Custom intercooler

[Thom]

Yes it's fast, but I don't think I'll comment anymore on the performance as I have been running out of descriptive vocabulary.

It is quite puzzling though that the compressor might be too small for that sort of boost level as folks with 2JZ-GTE engines seem to get serious numbers, north of 600rwhp, and I don't think I'm there, more likely around 550?
A stock 2JZ-GTE head seems to flow far less at peak valve lift with stock cams than I expected as I find figures of ~225 cfm @28", which I think is pretty miserable for a 4V head, the 968's flowing a whale-esque 330 cfm in comparison.
225x6 = 1350 cfm
330x4 = 1320 cfm
For one engine cycle the total flow is similar between both of these 3L engines, fwiw, but so far it makes sense to me that the compressor is not supplying the head with all the air it wants to keep making power up top. Not a real world issue though considering the massive mid range thrust but I can feel again that slippery slope with the tip of my toe... sigh.

[Bergerac]

No way you're getting anywhere near 39mm intake valves in a 86mm bore

[Boostfeen]

Did this last year and works great it was a speedforce eBay purchase used and while I first cut the bumper bar for in front of box frame I ended up cutting the box frame since it’s a full race car and mounted to protect from front bumps and bruises as it’s a endurance racecar.

2.5 16 valve turbo motor and makes 300ish hp

[Boostfeen]

Intake with a s2000 throttle body and custom cable

[Thom]

Good job. That should flow pretty well for a 2.5 engine.

Borg Warner's matchbot supports my impression that the compressor may be running out of breathe above 6,300 rpm when running more than 23 psi. The way the engine runs out of power abruptly is different from what I experienced with the 8V engine with the KKK hotside and suggests to me that the cause is not a sudden sharp rise in back pressure.

Any opinion for a replacement turbo? The GTX3582R Gen 1 would drop right in, or fork out for a recent G35-900 which will flow a little more than the GTX3582R but should respond as well as the GTX3576R, but require modifying the exhaust to fit the new hotside?

This is probably a holier than thou approach as the car really is fast enough by now. It does not need to be revved beyond 6k rpm at all to go ballistic and I'm not sure of what I'll be doing of even more power, to be honest, especially when 2nd rod bearing failure occurs when spending too much time above 6k rpm with the stock wet sump oiling system.

From another perspective I specced the engine around the turbo which seems to have exactly the room needed to display all of its performance, and that the head seems to "out breathe" the turbo at pretty much the end of the "reasonable" [RPM; boost] window should probably be considered a reasonable achievement, like there is virtually no "waste" of head flow or turbo in the whole set up.

ETA : apologies for sounding like a broken record as I realise I keep repeating the same stuff page after page, hopefully only because I am not "on it" as much as many years ago.

[Boostfeen]

Thanks Thom got some inspiration from your work at beginning and along the way. I built two of these 2.5 S motors at the same time that had great bores and pistons so I had the piston tops milled 1mm down and then coated from swaintech. Mild head work and new Audi lifters, stage 1 springs and I cut the Windage ports in blocks. All fresh bearings, trap door in pan, new oil pickups with sleeve and the Mishi support silver soldered on pickup tube. Then the largerest off the shelf Cometic head gasket. I believe we are at the top of 8 maybe right at 9-1 compression. With the longevity of motor in mind for endurance racing I’m using a slightly modified k26/8 and free intake and exhaust. We are limiting boost with Tial 46mm down to 10-12 psi. Electromotive Tec GT handles the engine with elec boost control and 300hp happens rather easy with lots of pre boost power to pump this turbo up!

[Thom]

The Borg Warner match bot is fun to play with and the calculated figures are very similar to what the butt dyno has registered, as the GTX3576R compressor reaches its peak flow of ~64 lbs/min at :
- 7000 rpm @18 psi
- 6300 rpm @23 psi
- 5900 rpm @26 psi
- 5400 rpm @30 psi

If we plot the GTX3582R Gen 1 its peak flow of ~75 lbs/min is reached at :
- 8500 rpm @18 psi
- 7500 rpm @23 psi
- 6900 rpm @26 psi
- 6300 rpm @30psi

... and the G35-900, flowing a max of ~82 lbs/min at :
- 9300 rpm @18 psi
- 8100 rpm @23 psi
- 7500 rpm @26 psi
- 6900 rpm @30 psi

This sort of explains why there was no perceived loss of power up to 7000 rpm when I initially ran the engine at 18 psi.
Due to oiling issues we probably want to avoid holding engine speed above 6000 rpm for "too long", which makes the GTX3576R still a very good choice @23 psi even I feel I have been maxing it out @23+ psi. Making the most of the GTX3582R would require running it between 26 and 30 psi, whilst the G35-900 will most likely not reach its peak flow at any pressure but run much closer to its peak efficiency area and at lower shaft speeds yet may spool as quickly as the GTX3576R as they use same size wheels. Again this is all assuming that the VE, BSFC and the various lengths of pieces of string as entered by default in the match bot are any relevant to a stock 968 top end.

However, let's consider the shapes of the efficiency islands for these 3 turbos, especially the right side of the respective maps.

GTX3576R :



With pressure ratios of 2.5 to 3.5 (23 psi < boost < 36 psi) the peak flow hardly varies and the GTX3576R will not flow any less than 62 lbs/min at most. That peak flow is delivered pretty much equally over such a wide range of pressure ratio is I think very good.

GTX3582R :



Here again the right side of the efficiency island of peak flow is remarkably "vertical" and the GTX3582R will flow near/at its peak over a wide pressure ratio range, say between 2.5 (23 psi) and 3.25 (33 psi).

G35-900 :



Here the right area of the map has a pointy edge and peak flow is reached at a pressure ratio of 2.4 (20 psi). We are still over 75 lbs/min over a pressure ratio ranging from 2.1 (16 psi) to 3.1 (30 psi), but as calculated above this is not reasonably happening within the rpm range of an engine like mine, but that's ok as there should be plenty of air flowing through below 6000 rpm already.

For what it's worth anyway, depending on how hot or cold air feels with the finger up.

[333pg333]

Great post Thom. Just for curiosity, I wonder what the next size frame up would look like? Say the G40-900.

[Thom]

Well, I can only comment based on what I think correlates more or less accurately between my set up (with the conservative stock cams) and what the match bot says, but I am pretty sure that a G40-900 would be way too large. The compressor map looks very good, better than the G35-900's above but the turbine would be way too big, even with the smaller housing. It seems to me it would work best on "large-ish" engines using "small-ish" cams... I just don't know.

[gruhsy]

Thom have you thought about twin scroll on any of the turbos?