What the EFR turbos should have been

[LuckyX2]

The GT2871 is poorly balanced but arguably the most popular turbo for an SR20DET.

I'm just curious to see what would happen...

Originally Posted by PoorMans180SX

The thing is, if you have big differences in turbine and compressor size, you lose turbine effeciency because the tip speeds are different, and power and transient response are sacrificed. Garrett has finally caught on with their new GTX3576. It's a GTX3076 compressor wheel with a 3582 turbine wheel. Spools exactly the same as a X3076 and makes somewhere around 50whp more.

Check out this article:

Turbo Tech: Compressor and Turbine Map Details

And this:

http://forums.evolutionm.net/evo-dyn...-gtx3576r.html

Thanks for the links!

[turtl631]

I'm still waiting to see what pans out when these are actually readily available with many dynos on different engines, not just a few early testers, etc. The GTX turbos are quite expensive considering they don't include a turbine housing, so there is certainly cost parity between Garrett and BW. I think the most potential for us cheap 240SX owners is the internally gated T4 twinscroll options, as you need only a simple divided manifold and v-band downpipe. Less connections, save the expense of a turbine elbow and the hassle of recirculating a WG dump. Of course the only option is the 0.92 A/R, so better hope it works well, and if you go to another snail you're limited to the T4 flange which is uncommon in our world. Garret now has several divided T4 turbine housings for the GT30 turbine however.

[LuckyX2]

BorgWarner EFR 7163 Turbo - Full-Race.com

These are finally available to the public! New turbine housings too, such as .80 T4 twin scroll.

[Motary]

Shame it is still sand cast as it looks quite rough, even compared to other manufacturers who use the same technique. Is anyone here on the list for one? Most likely to be a very popular turbo for the 500 whp crowd

[PoorMans180SX]

I know of a few Evo guys that are.

[LuckyX2]

Does the roughness in the casting really make that much of a difference? If anything, don't you want the compressor side a bit rough to reduce laminar flow?

And no S/Z chassis guys I know of with this yet but this R35 already broke a record with twin 7163s:



I'm gonna update the first post with some new info by the way.

[Def]

These look to be roughly $2100-2300 turbos. Yes they add some bits over the equivalent Garrett that's around $1300-1400 sans turbine housing - but that still seems quite the premium for very very similar performance.

It seems it all comes down to the typical "power capability vs. spool" argument, and while there are small differences across manufacturers, it seems between the good ones that there is not much between them if you pick turbos of the same power capability.


I do wish Garrett would update their GT28/GT30 turbine wheels. I know it takes a while, but they're really under no pressure due to BW *STILL* having supply issues with their EFRs. Haven't they been "out" nearly half a decade now, and now have a crappy looking turbine housing to boost production numbers?



And no on the rough turbine housing not making a difference - you would ideally want it to be a mirror finish. It is amusing in that they really touted the turbine housing finish when they were trying to investment cast them, and kept pointing out the crappy finish on Garrett housings... then they come out with some stuff that looks worse than the average Garrett housing a little while later. hah

[LuckyX2]

I'm not seeing the roughness you guys are talking about on the turbine side. Assuming the inlet is as smooth as the inside of the outlet in the picture, it looks pretty smooth to me. The compressor side is rough, yes, but as I was saying before, a rough compressor outlet would be good to promote turbulent flow, wouldn't it? I was agreeing before that the turbine side should be smooth but the compressor side is a different story.

Also, I don't think the numbers are as far off as you're claiming, Def, it's $1850 for a V band EWG 7163 and $2100 for a T4 IWG 7163 at Full-Race. Comparing that to the Garrett without a turbine housing isn't really fair either. Depending on the housing, that's going to up the price $100-250. So it's more like $1850-2100 vs $1400-1650. And that's without counting the money saved by the inclusion of a boost control solenoid, BOV and IWG that's good enough to not require an EWG. Not needing to buy wastegate(s) is a big cost savings and BOVs aren't cheap either.

I'm also going to have to disagree that they're very, very similar in performance.



Look at all that extra torque down low vs a 3076 while still matching it up top. This was on a prototype too. They've since tweaked the turbine geometry further and upped the power estimates by 50hp.

