Full-Race Geoff

this is not an urgent area of attention for us, but if there is enough demand we could consider it. We are working on the 996 and 997 EFR turbo kits right now

thanks for sharing those photos, that is exactly what andrew darud would want to build. i agree the twinscroll garrett turbine housings leave a lot to be desired, one major advantage of the borgwarner turbos is that their turbine sides are the best-in-the-biz and at less than half the price

correct - T4 is the way to go if doing a twinscroll. the T3 makes less sense in my experience, since you are stuck on limited turbos to choose from with inferior quality turbine hsg castings, and less room for a wrench to access the bolts

i would recommend no larger than 1.625" ID for this crossover

1bar at 3000rpm on a hta30R is very good, thanks for sharing

their statement is that availability will be greatly improved by the end of Q2 2012... not far away. let me know via email what you are looking for and im happy to keep you updated once we have a better handle on timelines

twinscroll and VGT are mutually exclusive.. you can't have both.


Im not sure what vitesse is doing/using, but to answer your question, 1 bar at 3000rpm on a 60lb/min hta30 compressor is very good. VGT aka VTG aka VNT is not suitable for gasoline use - and with the extra energy in a petrol engine's exhaust stream it is not necessary nor is it going to become "mainstream" for the aftermarket tuning communities due to many reasons which i dont have time to explain (certainly not in this thread). I have pushed on this ad nauseum with the BW team as I really thought variable vane was the ticket, but at the end of it all the twinscroll housings offers benefits that the singlescroll configuration does not. additionally the cost of a variable vane petrol turbo is likely to be in the neighborhood of 3000-4000$ usd and that is before the control system is implemented (vacuum actuated canister does not count)

AL951

If you think about it, a a/r of 1.06 is petty big for our engines, so 3K is not bad at all. I know a K26\6 is A LOT smaller the that for sure.

The thing would be to find a housing with a smaller a/r to bring there spool up down. Or install a valve to in one of the turbine in lets to increase gas velocity at low RPM.

Regards,
AL

Thom

Anyone on here has stock headers lying around to measure their inner diameter?

Darud35

Measured my stock headers last night.. If I recall correctly the ID was ~48.4mm I can confirm that later.

The downpipe off the turbo was ~60mm

Geoff, would it make sense to match the ID of the headers for the crossover?

Also, what route are you guys using for headers? My stocks are in rough shape, but I don't know about the $900 price I've seen for some kits.. If I'm making a crossover maybe I should do headers too? Obviously I can't put the R&D Porsche did into flow efficiency/matching engine flow rates.

Darud35

I was confused on the difference between VGT and twin scroll. I thought they were the same.

My understanding:
-VGT is variable geometry of the turbine blade. It adapts to different inlet flow rates to optimize spooling. It also doesn't need a wastegate.
-Twin-scroll needs two inlets as it points the two inlets at fixed angles relative to the turbine wheel. Optimizing the spool up over a broader RPM range

Can someone confirm or deny?

thingo

Good call on the pipe size Geoff, we did try different sizes, however the outlet of the headers is 48mm and we didn't want to go smaller than that, but larger went backwards in terms of response(no gains).
Yes we have melted a few T3 housings, next we are trying nos garrett housing.

333pg333

Rod on a separate note, you should do a bit of a build thread on the white car. It's got plenty of interesting stuff on it and a nice little something waiting to go in.

Thom

On a T3 Divided turbine housing, each inlet has an area of 1.824"², which translates into a virtual section with an ID of 1.524".
I assume the value of 1.625" mentioned earlier comes for a similar calculation?

Each outlet on stock headers has an ID of 48mm = 2.805" (thanks guys for having measured them).

In any case, it seems necessary to compensate the decreasing temperature of exhaust gasses, as they flow out of headers to the turbine, by increasing their velocity, which is achieved by reducing the global piping section between the headers and the turbo.

Considering the stock set up, the usual KKK turbine inlet has an ID of 1.88", So basically the piping cross section reduces from 2x2.805"=5.610" to 1.88" - that is a reduction of ~66%.

