Garrett Gt3071r Discussion & Results

[BHDave]

I finally managed to get mine road worthy again yesterday.

Swapped for another known working actuator that i pinched off a mate, went for a freeway run and had a play at low boost.

The problem is still there, but i found that if i bought boost on slowly it would stabilise at .8bar for about 1000rpm and then start to creep up.

I even had a play with the boost controller and i could bring boost on, get a 1psi spike, drop back and hold boost steady for 500rpm and then it would start to creep, so that seems to be able to effect the boost response and isn't just crapping out.

I'm starting to think the wastegate size is the issue, but dangerman had no issues, maybe because he runs significantly higher boost so less is by passed, maybe because he has a bigger compressor so isn't bypassing as much at low revs anyway.

This week i will be ordering a new actuator anyway. Somewhere around the 16-17psi mark i think. If that doesn't work then the turbo has to come back off and the wastegate port gets enlarged.

[Dale FZ1]

If boost is creeping as revs rise, it's starting to sound like wastegate size is inadequate. Dave was yours one of the production upgrade housings, with the bigger valve already fitted? Or do you know the size of the valve and port? It would be interesting to know how it compares to a HKS Pro S in that respect.

I also think there might be a few too many differences between the GT30 and GT35 to bring direct comparison with dangerman4's setup. The casting may be the same, but there is a much bigger vent area through the turbine. But I do see your point.

I know skylinecouple was/is running a 0.82 IW housing on his 3071, and don't believe there was any boost control issues there. But I'm not sure if he had the upsized wastegate.

[BHDave]

I'm not sure what you mean by production upgrade housing. I was under the impression that the garret IW housings were the same between the 3071 and 3076 as it's the same core and turbine, just different front end.

No idea on the relative sizes of the wastegate, I've posted some pics of the turbine housing earlier in this thread if that helps.

How much metal needs to be left to ensure a decent seal on the wastegate flap?

[Dale FZ1]

I believe Cubes had experience with the availability of an upsized wastegate port + valve after Garrett had a few of these GT30 housings running in service. I think there was a -M suffix put on the part number to indicate it. The size change wasn't massive, and the job on an existing unit isn't big either. But 3-4mm would be a large increase in cross sectional area. An upgrade would probably require a larger valve.

I would think that the valve would need around 3mm to be seating "nicely" and allow for the bit of float that occurs.

[discopotato03]

Dale I am interested to hear any details about a real GT3071R with a 0.82 A/R turbine housing as long as its the proper Garrett GT30 IW one .

A little OT but in a vague sort of way Garrett's GT2860RS is sort of along parallel lines to the real GT3071R - turbine size/flow a little higher than usual in relation to compressor size/flow capacity .

The Garrett production engineer who spec-ed the 2860RS wanted something with low and mid range punch for a Mazda Miata (MX5 here) BP1800cc DOHC 4 cylinder . BTW the engine was also free revving and not "choked" .

He used the best available GT28 turbine (~54mm NS111) with a large trim (62) 60mm GT compressor . The interesting thing is that he also used housings on both sides that are reasonably large in relation to the wheels .

The exhaust side housing is a 0.86 A/R GT28 where the majority of GT28's with 60mm diameter compressors use the 0.64 A/R GT28 turbine housing .

The compressor end housing is a T04B family one which on the surface is big for a 60mm compressor but it certainly would make for a free flowing path inside the volute and to its outlet . Also the current trend is to use a larger diameter back-plate on the compressor end which gives a larger diameter diffuser (air path between the exducer tips of the comp wheel and the volute section of the comp housing) . This tends to show up on the comp maps as wider islands along the horizontal axis so basically the surge line and choke flow lines move further apart . A recent example of this is Mitsubishi with their Evolution 9 Lancer , the comp housing/back-plate/diffuser section is larger in diameter than the Evo 6-8 .

It makes me wonder if these real GT3071R's "idle" or windmill a bit faster than some other spec turbos so do the boost and torque building thing fairly rapidly . I always reckoned that turbos like that GT2860RS thrived on the larger A/R turbine housing because its low restriction lets you run lots of light load ignition advance which really pulls up the low end torque and throttle response on many 2L fours .

Its impossible for me to say what the results of a 0.82 A/R real GT3071R would be on an RB25DET which is why if anyones done it I'd really like to know . It would be easy to presume that it would react like a GT2860RS on a 2L four but there are lots of little variations that could easily shoot down that theory in the real world . Power wise you think it would be good given the compressor and housing these GT3071R's use but the make or break is always the turbine response and boost threshold .

I go with the theory that removing restrictions in a production turbo engines gas paths is the way to make power - and to do so without high boost pressure . The way I see it a turbo engine on boost is making larger capacity engine exhaust flow so you have to deal with this if you want serious broad ranging torque and a reliable healthy engine . Its also why I like the idea of big enough turbines and their housings with modest compressor wheel capacity . Free flowing exhaust paths to keep the engine happy and turbine responsive + no excessive comp wheel to rob turbine response and low down engine power .

