GTX2863R vs GTX2867R vs GTX3071R
 

Heres how I chose my compressor.

I don't agree with the majority about choosing a turbo that is at its maximum to acheive the power you want. Sure you end up with more "area" on a dyno plot, but running a turbo well out of its efficiency creates heat. Heat is bad mmmkay.
I made this mistake on my ae86, running a tomei m8270 (which is comparable to a gt2876r, sort of) and sure it makes 430hp on pump and looks awesome ( and will peel 4th gear as the turbo spools on the street ) but the IAT's and the car in general runs way too hot. On the street its fine. On the track it only takes a lap for it to get too hot. I'm looking at upgrading to a gt30 frame turbo.

Don't select a turbo from someone elses dyno!!

This point gets lost on forums. Also, don't just read a review that says "OMG I have this turbo and its da bomb!" Have they tried another turbo either side? Probably not. They're probably just amazed at how much better it is than their previous turbo, and ignore the fact (and will probably get annoyed when you ask) if they considered a different turbo either side. You can't trust or read into someones review that has not reviewed/experienced anything comparable. /end rant.

AE86 dyno for reference. Great for street. Great for forum steez. Not efficient. Not good for track. Two lines are 0.9 bar, 1.1 bar.
https://fbcdn-sphotos-f-a.akamaihd.n...80231371_o.jpg

Anyway, heres the 3 compressors I shortlisted.

http://i13.photobucket.com/albums/a2...ps6e095e9b.jpg

http://i13.photobucket.com/albums/a2...ps70d92a52.jpg

http://i13.photobucket.com/albums/a2...ps5d5d074f.jpg

If you think my boost levels are too high, remember this is at the turbo, not the motor. I have assumed a system with 2psi pressure loss.

The gtx2863r I think it well suited to either a stock fuel system (260whp ish) or a 300whp car that is a street car, with light track work.

THe gtx2867r I think is well suited to this power level and boost. You can see how it passes through the islands where they are at their widest.

The gtx3071r I feel is oversized for 300whp on anywhere but track only. I would use this for track only even those the compressor efficiency is marginally lower because the gt30 turbine is much more efficient in this same range over the gt28. But i want my car (even though its a track car) to be decent in slower targa type racing as well. I also think to really see the gains of this turbo, you need to be aiming for more power, for me, that means built motor, box yada yada. I know some are getting 400hp - over9000 hp on the stock motor / box. But they're not tracking them. (Well I don't believe they are.) And if they are, thats great news for me, it means my selection has a good amount of redundancy.

I realise all turbos will do well more power, you can see this on the comp maps, but the efficicency falls away, and I'm not going there. The gtx2867r has its peak efficency basically right were I want it, basically smack bang in the middle of the revs I use at the track (5500 - 7500) If anything its marginally to low, and I'd prefer it a little higher, but theres so much variations/assumptions/estimations in this, that it would be wrong to look at it in that much depth.

Anyway, thats how I made my decision. :)

So after I checked with a buddy of mine that works for Garrett NZ about my initial estimations, he agreed that they seemed overly harsh, with respect to airflow requirements.

With a couple of his suggestions, I have gone back to trying to make the 63r work.

Basically picked two targets, 320whp on pump, 400whp on E85. Do I see myself using E85, maybe. But, as has been pointed out to me, good future proofing. If I could confidently say NO to the E85 option, it would be best to get a gt2860rs over the gtx2863r. It is basically the same compressor map at moderate boost levels, but much cheaper. But the gtx2863r has much more ability in the higher boost range. Plus its billet wheel should help spool slightly.

Some points:
I chose to make target hp at 7000rpm as this seems to be where most fa20's are making it, give or take. Certainly not at redline. Most likely to do with stock head/cams, amd/or losing VE past 7000rpm.
I pumped up VE from BA's suggest of 0.90 to 0.95. Reasoning being that the motor is high compression, direct injection. Yada yada.
I have plotted down to 3000rpm. Now this DOESNT mean that the turbo will spool and make full boost at 3000rpm. I don't think it will. But, it shows that if the motor is able to get the turbo to spool this early, the compressor can support it.

