 years ago, or am I missing something?

[serialk11r]

Dimman you're correct, twin scroll means divided turbine housing nothing more.

There's a difference between "consistent power across the rev range" and "best power at high rpm". The reason BMW is able to achieve high boost at low rpm is that their turbo had low-midrange focus, since they don't use variable vanes (or do they?). At any rate, just looking at a dyno chart you can see that the torque is flat probably because the boost is being limited, and the dying high end power points to that conclusion as well.

[madfast]

Originally Posted by Dimman

This thing is a Garrett VNTxx. My co-worker's TDI threw a rod at only 18x,xxx kms and he said something about replacing his turbo with a bigger one.

I heard "VNT" and offered him $50 for the old one, wanting to see what a Garrett variable geometry system was/how it worked. Didn't know the damned thing was attached to the manifold...

My understanding is that Garrett's is only a single big vane and is the least efficient (lots of turbulence) of the variable geometry setups.

Second best I've heard is the Mitsubishi which is similar but two vanes that open sequentially and 'backwards' compared to Garrett's VNT. Supposed to be very efficient and very reliable. Is this on USDM Evos? It is apparently on certain JDM models.

Porsche's multi-vane is supposed to be the best performing.

The twin-scroll thing puzzles me as to why it is even a buzz-word. My understanding is that it is just a twin-entry divided turbine, the same as Toyota was using with the CT26 ~25 years ago, or am I missing something?

modern diesels mainly use multiple vanes. afaik nobody except porsche use variable vanes in gas turbos. not because of a new design, but the proprietary materials/construction of the vanes, which can survive the temps.

the honorable mention, in recent memory, goes to the Acura RDX K23 which used a variable turbine housing, but not variable vanes a la porsche.

as for twin scroll, yeah thats nothing new as well. but companies like BMW try to sell it like it is. they went from twin turbo N54 to the "twin power" N55 single twin scroll turbo.

[Dimman]

RDX system reminds me of the 'DIY variable geometry' of the Quickspool valves that a few Supra guys have been using for a while. Uses a twin scroll turbine but manifold is set up as single entry, but a butterfly valve is on one of the scroll entries...

[Levi]

Originally Posted by tachi1247

+1. Bmw has been using the twin scroll turbo on the 335 for several years now and that car has almost no lag and makes great consistent power and torque across most of its rev range.

Drive the M3 E92 and then drive the 335i E92. You'll see how much lag it has. And unlike what people say, the S65 has a very flat torque curve with redline at 8.400 RPM.

[DoogenBurns]

Originally Posted by Dimman

This thing is a Garrett VNTxx. My co-worker's TDI threw a rod at only 18x,xxx kms and he said something about replacing his turbo with a bigger one.

I heard "VNT" and offered him $50 for the old one, wanting to see what a Garrett variable geometry system was/how it worked. Didn't know the damned thing was attached to the manifold...

My understanding is that Garrett's is only a single big vane and is the least efficient (lots of turbulence) of the variable geometry setups.

Second best I've heard is the Mitsubishi which is similar but two vanes that open sequentially and 'backwards' compared to Garrett's VNT. Supposed to be very efficient and very reliable. Is this on USDM Evos? It is apparently on certain JDM models.

Porsche's multi-vane is supposed to be the best performing.

The twin-scroll thing puzzles me as to why it is even a buzz-word. My understanding is that it is just a twin-entry divided turbine, the same as Toyota was using with the CT26 ~25 years ago, or am I missing something?

Twin Scroll turbos just keep exhaust slugs separated to prevent overlapping, and back pressure.

[MSTiFK8R]

Tomorrow I will bee meeting & interviewing the President of STI

Hiroyuki Karamatsu -san





I will try to get some details of the STI BRZ, which he would be able to unveil

"Wheeers the terrrrrbo?" - is definetely on top of my questions )

[F3dzo]

Booooooooooo

[serialk11r]

Originally Posted by Dimman

RDX system reminds me of the 'DIY variable geometry' of the Quickspool valves that a few Supra guys have been using for a while. Uses a twin scroll turbine but manifold is set up as single entry, but a butterfly valve is on one of the scroll entries...

That sounds like it's just an easier way to get "variable geometry" than having movable vanes, but the divided housing is useless if the exhaust pipes all merge together before it.

[Dimman]

Originally Posted by serialk11r

That sounds like it's just an easier way to get "variable geometry" than having movable vanes, but the divided housing is useless if the exhaust pipes all merge together before it.

If the valve is sort of angled so when it's closed it guides the exhaust into the open side it is ok as the entry to the turbo is smaller.

It will have flow losses compared to either a fully open single or twin, but they do improve spool.

Remember this is a 'DIY'-type solution, not from major turbo companies.

[Dimman]

Originally Posted by MSTiFK8R

Tomorrow I will bee meeting & interviewing the President of STI

Hiroyuki Karamatsu -san





I will try to get some details of the STI BRZ, which he would be able to unveil

"Wheeers the terrrrrbo?" - is definetely on top of my questions )

Turbo ga, doko desu ka?