Anyway looking good so far , cheers A .

[Dale FZ1]

It would be best to hear the impressions direct from skylinecouple, but I do know he has a pretty busy schedule and checks the forum infrequently.

He is running fairly conventional supporting system upgrades with the -23 60mm turbine 0.82 IW version of the 3071. Sourced from the US, and yes it is the same family of housings we are all talking about.

Numbers wise I think he indicated it's making around 260kW @ 17psi and running reasonably rich with conservative timing.

Evidently it runs quite linear delivery with maybe a touch higher boost threshold than his old GCG high flow, but the higher end breathing is freed up and it pulls very hard right through to above 7000 as boost builds quickly. Apparently wet roads and those steel joining strips on some roads can result in unexpected (unintended) traction loss that never happened with his high flow turbo setup.

We spoke at length, and he felt that there may have been no practical benefit of the 3071 + 0.82 over a 3076 with the same turbine housing. But the whole time he enthused over the linearity and overall feel of the setup. If you read his thread regarding this turbo, that was one of his primary aims along with removing turbine back pressure.

[BHDave]

I'm also interested in his results. I momentarily considered the bigger turbine but just couldn't do it. I am a self confessed response junkie. I am happy to accept a choked top end if i can put my boot in at 3k and get a decent amount of boost.

Actually, kind of related. 5th gear, flat road, i have positive pressure at 2k and .5bar by 2300. It's silly.

Anyway, back on topic. I had a look at the pics i posted earlier (page 2). I'm pretty confident that a larger wastegate could be fit in there easily. It looks like the current one doesn't fully cover the whole machined surface either so, as you said dale, not a massive job.

I'll be having a run on a dyno next weekend, so i can confirm that the actuator is opening the gate fully and not causing a restriction that way.

I really hope people aren't waiting on my results. I had hoped to be sorted by now.

[discopotato03]

Well then skylinecouple would be the one with the answers as far as the larger 0.82 housing is concerned .

I knw I'm comparing apples to oranges but I found the GT2860RS really nice on the FJ20ET I used to have . It really brought home to me how much you get when you crank up the ignition timing in the low areas , even at wide open throttle and low revs it makes an incredible difference . More torque to the point when you don't have to get into boost all the time just to have a little bit of squirt around town . That was always friendly as far as fuel consumption was concerned .

I know what skylinecouple means when comparing a real GT3071R to a real GT3076R , a few years ago people in the US tried these 3071R's out after using GT3076R's and unless they ran the 76R to its absolute limit with large A/R turbine housings they sometimes didn't see a lot of difference between the two . Its not impossible that they didn't advance the timing up and get the lower end advantages .

I would have to look at the maps/lists but I remember thinking that the maximum flow capacity of the 56 trim 71mm GT compressor was quite similar to the 76mm 52 trim GT wheel .

Something else to think about when comparing the GT3071R to the GT3076R is their compressor housings . They are both T04E series but the 71R's is 0.50 A/R non port shrouded where the 76R's is 0.60 A/R and is port shrouded .

Since they both use a comp end back plate (adapter ring actually) the same size the 76R effectively has less diffuser area because its OD is larger at 76.2mm rather than 71.1mm .

Now technically speaking the T3 comp housing/back-plate/diffuser set was designed for a 60mm compressor wheel , T04B set designed for a 71mm OD compressor , the T04E set was designed for a 76mm OD wheel and the T04S designed for an 82mm OD wheel .

When Garrett started using 60/71/76/82mm GT series compressor wheels they often went up a housing size for the same/similarly sized (OD or exducer sized) T series wheels .

Examples of this are the GT2860RS (60mm wheel in housing set designed for a 71mm T series wheel) , GT3071R's (71.1mm wheel in housing set designed for a 76mm T series wheel .

The GT3076R is different in that it has a 76.2mm wheel in a housing set designed for a 76mm T series wheel . You could probably put this down to it being a HKS spec cartridge and they need to be a bit space conscious for road car apps . Possibly to give its best upper end airflow the GT3076R needed a T04S comp cover and adapter ring set . In truck form these wheels are run in 0.58 A/R GT40 comp covers and they really dwarf the wheel .

If anyones been following what Forced Performance has been doing in the US they'd know about the "HTA30R" with its (I think) Borg warner 7/14 blade comp wheel in a T04S comp housing set .

Rambling again , long and short is that the 3071R probably has a more optimum sized compressor housing set for the size of its wheel than the 3076R does and it possibly may even out the two flow capacitys more than the size difference of the wheels suggests .

Cheers A .

