I have found the complete opposite when it came to sub litre applications in several different platforms from Hondas B18C and K20, to A853 SRT-4s. The fact that the average T4 flange system uses a turbine wheel larger than that of most of the T3s, even still, the documentation seen doesn't show the correct Boost /rpm data that shows the improvement on a dyno, nor (more importantly) have I experienced that on the circuit.

Again, I've found that with EFRs, that I've used and examine, they go well beyond the 90 degree point, to where the improvement seems to only lie with that particular change; that, and the turbine wheel itself has to be identical to see the real differences. Again, my experience comes mainly from circuit racing both the EFR, GTX and GT-R turbos that stay within a relatively decent power level of about 550whp and less. When those that used a divided smaller T4 .68 and 1.03 housing compared to the T3 undivided, utilizing the same Turbine exhaust wheel (such as the GT35R's N32 68mm exhaust wheel), I saw no improvement, and in fact, less improvement in the overall powerband. Sorry.. over the last 13 years working with Honeywell and Garrett, I never saw any "improvements". I just didn't see it.

Actual twin scroll improvements (for smaller turbine wheels) = True
Divided housings that are T4 based = Fiction

 

I'm sorry, but this claim conflicts with dozens of first hand observations reported here (including my own) in going from a .63 open T3 to a 1.00 or larger divided T4. No wonder you find this interesting. Either your assertion is inaccurate, or we have dozens of inexplicable mysteries on our hands.

 

The Shodan,

Please to explain why I found substantial increase in spool with a 1.06 T4 over my 0.82 on a Non MIVEC motor vs my MIVEC motor?

https://www.evolutionm.net/forums/ev...l-hta3582.html

To wit-

 

So here is an interesting comparison since it is more Apples to apples than an VIII to a IX.

Following cars are both the same as regards -

Evo VIII
Stock 2.0L
Stock port head
Stock Intake manifold
Kelford 280/276
RC1000s
SD tune
3" exhaust
44m WG
35R turbo kit



The differences-

Blue is a 0.63 T3 single scroll and by classical definition is nearing backpressure issues based on the torque fall off (not logged though).

Red is the same 1.06 T4 as before.

This fits the Garrett white paper very close though, Divided/Twin scroll spools like 60% its size. In this case 1.06 x 0.6 = 0.636 A/r

 

^ Those are nearly identical. The power difference could just be a difference between cars or weather. The torque falls off on the T4 the same as it does the T3. And the T3 spool'd faster (and was started later).

If that is trying to show T4 TS as being better, it really doesnt.

 

Even if the TS setup didn't deliver an additional 20ft/lbs and 35whp, even if the dyno charts were identical (and they aren't), the difference from behind the wheel is plenty obvious. We've said this before, but a dyno chart represents only a static run in a single gear. It doesn't illustrate the manner in which TS quickens transient response. That's the first difference one notices with TS in real world driving.

 

Well the curves are the same. So if the T3 is "falling off" so is the T4.

Also, those charts are uncorrected. Without weather data, we dont know if the tq/hp difference was because of setup or weather or just car to car variance (since this is still not apple to apples, since its two different cars).

Yes, dyno charts dont show where TS T4 shines. I never said they did. I just said if that chart was to "prove" TS T4 is better then OS T3, it does a horrible job of it.

I had a car w/ a GTX3576 that had OS T3 .82 and switched to a TS T4 1.06. Only change. While the car had better off boost and transient response, it spool'd the same, made the same power and made the car more knock prone on 91oct. I have charts/logs of all this if I ever remember to bring my tuning laptop home from the shop.

IMO, while TS T4 has benefits (better off boost/transient response), I dont think it as great as the TS T4 supporters try to portray it as. I have yet to see a REAL apples to apples (same car, same setup, comparable weather, just OS T3 to TS T4 changed and nothing else) comparison that shows TS T4 to be so much better for the "normal' EVO owner. I believe its beneficial on a road race build though

 

This was my experience with going from a OS T3 to a TS T4 setup. The T3 OS was done with a RevHard manifold and the T4 TS was done with a short runner manifold I built. No other changes to the setup and the turbo was a GTX3076r.



