Turbine size matters

[jgkurz]

36 minutes ago, rossman said:

I certainly would appreciate it!  Do you know if going from .63 to .82 moves the exhaust v-band relative to the T3 flange?  I assume it does.  If so I may have to modify my down/wastegate pipe when I eventually get there.

 

Yes, the downpipe needed to be completely re-fabricated. Like you, I hoped it would be a simple flange change but it was not in my case.

[jgkurz]

Hi all, I finally got time to install the new .82 housing and dyno test. To close out this thread I wanted to post my results. The new Garrett GTX3576R finally was able to produce 467hp and 496ftlb on a DynoJet chassis dyno. The .63 housing maxed out at 404hp and just under 400ftlb. Boost was 23lbs. The boost lag worsened a fair amount with the larger housing. The boost did come up far smoother which was a positive effect. Also, the old turbo with the small turbine housing use to come up on boost quick which was fun, but it would also reek havoc on clutches. 

 

Here's the interesting thing about my upgrade. The old T3/4 hybrid made 451hp and 467ftlb at 27psi WITH a .63 turbine housing and quick spool. My new EXPENSIVE turbo did make more power at a lower boost, but was it worth it over a traditional T3/4 journal bearing? Not really. I started with a .63 housing on new GTX3576R, but it caused 10psi of back pressure and only made 402hp. The old turbo made 451hp with a .63 turbine. The 16hp gain with the GTX3576R wasn't worth the trouble given the boost lag. My cylinder head and intake are near stock which is certainly limiting my engine. Notice the peak power is at 5400rpm. All-in-all, 550hp at the crank is pretty good for an old L28.

 

 

 

 

 

Edited April 24, 2019 by jgkurz

[bradyzq]

Not worth it? You're hard to please!! Notwithstanding the mystery issues on the GTX .63 hotside, you now have virtually the same boost threshold as your old turbo with a huge, almost 40 lb*ft gain, all while running 4 psi less boost!

 

To compare apples to apples as much as you can, you really should run 27psi on your new setup and see what happens. I bet the difference will be huge.

 

BTW, journal vs. ball bearing differences are not that huge in terms of absolute power. Transient response should be better though, especially coupled with the more modern wheel design.

[jgkurz]

43 minutes ago, bradyzq said:

Not worth it? You're hard to please!! Notwithstanding the mystery issues on the GTX .63 hotside, you now have virtually the same boost threshold as your old turbo with a huge, almost 40 lb*ft gain, all while running 4 psi less boost!

 

>>Yeah, more power on less boost is always good but the cost to do this project was crazy expensive. Having a part custom CNC'd on top of an expensive BB turbo added up to big $.  The $/hp just didn't work out well. To some degree that is the nature of any car project. I like the car and it is fast so it's all good. 

 

>>I think the GTX hotside issue can be attributed to the Tial housing. Their .63 visually looked smaller than my T3/4 .63 housing. I should have taken that cue. 

 

 

Quote

 

To compare apples to apples as much as you can, you really should run 27psi on your new setup and see what happens. I bet the difference will be huge.

 

>> I may test 27psi some day, but I will need to beef up my valve springs. My Crane springs are 90# at the seat so only 63# if I run 27psi. I think my peak RPM of 5400rpm may be caused by weak valve springs. That and a near stock head and N42 intake.

 

Quote

 

BTW, journal vs. ball bearing differences are not that huge in terms of absolute power. Transient response should be better though, especially coupled with the more modern wheel design.

 

 

>>I have learned this as well. A journal turbo from a quality manufacturer is still a great option.  Thanks for all your comments!

Edited April 26, 2019 by jgkurz

[bradyzq]

You can always throw the old turbo back on and run it at 23psi and see what it does.

 

You know, in case you want to spend more time and money on this in the name of science...

[Chickenman]

One of the interesting things the Audi BT guys are doing is putting a 4" expansion section on the downpipe. This really results in a drop in Exhaust back pressure.   Quicker spool and higher WHP. They are transverse mount, so limited by space.  But even a shortish ( 12" ) expansion pipe made a noticeable difference.

 

I suspect a 4" downpipe on an L-6 then taper to 3" or 3.5" exhaust would certainly help. 

[240Z Turbo]

On 4/21/2019 at 11:53 PM, jgkurz said:

Hi all, I finally got time to install the new .82 housing and dyno test. To close out this thread I wanted to post my results. The new Garrett GTX3576R finally was able to produce 467hp and 496ftlb on a DynoJet chassis dyno. The .63 housing maxed out at 404hp and just under 400ftlb. Boost was 23lbs. The boost lag worsened a fair amount with the larger housing. The boost did come up far smoother which was a positive effect. Also, the old turbo with the small turbine housing use to come up on boost quick which was fun, but it would also reek havoc on clutches. 

 

Here's the interesting thing about my upgrade. The old T3/4 hybrid made 451hp and 467ftlb at 27psi WITH a .63 turbine housing and quick spool. My new EXPENSIVE turbo did make more power at a lower boost, but was it worth it over a traditional T3/4 journal bearing? Not really. I started with a .63 housing on new GTX3576R, but it caused 10psi of back pressure and only made 402hp. The old turbo made 451hp with a .63 turbine. The 16hp gain with the GTX3576R wasn't worth the trouble given the boost lag. My cylinder head and intake are near stock which is certainly limiting my engine. Notice the peak power is at 5400rpm. All-in-all, 550hp at the crank is pretty good for an old L28. 

 

I know it is easier said than done, but you need to think about having someone fabricate a twinscroll manifold for your setup.  The open scroll is great for a drag car, but in my experience it sucks for a street car because of spool.  Below is an example on my EVO from back in 2011 when I was running a GTX3076r.  I had a cast manifold using a T3 open scroll housing, 0.63 vs 0.82, and then I fabricated a twin scroll manifold and went to a1.06 a/r T4 twin scroll.  The twin scroll crushed the open scroll in spool and made the same power on the topend, but killed the open scroll torque everywhere.  I'll post a link below my new setup going from a GTX3582r (6262) to a BW S362 SX-E (6268) both T4 twin scroll.  You'll be shocked the car spooled the same (on the street) and picked up 50hp@wheels...601hp@wheels on Shell 93 pump gas at 33psi!

 

https://www.evolutionm.net/forums/evo-dyno-tuning-results/754630-gen-i-gtx3582r-vs-s362-sx-e-curt-brown-2-15l-pump-gas.html

 

[240Z Turbo]

GEN I GTX3582R vs S362 SX-E spool up on the street.  The dyno does not load the engine as much so that affects the spool of the 68mm turbine wheel in the S362, but on the street I have countless logs showing it spools the same.

 

 

 

Edited June 22, 2019 by 240Z Turbo