Anyway, this is what I ended up with.

http://i13.photobucket.com/albums/a2...pscc33102f.jpg

Looking at the results and I'm liking what I'm seeing.
The pump line is showing good efficency. Dropping the boost would result in the same or less efficiency as the islands narrow. The lines past the optimium island are also at about 45 degrees at this point, so less boost, less airflow, but you'll probably be in the same islands.
You could maybe go a little higher on the pump line, to 14 or 15psi. The islands get slightly wider and the past-optimium lines start to steepen. 17-18psi seems like it would place the line where the islands are at their widest, so I would estimate thats where the best efficiency would be found. However, I haven't plotted that, as I think the pump line as is, is pretty much where the knock threshold is (thereabouts) due to the high static compression.

The E85 line is there as sort of the limit as to how far I would push it. Looking at the map, the past optimium lines are near vertical, so reducing hp target, would lower the boost and airflow req'd and lower this line. It would bring the plots into better past-optimium zones, as again, they're near vertical at this point.

So after seeing these results, and getting reassurance that while these are still only estimates, they are most likely more accurate than my own initial estimations, I'm sticking with the gtx2863r.

Just quickly about the turbine housing, I'd already decided to go 0.86 instead of 0.64, purely based on the overwhelming preference of tuners already on this platform that 0.86 performs better. But my buddy pointed out that again this turbo is very similar to the gt2860rs (disco), and the same arguement over 0.86 and 0.64 has been debated to death when talking about the disco. He pointed out that comparing the two spooling on an sr20det, the 0.64 spooled sooner but slower, the 0.86 spooled later and faster. They both reach maximum boost at basically the same point. 0.86 will also cope with the higher rev range much better, and certainly with the e85 line.

i went with the gtx2867r... kinda wishing i would've went with the 3076r after riding in @Sportsguy83's car. actually not kinda, definitely :).

Problem with the 3076r is its a big turbo, made for big power on a 2L.

2863r can get to 400whp on e85 as above, so hard to jump up to a 3076r unless you're talking much higher than that.

No doubt it feels fast though. When it eventually ramps up it'll be like a rocket coming on boost. But at the sacrafice of so much mid range.

I'm sure if it was plotted, you'd see how under utilised the 3076r is. You can see it on the gtx3071 graph above.

This graph gives a good visual to what I'm talking about. You can also see why another guy I know is really wanting to try an efr6758

http://ec1.gorytus.datumweb.com/1/calibra460.jpg

I don't know the specifics of this graph. But the way each turbo comes on says a lot about how they'll be like as a daily.

I'd love to see some real-world boost plots from people. PTuning as I'm sure you are aware is getting good results out of the GT3076R but I don't know how good the on/off throttle response would be with that turbo even in their layout, I can't imagine it's great when not at high RPM. For me it would be a tossup between the GTX3071R and GTX2876R but not sure what size housing I would use. Fuel makes a big difference as well, if I go FI on a stock motor the car will likely only ever see E85 which helps as well with spool and response when tuned correctly.

you'd think it'd spool slow... but it really doesn't at all. it's hardly noticeable compared to my 2867r, and the advantage in overhead is just immense.

i'll be swapping soon.

Got a ride in prototype form and it's surprisingly not laggy at all. Not at 3K rpm in any gear; remember we're talking 12.5:1 compression here.

Agree the overhead is definitely an advantage. For me, I just looked at 400whp on e85, saw the 2863r can do it efficiently, and knew I would never be looking for more than that.

That's awesome. Not only for him, but its all realitive. If a 3076r is spooling that early, then the 2863r will be redonkulous

I had a built 2.5L WRX with a 9:1 CR motor and a fairly small turbo for autocross that was like an on/off switch. It made 17-18PSI almost instantly when tuned on race gas. I'm not sure I'd want that in a RWD 86, I'd want a little more linear boost build up since you can't just mash the gas like AWD... PTuning was supposed to be tuning/dyno'ing a car with the 3071R, I'd be really interested in seeing that especially with a .63A/R housing.

On a vw 2L (10.5:1 DI)I went from a T25 2871r to a vband 3071r w/ a billet wheel and Tial housing. Additional lag was very minimal. The 2871r was great at around 325whp on pump. But the 3071 made over 400whp at the same ~20psi with a bit more timing and 460whp @ 27psi and w/m. Both were driveable, it was just the supporting mods and lack of traction that made all the horsepower not that much fun. But you never realize that till your there.