When I swapped to the GTX3576r on the stock bottom end the turbo would surge in 3rd gear because it spooled so quick. Here is the spool data on the stock bottom end. 31psi @ 4156rpm in 3rd.

 

Actually, that's kinda late, regardless. We're looking at a couple of hundred rpms difference here in a 500whp application in one SNAPSHOT of a gear. As I stated in my last post, I'm looking at this from a transient response perspective in several gears, not just top end, which is what is being shown here... In addition, both are falling off at similar points, which kinda goes against your argument a little.

For a T4 1.06 to make more top end power with reaching maximum boost pressure 200 rpms earlier, That's expected for a T4 divided housing with a larger volute to allow a higher volume of exhaust energy over a T3 .82A/R as a single volute. It doesn't mean that it was a "twin scroll" that made it make better top end power, its the fact that its a T4 that uses a larger volute circumference than a T3 .82. With the larger volute, comes higher exhaust energy volume (notice I didn't say exhaust velocity), if you're using an HTA at over 25psi boost pressure levels. Put the two together on the street for acceleration in its proper gearing, You'll see the T3 .82A/R recover a bit more quickly (because of higher exhaust velocity) over the T4. This is nothing new here.

 

I completely agree with this, which was part of my original point.. But what I think is being MISSED here is the fact that a T4 divided housing is NOT A TWIN SCROLL design, like the factory volutes were designed by MHI, or in newer cases by the EFR.

 

If you could, please show a photo of both your revhard manifold AND the one that you put together yourself. I have a feeling I'll easily be able to explain this as a matter of exhaust backpressure at the manifold...

I'll wait for it..

 

If you could, please show a photo of both your revhard manifold AND the one that you put together yourself. I have a feeling I'll easily be able to explain this as a matter of exhaust backpressure at the manifold...

I'll wait for it..

 

You know... there's this thing called the 'internet' which has multiple search options. I decided to use something called 'Google'.

Found in 73 seconds:
https://www.evolutionm.net/forums/ev...onversion.html

https://www.evolutionm.net/forums/ev...xperiment.html

Having looked at the pictures, the Revhard is basically a log with a nasty sharp 90 degree turn into the turbo which is killing it on the topend. However, it is lower volume than the custom short runner manifold 240Z fabricated which does better manage the flow into the turbine housing. So 240z's manifold does have better flow which makes up the power gain, but the twin-scroll is what's improving the spool-up even with the larger volume turbine housing.

 

Your argument really doesn't make sense. Exhaust energy doesn't come in "volume" and the turbine housing doesn't effect that much anyway, as volume of exhaust is fixed with the engines VE. A divided housing works by separating the kinetic energy of the exhaust pulses so they hit the turbine wheel individually, rather than mixing up in one big collector. This is why they spool faster. Whether they make the same power or not is based on the volute design/volume and whether or not it can support the engine's flow.

Seriously have you ever driven two cars with similar setups where one is single and one "divided"? Transient response and off boost throttle response is better with the divided housing. The powerband is usually bigger if it's sized right as well. While that might not make a difference to a drag racer, anyone who drives their car on the street or road course is much better off with a "divided" housing.

Call it "divided" or "twin-scroll", it does the same thing.

Yup. Pretty much.

 

You TOTALLY MISSED the point

Corrected numbers-



The Shodan claimed that a Divided couldnt or wouldnt make the same results as a true twin scroll. The Garrett white paper says that a twin scroll housing (I looked it up, it doesnt use the word Divided) will spool like its 60% of its size and then make 5/6s the power of its size. It actually says that you need a twin scroll thats 1.2x larger than the single scroll for the power level, but thats inverse math at work, 1/1.2=5/6.

The slight difference in spool is that the 1.06T4 still has larger volute volume so it will still spool like a 0.63 T4 not a 0.63 T3. Since the 0.63 T3 was starting to run into flow issues this is where the 1.06 being able to make the power of a housing 5/6 it size comes in. 1.06 * 0.8333 = 0.88 T4. We get the best of both here, spool similar to a 0.63 and the power potential of a bigger 0.88 single scroll.