Now, to get even "pulses" through each scroll I assume that the volume of the piping for each scroll should be the same, and the reduction of piping from 2.805" (each headers outlet) to 1.524" (each T3 turbine inlet) should be "evenly" located on each scroll, to make sure that the same quantity of exhaust gas enter, and at the same speed, each turbine inlet.

For people who have stock headers lying around, could you fill them up with water in order to measure the volume of each pair (1/4 & 2/3) ?

Interesting to hear that a T4 Divided turbine housing should be a better choice than the T3.
On a T4 Divided turbine housing, each inlet is 2.482"², the corresponding cross section would be 1.778", which is obviously slightly larger than the T3's 1.524", but since the turbine is, on our cars, that far from headers, can we really conclude that bigger turbine inlets is necessarily better?

This all sounds like fun

Full-Race Geoff

hey i wont be around for the next 2 days, but i wanted to mention that you should not focus on the 2D area of a turbine housing inlet, you really would need to cut the housings in half and study the "nozzle shape" and area at the end of the tongue (A) as well as radius from tongue to centerline of the turbo (R).

The t4 twinscroll is definitely the way to go, even on smaller turbos IMHO. ill get into more detail when i have some free time, got to finish a massive build here in the shop. my buddy wrote this article the other day, you might find it helpful in understanding turbines a little more: http://www.motoiq.com/magazine_artic...p-details.aspx

Thom

Forgot to mention in my reflexion above that I am considering the same A/R ratio (nevermind which of the ones available) between the T3 and the T4 turbine housing, for comparison's sake.
Since the T4 Divided is apparently only available with the A/R ratio of 1.06 anyway, let's say we are talking about an A/R ratio of 1.06 for both turbine housings.

blown 944

An AR of 106 on a t4 would well beyond the capabilities of this engine. I set mine up with a .48 t4 and it was very nice up to a point where the pipe size allowed for too much expansion (theory).

I havnt looked to far into what sizes are available but I don't think I'd go over a .58 ar t4 for anything ran on the street.

Iirc my .48 t4 was very close in size to a .82 t3 in the nozzle area.

I do think there are some benefits to twin scroll, but having built a couple turbos that just flat out spool fast I'm not sure it's worth the work.

Thom

My bad, I was only looking at turbine housing sizes for the GT35, T4 Divided is indeed available in smaller sizes for other Garret turbos.

Darud35

I'm thinking about ordering my turbo in the next day or two. I'd like to get to about 350hp with minimal lag in boost.

Seems like this number is pretty attainable with the Garrett hybrid options. Many are having success with a ~50trim compressor wheel and .63 AR turbine.

My question is how will the BW being discussed with an AR of 1.06 compare to the Garrett? Seems like that's quite a jump in AR. Does the twin scroll reduce spool lag so drastically that I can up the AR so much? I haven't seen a compressor map for the BW on our 2.5L.

Full-Race Geoff

It seems you do not realize that twinscroll is literally (2) seperate scrolls. Try and think of each scroll like a nozzle on a water hose, spinning a pinwheel. In the case of singlescroll you have 1 hose spinning the pinwheel, in the case of the twinscroll you have 2 hoses (nozzles) spinning it. The A/R must not be thought of the same as (1) scroll a/r sizing. IE: a twinscroll 1.0 a/r is actually (2) x 0.5 A/R nozzles, spraying exhaust gas at the turbine wheel

again - this would be common knowledge for those who are familiar with twinscroll, but your particular platform has been dominated by singlescroll configs for a long time. Nothing wrong at all with a proper singlescroll setup, but the torque/midrange/powerband can not hold a candle to a properly constructed twinscroll performance

you can not have the same A/R for an accurate singlescroll vs twinscroll test, the singlescroll would be wayyyy too lazy and laggy at 1.06 a/r OR conversely the twinscroll would be way too restrictive at .63 a/r

the question you need to decide is whether you want something bolt-on: T3/T04E

OR

a fabrication intensive project with greater performance potential: twinscroll

Darud35

Geoff, when you say greater performance potential can you give an idea of what to expect vs the Garrett single scroll option?

Will I have a significant increase in peak power, where peak boost will be achieved, transient response, etc? If you can give some expectations vs the known dbb T3/T04E results: http://refresh951.com/Dyno.htm