I'd love to see a head to head comparison with a GTX2867 or GTX3071/76 on the same turbine housing. Measure time from full vac to full boost starting around 3-4k (to test response coming out of a corner) on all of them and I bet the 7163 would get there much quicker than both while making more power than the 2867 and matching the 3071/76 in power. To quote the Madzdaspeed 6 tester again, "The spool is unreal. We went from vacuum to 22 psi in 400-500 RPM ANY time we hit the gas above 2400 RPM." I don't think a GTX could do that. You just can't ignore the inertial savings of the Ti wheel, and now BW has a geometry advantage too. I don't see how they could be very close with a distinct physical material advantage for BW. But currently there's very little info, so it's a lot of educated guessing and extrapolating the little data we have.

[PoorMans180SX]

Originally Posted by Def

These look to be roughly $2100-2300 turbos. Yes they add some bits over the equivalent Garrett that's around $1300-1400 sans turbine housing - but that still seems quite the premium for very very similar performance.

I'm not really going to argue here, both turbos cost a lot of money, and I think there is a correct application for them all. But I've never seen a single-scroll turbo on a stock STI spool dang near like the stock turbo and be capable of 500+whp. I know I know dyno's don't show transient response, but I'm pretty sure the power curves don't lie.

Originally Posted by LuckyX2

I'm not seeing the roughness you guys are talking about on the turbine side. Assuming the inlet is as smooth as the inside of the outlet in the picture, it looks pretty smooth to me.

I do believe the external gate housings are still investment cast across the EFR line. They only had problems investment casting the more complex internal gate housings.

[LuckyX2]

Ah, that's a shame about the castings on the IWG housings. Would be nice to see a back to back between the internal and external V band housings. Probably not a huge difference but I'd expect a difference none the less.

And speaking of stock STI turbo, the Perrin test shows that well too:


Pretty much identical below 3k but blows it away above that. Too bad this Perrin test is really our only professional source of info on these. And even so, it was with an older prototype. Only a little bit of info out there besides this test and it's all just random forum people.

Another random note. Indy is switching to this turbo for 2014. In 2013 they used the 6758 so, keep an eye on Indy to see a good indicator on how this compares to the first gen stuff.

[PoorMans180SX]

Let's not forget the aluminum bearing housing, that is pretty cool too.

Well, the Evo guys are super big on dyno'ing their cars and whatnot, I'll make sure and relay any info I get back here. There is at least one T4 twin-scroll version going on a 2.0 liter Evo.

[LuckyX2]

Along with the aluminum bearing housing (which is also much longer) they have ceramic bearings vs Garrett's steel, metal retaining clips vs Garrett's plastic and dual, redundant seals on compressor and turbine side vs Garrett's single seals. Assuming they took care of the shattering turbine issue from a producer not following spec on their alloy, these are looking to be extremely reliable.

Looking forward to those Evo results too.

[McRussellPants]

lol at the GTR breaking records and being more than half a second off of English Racing and Boost Logic.

:TLDR

[Kingtal0n]

one day when photobucket is healed


I just love looking at compressor maps, but I feel like not everybody can get the info they want out of them. So Instead of studying like I should be I took a minute to add what I feel every compressor map should come with...

And I am not sure you can call the 500 horsepower sr20 community a "crowd". Maybe more like a "pond" or "swamp"

[LuckyX2]

Originally Posted by McRussellPants

lol at the GTR breaking records and being more than half a second off of English Racing and Boost Logic.

:TLDR

I'm not sure which record it was they broke but they claim to have some kind of record and that video was just posted two weeks ago. I know there are faster GTRs but perhaps that is the fastest with twin 71mm turbos?

Thanks for the points on the map, Kingtal0n. I put a link and screenshot of my matchbot map results at 25psi and 7500rpm max on the first post but your map shows some points mine does not.

[PoorMans180SX]

Originally Posted by LuckyX2

I know there are faster GTRs but perhaps that is the fastest with twin 57mm turbos?

Fixed that for you. People talk about compressor inducer size when they talk about turbos in millimeters .

[Def]

Originally Posted by LuckyX2

I'm not seeing the roughness you guys are talking about on the turbine side. Assuming the inlet is as smooth as the inside of the outlet in the picture, it looks pretty smooth to me. The compressor side is rough, yes, but as I was saying before, a rough compressor outlet would be good to promote turbulent flow, wouldn't it? I was agreeing before that the turbine side should be smooth but the compressor side is a different story.

Also, I don't think the numbers are as far off as you're claiming, Def, it's $1850 for a V band EWG 7163 and $2100 for a T4 IWG 7163 at Full-Race. Comparing that to the Garrett without a turbine housing isn't really fair either. Depending on the housing, that's going to up the price $100-250. So it's more like $1850-2100 vs $1400-1650. And that's without counting the money saved by the inclusion of a boost control solenoid, BOV and IWG that's good enough to not require an EWG. Not needing to buy wastegate(s) is a big cost savings and BOVs aren't cheap either.