[juggernaut1]

Interestingly, I been having email discussions on the 3071 vs 3076 albeit in a twinscroll configuration with Geoff Raicer of Full-Race fame.

Essentially, it comes down to whether the engine in question can swallow the airflow afforded by the 3076 compressor. Some engines just can't make efficient use of the airflow of the bigger compressor due to basic head design and set-up. An indication of this is that the 3076 will only have the same or marginally more power than the 3071.

Another point raised by Geoff is that while the spool difference between the two might only be 100 - 150 rpm (in a twin scroll ATP .78 a/r ex housing) the 3071 is the more responsive of the two. I wonder how much the port shrouded compressor cover of the 3076 hurts it's efficiency.

[discopotato03]

Yep Jeff has lots of hands on experience with RB's for someone in the US . He put lots of work into that GTR engine and full driveline conversion in his S14 240 SX . He did have an R33 GTS25T at his shop for a while too .

The twin scroll theme changes things markedly on the exhaust side and in the wallet area , most people here can't justify alternative exhaust manifolds and external wastegates and are more likely to use a larger engine I'd say .

If you go to the turbobygarrett site you can see the compressor maps for the real GT3071R/GT3076R and they are good in their own respects .

If you wanted high Hp/Kw numbers from a 71R you have to spin it faster than a 76R but in the middle pressure ratio (boost pressure) range the 71R is wider from an air flow (mass flow) perspective .

I think to run the 76R to its air flow limits would require a larger diameter compressor cover set .

The port shroud is an each way bet meaning it can bypass excessive air flow if the engine can't swallow it all . My understanding is that when the engine can't accept extra air that extra amount that the higher or full height blades are trying to pump goes out through the radial slot and back to the snout . So because the wheel is running in a partially unloaded state its not demanding excessive drive power from the turbine , so the turbine can spin more freely and be less of a restriction to the exhaust gasses flowing through its blades - reduced turbine inlet or backpressure .

Brett Lloyd mentioned once that when the engine and compressor get beyond their surge window the radial slot can actually feed slightly more air into the wheel than would be the case if the slot wasn't there .

Anyhow the GT3076R's comp map is quite reasonable , the burning question is does the GT3071R do a similar job and be more responsive by being a little less taxing on the turbines limited shaft power supply . The only way to find out is to test them back to back and tune the engines to the nth degree .

Cheers A .

[juggernaut1]

Yes the port shrouded comp effectively recirculates the air back out of the comp cover through the slots if the engine can't otherwise swallow the airflow being supplied (usually at lowish engine rpm and big compressor/smallish turbine).

I agree that the slots in theory could supply more air but a discussion I've had with someone who has airflow tested port shrouded comp covers shows the slots actually cause some turbulence in the compressor housing.

[-33-]

I'll be interested to see whether machining out the gate does anything. I'm not so convinced in my case - if it was a steady creep all the way to 7k i'd have machined it already, but apart from actuator, i still can't see any other feasible explanation as to why my boost curve drops back down to proper levels near redline. My wallet is drained atm, and i'm already needing new coilpacks (i was hoping the stockies would last a bit longer) so i'll hold this one off for a while. My tune seems to be a little too on the rich side as well, although i'll see whether thats truly the case when i put the new coils in - can't really tell amongst all the misfire.

Regardless of the little nagging issues i still reckon the turbo is an animal of a thing. Response junkies look no further. With nice cams and cam gears (later), who knows what it could do

[Dale FZ1]

skylinecouple would be the one with the answers as far as the larger 0.82 housing is concerned .

I know what skylinecouple means when comparing a real GT3071R to a real GT3076R , a few years ago people in the US tried these 3071R's out after using GT3076R's and unless they ran the 76R to its absolute limit with large A/R turbine housings they sometimes didn't see a lot of difference between the two . Its not impossible that they didn't advance the timing up and get the lower end advantages .

skylinecouple actually seemed to have an opinion from the other angle, and last I heard from him he was about to replace the 3071 with a 3076.

If you go to the turbobygarrett site you can see the compressor maps for the real GT3071R/GT3076R and they are good in their own respects .

If you wanted high Hp/Kw numbers from a 71R you have to spin it faster than a 76R but in the middle pressure ratio (boost pressure) range the 71R is wider from an air flow (mass flow) perspective .

I think to run the 76R to its air flow limits would require a larger diameter compressor cover set .

Anyhow the GT3076R's comp map is quite reasonable , the burning question is does the GT3071R do a similar job and be more responsive by being a little less taxing on the turbines limited shaft power supply . The only way to find out is to test them back to back and tune the engines to the nth degree .

Looking at the maps, picking an arbitrary 2.5PR that would be necessary to pull the sort of hp these things are capable of, the 76 actually has a wider map than the 71. The difference is that the 71 actually hits on its peak efficiency through the midrange rpm (say at peak engine torque) while the 76 just misses it.

