How do you think the results would play out had you used a single scroll manifold on the twin scroll turbo instead of the other way around?

 

I do not remember insulting you, but if I did then I apologize. Just confused as to how your listing relationships with various folks within the industry has a bearing on the test or its results.

First and foremost we are all friends and this is just a cordial debate among car enthusiasts. Additionally, I appreciate that you tried to perform a legitimate test comparing twinscroll vs openscroll results.

I understand you want to stand by your test, but the reality is that reasonable doubt regarding the results exists because you chose to run the cast vband twinscroll flange on an open scroll turbine housing. You want someone to prove that your test is flawed, when you cannot prove your test is not flawed.

One last question, if you were building a header for your personal car (assuming 4cyl) for an open scroll turbine housing, which of the merge collectors below would you choose and why?

4-1 vband merge collector


or

Garrett cast vband twinscroll merge collector

 

I would have liked to see a straight-up T4 TS vs T4 SS comparison, I'm not a fan of Garrett's TS v-band setup for the abovementioned reasons.

 

letsgetthisdone:

You statement "The flow of a system (turbo and manifold for example) is only as good as it's most restricted point" isn't correct.

A more accurate principal is tha the PEAK flow of a system (turbo and manifold for example) is LIMITED by the most restricted point during SUBSONIC FLOW . I have to run some number to see the flow velocities through each runner to see the mach numbers. Of course, these are dependent on the power output, pipe diameters, EGTs and a number of other factors. I would doubt that the 20% reduction in area constitutes a choke where the flows reaches a sonic speed.

What is interesting is that supersonic flows are accomplished by using a channel that has a converging then diverging nozzle. So the choke area is important, but it may not be a factor is the power levels (mass flow through the system) are nowhere close to reaching sonic speeds.

If you aren't familiar about Carl de Laval and the 'de Laval' nozzle or converging diverging (CD) nozzles, you should check them out. A bit of 19th century magic.

 

RS200

Sounds like a cool plan. Perhaps we'll do that test in the future. Maybe we'll try it on an inline six cylinder to see if the results are more or less dramatic. Maybe we will also try with both types of manifolds too to prove that it has little effect.

 

As a note, DSPORT made 630whp on a dynojet with what looks like a GTX3576R. I generally consider that a 600whp turbo (leaving a little margin so as not to overspin the turbo) on a dynojet typically with a T3 flanged 0.82 (the 1.06 usually does not make more peak power, just spools slower). So 630whp is really at the limit of the turbo.

Mike Essa used the this same Garrett v-band divided turbine housing and adapter configuration with a GTX3582R in his BMW E46 M3 and won the Formula Drift championship that year. His car was reported to have made 750hp. I haven't seen a dyno sheet, do not know if that's crank or wheel, but I typically consider the limit of the GTX3582R on a dynoject at 725whp pushed to the limit.

So it seems the Garrett v-band divided turbine housing is not a major choke point in making power based on the limited data openly easy to find.

 

240Z Turbo

You said," You want someone to prove that your test is flawed, when you cannot prove your test is not flawed."

There are potential sources of error in any test conducted. We did everything possible to minimize these. In one sense, you could argue that every test is flawed.

Of course, you didn't mean "flawed" it in that way. You called it flawed because you believed one element of the test should have been done differently. We disagree on that point. In our opinion, we reduced the possible sources of error conducting the test with the procedure that we used.

Now if you want to call the test flawed, you should be able to prove that doing it with your procedure would accomplish two things:
1) Show a different end result
2) Show a result that you would be able to prove is more accurate

The conclusion of the results is a twin-scroll will outperform a single-scroll in terms of boost response and power output with all things being equal. There is no dispute that this test shows the exact differences if a twin-scroll manifold is used on a single- or twin-scroll turbo.

 

Your conclusion is not accurate because all things were not equal and the results are specific to your test that used a vband twinscroll flange on a vband open scroll housing. Let me write it for accuracy.

The conclusion of the results is a twin-scroll will outperform a single-scroll in terms of boost response and power output when using a header fitted with a vband twinscroll flange.

You chocked cross sectional area entering the open scroll housing by 20+% so it is up to you to prove that doing so does not affect the performance of the open scroll housing. If you had made a proper merge collector for the openscroll turbine housing I suspect the twinscroll would out spool the openscroll, but the openscroll could potentially make more power.

Curious as to why you chose not to answer my simple question.

 

240Z TwinTurbo

What simple question didn't I answer for you?

 

my opinion of TS is it is indeed awesome when you want to maximize the power band on a street car. but pretty much useless on a drag car. should be the end of the discussion as that is it in a nut shell.

there is no such thing as a good back to back test that proves anything. the dsport test is flawed because they used a twin scroll manifold on a open scroll turbo so power on the open scroll was less than it should have been. there is also no such thing as saying turbine housings of the same size when talking TS and open scroll. you just cant do a good back to back. too many variables will change.