I'm also going to have to disagree that they're very, very similar in performance.



Look at all that extra torque down low vs a 3076 while still matching it up top. This was on a prototype too. They've since tweaked the turbine geometry further and upped the power estimates by 50hp.

I'd love to see a head to head comparison with a GTX2867 or GTX3071/76 on the same turbine housing. Measure time from full vac to full boost starting around 3-4k (to test response coming out of a corner) on all of them and I bet the 7163 would get there much quicker than both while making more power than the 2867 and matching the 3071/76 in power. To quote the Madzdaspeed 6 tester again, "The spool is unreal. We went from vacuum to 22 psi in 400-500 RPM ANY time we hit the gas above 2400 RPM." I don't think a GTX could do that. You just can't ignore the inertial savings of the Ti wheel, and now BW has a geometry advantage too. I don't see how they could be very close with a distinct physical material advantage for BW. But currently there's very little info, so it's a lot of educated guessing and extrapolating the little data we have.

I'd sure expect a whiz-bang EFR to outspool a decade+ old GT3076R that has a pretty bad wheel diameter mismatch from the outset.

The EFRs aren't junk, but I've seen a ton of tests pitting them against old old Garrett stuff. The few comparisons I've seen between modern turbos basically shows very very little difference between them. As you get billet compressor wheels on both which up the flow for the inertia and wheel diameter, which then helps out with the spool, lower backpressure etc.


Don't forget that inertia is proportional to radius^2. So while BW might be using slightly bigger turbine wheels to help generate torque and help spool, it also puts the inertia much closer to the more compact inconel wheels that Garrett uses. The new turbine material does have advantages, but it's very very small increases - nothing drastic like it's marketed to be.

[LuckyX2]

Yes, inertia is linear with mass and quadratic with radius but even so that still means BW's claimed halving of mass allows them to make the turbines 1.4x larger and have the same inertia. A 40% increase in turbine size at no inertial cost is pretty dramatic to me. But this is also assuming the material is distributed the same in the turbine design. If BW's wheels have more material at the edges than Garrett's then obviously some of that advantage is lost. Even so, a 10-20% increase in turbine size would still be a formidable benefit. Combine that with the new geometry and twin scroll housing and I don't think anyone has anything to match the 7163.

Where are these comparisons you've seen? I've only seen a handful of comparisons with the first gen stuff and next to nothing on the 7163.

Edit: Found a GTX vs GT comparison


Looks like the GTX isn't much better than the GT. The 7163 that Perrin tested would still beat it pretty handily below 3500 and be about equal above there.
And if anything, the GTX spools slightly later. That makes sense considering they have a more aggressive compressor but the same turbine powering it from the GT.

[Kingtal0n]

But my physics book says that a cockroach on a spinning disc that suddenly stops to eat a bread crumb... ah forget it

[Def]

Originally Posted by LuckyX2

Yes, inertia is linear with mass and quadratic with radius but even so that still means BW's claimed halving of mass allows them to make the turbines 1.4x larger and have the same inertia. A 40% increase in turbine size at no inertial cost is pretty dramatic to me. But this is also assuming the material is distributed the same in the turbine design. If BW's wheels have more material at the edges than Garrett's then obviously some of that advantage is lost. Even so, a 10-20% increase in turbine size would still be a formidable benefit. Combine that with the new geometry and twin scroll housing and I don't think anyone has anything to match the 7163.

Where are these comparisons you've seen? I've only seen a handful of comparisons with the first gen stuff and next to nothing on the 7163.

Edit: Found a GTX vs GT comparison


Looks like the GTX isn't much better than the GT. The 7163 that Perrin tested would still beat it pretty handily below 3500 and be about equal above there.
And if anything, the GTX spools slightly later. That makes sense considering they have a more aggressive compressor but the same turbine powering it from the GT.

The GTX3076R is a crap turbo, as the GT3076R was already a compressor/turbine mismatch, and putting even MORE compressor on the small 60 mm turbine is just not going to go well. Of course it spools up slowly and barely outpowers it. Get to a more reasonable compressor/turbine wheel speed ratio like the GT3071R to GTX3071R and a better story emerges.