315whp and sudden boost sounds like heaven.

I wont have to run 55psi in my rear tyres anymore.

I've had a few silvias with similar power. 300whp, no lag, and light rwd car, running 215's. Tha best.

I wanted to see this too, and really rated this kit. But I just don't buy into all that piping not affecting response.

I know when I had my sti, and I went from top to front mount IC, the extra IC piping gave a noticeable decrease in transient response. I didn't like it.

I would love to see this too since im set on the PTuning kit. curious tho, you mention if you ever go FI on a stock motor it would be strictly E85, isnt that incredibly damaging to the fuel and cylinder heads?

I still would love to see someone make a low turbo mounted kit with a TMIC. Sure some hood modification is required but...that response tho...

I doubt it. People have been running E85 through EJ engines for years with no ill effects. By the time I go FI hopefully others will have figured out the effects on the FA20. :D

On my old 2.5L WRX I went TMIC to FMIC and it didn't make much difference with a smaller FMIC. I later switched to a top mount AWIC IC to improve the response and it was totally not worth it.

i can confirm that the 2867r spool is in fact ridiculous. it's essentially instant above the boost threshold, and comes on very, very quick from down low. the high compression really helps a lot.

OP what's pump gas for you? Nice read by the way

Pump for me is 98RON, which I believe is equivilent to US 93-94. But I didn't use that in my calcs, just chose the pump gas setting the boost advisor app.

Ethanol isnt availabe in NZ at the pump (not in the South Island anyway) But we can buy it buy the drum - its about $4/L though.

I'd be more likely to use av gas. At least I can go down to the airport and pump it in. And its only costs 10% more than the 98 anyway.

Might be a good race tune option. Airport and race track are super close, and I think they sell avgas at the track anyway. (I haven't bought it in years, so not 100%)

Nice, I've spent lots of time crunching numbers on turbos and its all great to see it on paper but nothing replaces actual trying it out. Its to bad the cost prohibits most people, including myself, from trying on a few sizes before settling. And no equation factors in preferences.

I've done similar analysis on the turbo options you mentioned and it looks like the GTX2867R would be excellent for this car. Combine it with a twin scroll T3 turbine housing and you would have a very responsive setup with great top end power: http://www.atpturbo.com/mm5/merchant...egory_Code=GTH

At least, that's what I'd do if I were going for a turbo setup on this car.

Yea for sure. You can analyse and over analyse it all you like. Real results are much better.

Best thing for anyone is to go for a ride or drive in others cars, you may find that you really prefer how one setup performs, and want the same thing etc.

I remember when i asked you if youd change a while ago and you said no :cheers:

I happy that i went with a 3076 .63 :)

Everyone is going big these days making me jealous. The sound alone is worth the upgrade cost. I thought I'd be happy with my 2867 forever lol.

Stick with the GT3076 .63 and you'll be happy for a while even with upping the boost down the road. We did not see any gains in power or spoolup on the GTX3071R vs the GT3076R both .63 A/R at around 310-320whp on the dyno. On paper you would expect the transient response to be slightly better with the 71 but it's not enough to tell with the butt dyno. The 30R really shine when you get into the 400whp range, where as the 28s will start to run out of steam.


I'm currently running the GTX3076R with my setup and the spoolup is a tad slower than the GT3076 but again transient response and spoolup seems the same on the road. The GT30R works great on this motor and gives you a lot of room to grow when you get that itch for more power. There's no one magic turbo to fit everyone's need. You just have to pick and choose your weapon based on the type of driving/racing you plan for your car.


You can't really choose a turbo based on how it performed on another motor/vehicle. For example, a GT30R on an Evo 8 4G63, is mediocre, where as that same turbo on a S2000 make awesome power--with the right turbo system setup of course. So I don't know if it makes sense to expect the same result from the same turbo on say an EJ25 vs FA20.

with the fa20, i'm quickly learning to pretty much throw out everything i thought i knew about this stuff lol. i never would've expected the spool that we got out of the 3076 on @Sportsguy83's car.

With the 50% difference in CR it doesn't make sense to, either. This engine when boosted, will be much stronger off-boost than an EJ25 with a big turbo, even if the EJ has .5L more displacement. I'm a fan of the GT28, but having been in a GTX3076R equipped car, the difference is really minimal.