Running a 0.7A/R TO4S type cover might yield a higher flow capacity, but have a bad impact on spool/response. There's a lot more internal capacity to pressurise.

I'd say it's more likely that to pull every last pound of airflow from the 76 it would probably need the 1.06 turbine housing and there goes your bottom end spool. I can advise that the 3037/3076 running a 0.87 HKS turbine housing does not feel compromised on a stock RB25 for overall response, boost threshold, or maximum power output. I know it is a very good unit, but there are varying preferences out there.

I think BHDave is on the money with needing to spin the 71 harder, earlier, in order to make it respond and get the theoretical response improvement over a 76. But the smaller 0.63 housing will choke a little earlier. I'd say that driver preference and maybe intended use of the car would be a deciding factor here.

Another point raised by Geoff is that while the spool difference between the two might only be 100 - 150 rpm (in a twin scroll ATP .78 a/r ex housing) the 3071 is the more responsive of the two.

Regardless of the little nagging issues i still reckon the turbo is an animal of a thing. Response junkies look no further.

It would be prudent to look at "response" from a couple of angles. You have boost response, engine response, and the intangible "feel". It might be that the 71 particularly with a 0.63 housing is a bit too touchy for some, and gives more/quicker torque delivery than desired. Personally I want tyres that are turnin', not burnin'

Playing around with jetting on bikes a couple of years back, one of my dyno operator's tricks was to review an acceleration run through the gears, vs time. That test could be done using any particular rpm for change point, and was not only a way of looking at power output but type and rate of delivery. Believe me that is an important factor on two wheels.

That might just put some of skylinecouple's choices regarding the 0.82 housing and a 76 v 71 choice in perspective.

Too much off topic, and apologies to BHDave. I am pretty interested to see what happens over the next week for him, and hopefully a bit of work around the wastegate area will get the boost controlled properly.

[-33-]

Personally I want tyres that are turnin', not burnin'

Haha, well some good old pedal control (and fat tyres of course) will solve that...

[Dale FZ1]

Without a doubt, and point taken.

Where I saw the advantage of the 71 was right down low. The one weakness I perceive in the 3037/3076 after stepping up from a 2871 hybrid is from 2000-2500rpm. Previously it was very athletic in that range, but now it feels "torquey" while gathering momentum. By way of comparison with Dave's observed rate of spool, I will see around 4psi @ 2400 if I accelerate it in 5th from 2000rpm. Not weak or laggy, just not ballsy in that range. From 2700-7000 is just a massive torque wave.

I have a strong interest in the outcomes here, mainly because until this HKS came along, my intention was to order the 3071 / 0.63 combination.

[BHDave]

Mine was by no means ballsy at those revs either, just thought it was funny how low the thing got going and had still made 260rwkw without even trying.

It's not something i do regularly, i was just sitting in traffic and thought i'd play, for science, and because it's all still shiny and new.

[BHDave]

I know it's a 3071 thread dale, but i'd be interested to see your 3037 results too.

more food for thought for prospective upgraders.

[BHDave]

It's confirmed. The wastegate was wide open and boost kept on rising

Absolutely useless half throttle run below. Even though it isn't shown, shoot 6 uncorrected for air temp. It at least shows the response (and top end if you're that way inclined) potential. Midrange power is down to buggery.

Edited April 13, 2008 by BHDave

[Dale FZ1]

I know it's a 3071 thread dale, but i'd be interested to see your 3037 results too.

more food for thought for prospective upgraders.

Well this should give an indicator of what a 3037 can do with reasonably good tuning on an internally stock engine and running a free flowing (reasonably noisy) exhaust. That is ~ 290kW for the mathematically challenged, with boost peaking @ 18.6psi and tapering to 18.0psi using a Turbotech adjuster. AFR is held steady @ 11.95-12.1, and max reported knock was 16.

Slightly shaky looking graph at the top end indicates it wants more ignition, but that results in higher knock figures. I have a solution to that issue.

Running with an extra baffle, power drops 24hp to ~ 270kW, and the increased backpressure holds boost back to 17.3psi.

There are no complaints about either delivery or quantity of power.

[BHDave]

The enlarged wastegate.

39mm flap, and as best i can measure 34mm port at the top.

They couldn't fit anything bigger without machining into the turbine outlet to fit the wastegate arm (you can see where they had to take a bit out already)

It's a shame about the bit of a lip at the bottom taking up some of the area, but unfortunately you are limited by the housing design and taking any more out there will reduce the wall thickness significantly.

Anyway, i'll give it a go.

I have a new 16psi actuator to fit up as well so if i still have boost control issues then garrett may get a nasty phone call regarding selling a housing that can't even bypass enough to maintain stable boost on the smallest of the turbos that it fits...