As for TiAl wheels - it's not just all a huge party. The turbine wheels need a larger cross sectional area on the blades to handle the stress since it is a weaker material at high temps than the very nice 718 Inconel of Garrett wheels. Ideally you'd have an infinitely thin turbine blade, so having a thicker blade cuts into your turbine efficiency. To combat this BW went with a full back plate on the turbine wheel like you see on compressor wheels. This enhances efficiency at the expense of a lot of inertia, and also helps give structural stability to the blades by supporting their back end. This all sums up to way more inertia than simple weight measurements would indicate, and a slight hit to turbine efficiency. Which probably explains why most EFR designs are going with higher turbine wheel major and minor diameters for the same "power" compressor wheels vs. the GTX lineup.

This is some good info on TiAl wheels (and other good info on speed ratios, and steady vs. unsteady behavior):

On Mixed Flow Turbines for Automotive Turbocharger Applications

[PoorMans180SX]

Yeah, the 3071 is a decent match, pretty cool that they came out with the 3576, which is a really nice match. Seems like the 3582 needs a bigger turbine, I helped with a GTX3582 built Evo that really didn't like anything past 30psi.

Some of the EFR's seem really great and others seem average. The 6258, 7163, and the 8374 seem to have really great spool and performance for their size. The others seem more "normal" I guess. Almost everyone that runs one, regardless of what turbo they switched from, raves about the transient response though.

[McRussellPants]

There is no magic.

The only reason EFR looks better or worse is its a half size smaller or bigger than the Garrett you're comparing it to.

EFRs always seem to have exhaust wheels 2-5mm bigger than a garrett with the same compressor inducer. So every review of a EFR ever: "it only spools 100-200rpm later and makes a bit more power and it doesn't hit transitioning into boost."

Shocking.

[PoorMans180SX]

Ummm none of these tests are pushing the Borgs hard at all... We regularly run Borg and Holset turbos to 40+ psi on E85, and they keep making power. Garrett's, not so much.

I don't see why people are so pro-Garrett or pro-Borg. They both make great products, and they all have different characteristics.

Let's not get into oversimplifying for the sake of argument.

[LuckyX2]

Thanks for the research link, Def, it was a good read. And sorry for all the 3076 comparisons. It just seems like I keep finding comparisons done against that; I'd love to see a 7163 against a GTX3071.

I'm getting a vibe that you're thinking mostly of the first gen EFRs though. I realize that the first gen models weren't all they were hyped up to be; the title of the thread even implies they were a letdown. The first gens delivered in terms of transient response but weren't anything spectacular in terms of power, flow or efficiency for the reasons you just stated. But I think the 7163 is going a long way to fix that and the link you posted seems to support that.

The negative you pointed out about the TiAl wheels needing more backplate material is one thing countered by a mixed flow design. To quote the paper: "In addition to this, mixed flow turbine wheels offer the chance to design turbine wheels with reduced inertia. One key benefit is that the back disk is clearly reduced in diameter." So there's some mass at a large radius removed but even before that mass was removed on an MFT, the paper still claims a 46% improvement in inertia for the turbine alone and 30% improvement for the entire rotor assembly with the regular radial turbine.

Another disadvantage is countered by mixed flow: "To achieve the same flow capacity, the blade angle distribution needs to be modified for readjustment of the throat area of the wheel ... because of the aforementioned γ-TiAl design constraints a turbine efficiency penalty of 5% is assumed." With a mixed flow design and it's greater swallowing capabilities, this compromise doesn't have to be made and the efficiency penalty that you mentioned is present on the first gen EFRs isn't incurred with the mixed flow design of the 7163.

Seems like TiAl and MFT tech goes hand in hand. The combination of the two technologies enables what seems like a no compromise turbo. Lower inertia, greater efficiency especially at lower velocity ratio and greater flow. The inertial savings also gives BW a lot of room to play with sizing and the greater air scavenging ability of the MFT means greater torque production which lets them go with a more aggressive compressor design too.

Just look at the 7163 vs the 7064. Nearly identical sizes but the 7163 flows 4lbs/min more. The MFT clearly improved things on the overall flow/power front and data would suggest transients will be improved too from the further reduction in inertia along with low down power from the increase in efficiency at low velocity ratios.

Originally Posted by McRussellPants

There is no magic.

You're right, it's not magic, it's technological advancements. Just in the same way carbon fiber enables better parts, TiAl and MFT enables a better turbine. It's about time we stop using a turbine design from the 70's don't you think? Also, you're still comparing to the first gens which were admittedly underwhelming. Yes, they sized up the turbines slightly but that's because they could with the weight savings from TiAl. Have you ever been in a car with an EFR? In overall power it's going to be similar as you said but the transients of the EFR are unmatched. BW made a design compromise on the first gen turbos. Depending on their wheel sizing, they could have made either a quicker or more powerful turbo but went for balance while improving on transient response instead. They didn't have to make that compromise with the 7163 though thanks to the MFT. It responds like a 6758 but yet it flows more than a 7064. I find that to be a great improvement and am surprised to find little agreement with that in this thread.