FP just started selling 2yrs ago bolt on turbos for evo 8/9 with their own design open scroll housing instead of the factory TS housing. there is enough dyno data on the hta green,red and black open scroll and TS to conclude quiet easily the TS powerband starts about 500-600rpm sooner for those size turbos. and the new open scroll makes 5- 10% more high rpm peak power for the same turbo size.

I think this test is ruffling everyones feathers because MOST know a open scroll makes more power than a properly designed twin scroll. and this test doesnt show that. I personally think that garret cast inlet should be trashed and redesigned. the inlets should be oval and it needs to be made longer with a gradual transition. better yet trash the TS 3in vband inlet housings and redesign for 4-4.5 vband. the inlets are just too small for the turbo sizes used.

 

What data do you have that says the inlets used on the Garrett v-band divided are two small? DSPORT just showed 630whp out of a GTX3576R on it and that's really all the power you should expect out of that turbo. Pick any car, engine, exhaust manifold and turbine housing match you can find anywhere on the internet, and if you see one making significantly power on a dynojet, post it up. Quite frankly, 620-630whp dynojet are more typical GTX3582R numbers with peak numbers out of the GTX3582R being 700-725whp at the limit.

Why do the inlets need to be oval? Should every turbo manufacturer change the inlets on the T3 and T4 divided flanges from rectangle to oval?

Say you do change the flange size to 4-4.5" diameter and made it longer for a more gradual transition. That is going to make packaging on a lot of cars very difficult.

While I do not disagree a larger flange with longer transition would be more ideal, the gains in efficiency on a converging nozzle by having the longer transition are not nearly as big as on a diverging nozzle such as on the turbine side.

As for the area of the ports, while they may appear small to you, they are still larger than the Area leading into the volute of the turbine housing on the largest 1.01 A/R housing (smaller options being 0.83 and 0.61). So you still have a constantly reducing area into the volute of the turbine housing.

 

if you measure the area of TS inlets on T3, T4 and vband garrett you would have an idea why I think they are too small. and the transition is flat too sudden. I dont have any testing to prove anything its just my opinion and I stated it that way.

630whp is not that great for a Billet 58mm compressor 35R. many evos have made 750whp on the 3582R by FP. same 35r turbine, 59mm compressor. nothing you cantered with changed a thing of my opinion of what I see here. that collector is a 50-100hp restriction when maxing out a turbo size. these housings have been out for 3 years now. this 630whp is the biggest I have seen.

 

The inlet area is of less concern than what is often referred to as the T-T section which is the area highlighted in the graphic; as long as that inlet area is not smaller than the area at T-T or you would end up with an expansion before reaching T-T and flow loss. The T-T is where the volute begins. And again, ideal world, yes, the transition would be longer and the area larger, but real world concerns dictate the packaging compromise as the impact on performance is relatively low.

Again, you can google Mike Essa and his E46 drift car where it is stated the car made 750hp on this v-band divided turbine housing on a GTX3582R. And as I said before, I do not know if that's wheel or estimated crank, but that's in the range expected for a GTX3582R. 750whp on a HTA3582? I have yet to see that! On a HTA3586, yes, but not the 82. Why do you think FP had to come out with the HTA3586? It's because the GTX3582R out-flowed it.

And again, google any GTX3576R dyno you want. STIs, Evo Xs, S2000s, and you'll be extremely hard pressed to find one over 630whp. Vast majority are in the mid-500whp range as most people, and rightfully so, do not push them to the limit.

 

The compressor wheel for that turbo is rated at 64lbs/min so it is hard to understand how someone can make 630hp@wheels regardless of boost. Using the GTX3576r and the 1.01a/r vband twinscroll turbine housing I was able to make 572hp@wheels max effort 37psi tapered to 33psi(wouldn't hold boost due to flow). Because I log turbo shaft speed I could see that it was riding the choke line. In the GTR community a member recently pushed a pair of GTX3576r turbos and the car made 1177hp@wheels, which is 588hp@wheels per turbo.

Regarding the cast vband twinscroll flange, its transition is fairly abrupt. It literally transitions from cs area of ~3.14sq-in to an cs area of ~1.48sq-in in a length of 1.5". From an equivalent tubing, it goes from a 2.000" ID to 1.375" ID in a length of 1.5".

 

Agreed. Its flawed because of this. The true power potential of the open scroll is limited because of this. This is obvious. It may not spool as well, but the OS would have made more top end power if the correct header and collector were used... That is if the compressor had the potential to actually move the air.