I just can't see how a gtx3076 is feeling just the same as a gtx2863/7r

The 76r graphs I've seen don't hit peak torque until 5500rpm or later. The 67r is closer to 4000rpm

That's basing my understanding that the point where a turbo has reach max psi (in other words, fully spooled" is where max torque is. Or at least where it stops significantly ramping up.

Happy to be proven wrong.

But at this point I can't swallow that stepping up to a 3076r from a 2867r is "not all that different"

My car is hitting full boost at 4300 RPMs (10.5 psi)

Dyno?

Not that I don't believe you. Just want to see it.

No problem, I know where you are coming from :w00t:

Boost plot here: http://www.ft86club.com/forums/showp...postcount=1707

Dyno here (it's a work in progress, but we ran out of time on the Dyno, DISREGARD AFR, was off by more than a point): http://www.ft86club.com/forums/showp...postcount=1701

Hmm interesting.

Based on the dyno, I'd say you're not making full boost till 4800, as that's when peak torque is basically reached. Say max boost was reached at 4300, then I can't see why torque would continue to climb between 4300-4800, if boost stayed constant.

I am counting full boos the moment I go over 10.15 which is the WG spring pressure. That moment is 43XX. I guess boost is not staying constantly at 10.15 and that is why torque climbs up a bit more.

But still, even if it were 4800, it is not 5500 either. I will look at the log again and see if I can point out why 4300 not 4800.

Edit: @diss7

I will also add that the moment where you reach max torque does not automatically means it hits full boost there. There are different tuning strategies and one of them is to ease the torque hit via cam timing, and ignition timing and that peak torque can be phased to a different point than the first hit of full boost.

Agree 4800 is better than the 5500 I was stating. Can't recall who's dyno I saw around 5500, must have been ptuning, but I think they're gt3076r, not gtx. Which would account for the difference between 4800 and 5500

Full boost is reached at 43XX and 44XX in both. I believe the matter of reaching full torque later is a matter of tuning. I've seen JR tunes with a GTX28 where torque peak is at over 5K, where he explained specifically he delayed the torque hit to make it a smooth progression and being easier on the motor.

.82 Backhousing

Looking at Log one

4085 7.34
4117 7.92
4163 8.355
4213 8.79
4277 9.66
4319 9.66
4322 10.095
4424 9.805
4442 10.385
4512 10.24
4585 10.095
4614 9.66
4712 9.805
4734 10.095
4870 10.385
4868 10.53
4970 10.53

Second log:

4109 6.76
4119 7.195
4135 7.195
4193 7.63
4231 8.5
4314 9.08
4354 9.66
4436 9.805
4444 10.385
4471 10.095
4540 10.095
4622 10.095
4640 9.805
4672 9.95
4765 10.24
4819 10.53
4874 10.24

It really depends what your full/target boost is.

Very true. full boost targeting 10 psi will be reached at a different RPM than 20 psi.

Yep that's where it's hard to compare numbers as well. People running these turbos on EJs and SR20s, etc. are generally running at least 18psi or so on a GT30 series turbo, most times more than that because they are looking to make 400+ WHP. Those looking to make 300-350WHP are usually using the GT28 series stuff since it accomplishes their goals with the least amount of lag.

It's difficult on this motor with the 12.5:1 CR. It can help mask some of the lag and if you are willing to deal with a little more lag you can make the power on a GT30 with a lot less heat than the smaller turbos by running less boost.

I can't wait to get my gtx2863r on and get some comparisons. From the performance above of a gtx30, my turbo is going to be awesome as a daily / track hack.

Edit - I'm hoping for ~250ft/lbs from 10psi, from 3500 (or less :D ) to 7000rpm. If it starts to taper at 7000rpm, then I can live with that. Although I think the gtx2867r graphs I am seeing this on, are more a reflection of the gtx28 0.63 turbine housing becoming restrictive. I am hoping my 0.86 with 3" exhaust (with no cats or mufflers) will help/solve this issue. Its got to be the turbine, the compressor (even the 63r) can provide the airflow, somewhat efficiently at 7500rpm. Its certainly very efficient up untill 7000rpm.

Important to note that my backhousing is .82. The .63 would spool faster. :cheers:

Much impress. wow.