Originally Posted by PoorMans180SX

Almost everyone that runs one, regardless of what turbo they switched from, raves about the transient response though.

Check this link out:
http://blog.perrinperformance.com/bo...-tuned-part-2/
The 7670 isn't one of BW's best EFRs but I think it still showcases your point nicely. On the dyno it's looking like the 7670 is trading blows with the 3076 and spooling slower but when he logged it by time instead of rpm, the 7670 hit full boost a whole half second faster than the 3076. Add up those half seconds coming out of every turn on your favorite race track and you have a significant lap time improvement. Also, even though the Garrett arguably dynoed better, he still preferred the BW for the "feel". Driving just felt snappier and I think that's the whole point of the BW's that people missed on the first gen. To quote a friend who drove a 1.8 vvt Miata with a 6758, "I can't imagine a turbo spooling faster, it was already transparent. It felt pretty much like a v8." He owns an LS1 swapped FC RX-7, I imagine that's what he was drawing a comparison to.

[ultimateirving]

Yea based on all these posts the turbo sounds like it could be awesome. But until I drive a car with one no one can really comment on its feel. Also with the current price point I don't think I'll be trying to afford one anytime soon. And for those whom are going with full builds might shell out the extra cash for it, but the cost of the turbo is a major factor for me.

[Mannykiller]

not much tech to add here but I've been building my drift car for about 3 years now. Had a series of engine failures due to ignorance. But I came from a Gtx3582R and will now be running an EFR8374 on my 2 rotor. Hoping to have some comparison dyno charts etc here in the next couple months.
GTX3582R set up w/dual synapse gates

So far only started on my manifold but I'll be running dual Turbo Smart pro comp 40's on an updated design of the long runner above.

[Kingtal0n]

Originally Posted by Mannykiller

Had a series of engine failures due to ignorance.

Its not your fault dude, thats what the rotary engine does. I've been trying to get them to add the definition of "hand grenade" to rotary in the dictionary.

[LuckyX2]

On a somewhat related note, has anyone seen Honeywell's dualboost turbos with axial turbines and dual compressors? They announced it a while back but just read about it today. Pretty cool tech.

Gasoline DualBoost Turbochargers | Light Vehicle Gasoline Engines | Our Technologies | Honeywell Turbo Technologies
http://turbo.honeywell.com/assets/pd...esentation.pdf

Apparently they're looking into incorporating VGT and TiAl too. But even in its "basic" form they're seeing a 50% inertia reduction and a significant movement of the peak efficiency to a lower blade speed ratio (~0.6 vs ~0.8) while keeping peak efficiency the same. Steady state performance is the same but they're seeing 50 more NM of torque 0.6s earlier on a 1.6L engine. The mixed flow turbine in the 7163 is a cross between this and the radial turbines we're used to. It would be reasonable to expect some of the same benefits Honeywell had with their axial turbine.

Awesome time for turbo tech right now

EDIT: Added one more compressor map overlay with a GTX3071 to the first post.

[Def]

Neat technology... probably won't see the light of day for a decade+. Unfortunately BW dropped the ball to really push Garrett in the marketplace, so Garrett is keeping investment pretty low on R&D and actual product roll out.

Before you guys jump all over me, BW has made some neat turbos, but actually GETTING them in your hands many years after their launch is still a joke. They're also priced on the very "premium" side of things regardless of the arguments of how a $4 plastic BOV cap or cheap boost solenoid with a IWG change things up.

[LuckyX2]

I completely agree with you there. This thread is proof of that; I started this in 2012 and we still don't have a steady supply. I think I read that the waitlist for the 7163 extends into March... Apparently in the time since, extra R&D has squeezed another 50hp out but come on already, some of us just wanna buy the thing. And yea, probably a while before the dualboost starts showing up. It's an OEM product so I'm sure it will have to go through years of bullshit before it starts to trickle into cars anyone can buy.

Also, Honeywell claims their dual compressor design doesn't make any thrust along the axis of rotation and market that as a feature. Makes sense due to the symmetry cancelling the two thrusts out but wouldn't the new axial turbine create thrust? I imagine having the gas flow across the turbine axially would put a thrust on the axle towards the turbine housing outlet, wouldn't it?