GTX3582R and compressor maps

I originally posted this HERE, but it died a death. We got into some interesting talk about split housings, twin scroll stuff and so forth, and it got me wondering - if you combined the right components, just how responsive a setup could you get while still achieving around 550 to 600bhp?

So I started looking at a GTX3582R as they are supposed to be quite the improvement over the previous generation, and they are a close analogue to the T67dbb I currently have. I attempted to plot the Supra on the compressor map but I fear I may have gone wrong, as it appears to me that any faster spooling than what I currently have will simply result in turbo stall, boo.

Have a look at the attachment and see what you think. I'd love to get significant boost at 3000rpm

-Ian

I'm going to bump this as it looks like an interesting subject. Sorry I dont have any input though.

Sorry i didn't came back on your post, it is indeed very interesting, but am with bigbloody turbo on this, i don't have any input.

Thanks for the bumps

Well, it being Christmas and New Year, a lot of people probably aren't knocking around here at the moment. I'm in on nightshift again so I might have a deeper look into this if I get the chance. The ability to plot a Supra engine onto a compressor map would probably be quite useful to the club.

What sort of level are you looking for regarding significant boost? I'm pretty sure that the older GT35r managed to make a decent amount of positive boost at 3k rpm, did it not? I know that that particular turbo was really gassed at around the 600hp mark but apparently the GTX can push a fair bit past that and yet there is no drop in spool.

Will traction not be an issue though? Wouldn't want to make the car undriveable, unless of course you have in gear/RPM boost control etc

I suspect I am getting turbo stall on my gt35 with my new manifold. I'm waiting to get a dry day and a friend with a windows laptop so I can run scal and get some logging data off to have a look at. Hopefully Ryan may have some time to look and analyse for is, he didn't mention any occurring when mapping at SRR last month.

Only happening lower down, sub 4k rpm so far. You've seen my last dyno graph.

Scott I do start making positive boost quite low down, this happens on either 1.0 bar low boost or 1.6 bar high boost where it hit 579bhp, torque was the more phenomenal gain for me. My dyno plot is over on the ecu register thread

The only time ive had problems with turbo stall/surge is if i went to small on the hotside, Ryan got over it on my car by limiting boost lower in the rev range, It dont always show up on the dyno (rolling road) as the car wont be loaded as hard, thats one good thing about hub dyno's, they replicate road conditions better with longer runs.

I changed to a bigger ar and it was fine after that, I just looked in your garage and it says .72, id change to the next size up, i suspect the turbo Ian is looking at would be fine as that will be a .82ar, cant say for sure though.

What sort of level are you looking for regarding significant boost?

I don't know really. I'm interested in what could be achieved. Having pored over various articles and th maths involved, a lot of it is "suck it and see" and "finger in the air", especially when it comes to working out engine VE and so forth. I've still got all my data from the T67DBB tests I did, and I know that's a GT3582R core (CHRA), which is brilliant as I can get the compressor map for that. 6th gear would hit full boost at 3400rpm on the flat, from a 2000rpm roll-on, but it'd be at 0.5bar by 2900rpm which is significant boost IMO. 4th gear was 4000rpm for full and 3200rpm for 0.5bar.

I'm pretty sure that the older GT35r managed to make a decent amount of positive boost at 3k rpm, did it not? I know that that particular turbo was really gassed at around the 600hp mark but apparently the GTX can push a fair bit past that and yet there is no drop in spool.

Funny you should say that, as I've been looking at the aforementioned compressor map for the GT3582R, and it does look like it gives up at about 600bhp. Given a VE of 93% (modified breathing 4 valve motor, according to A. Graham Bell) we need about 810CFM of air at 6800rpm, or 66lb/min. The map for the GT3582R doesn't even reach 66lb/min! It enters choke at 60 I dunno what's going on there because I can't say mine feels anything like choked up so maybe my CFM calcs are a bit off? Perhaps I'm on the raggedy edge but this is only at 1.4bar of boost, I'm sure people have pushed them to 1.6, 1.8 even?

Anyway, the GTX3582R has way more flow capacity, the map goes just over 70lb/min for my chosen pressure ratio, so it's comfortably within what I'd be after. This means the compressor efficiency would go up (i.e. it would heat the air less = more power). So that's good. Also, at 1.4bar of boost it'll flow 29lb/min minimum before it hits stall, which roughly equates to 3000rpm. So I think the GTX3582R would be able to cope with as fast a spooling setup as you could create with housings and so forth and not trip over itself

The last thing I checked as an idea was the turbine RPM. Taking the 6th gear plot, the GT3582R needs about 95,000rpm to supply the air needed for 1.4bar at 3500rpm. The GTX3582R needs about 92,000rpm. This says to me it'll take slightly less time to deliver that boost, all other things being equal, so a bit less lag and full boost earlier in the rev range.

Will traction not be an issue though? Wouldn't want to make the car undriveable, unless of course you have in gear/RPM boost control etc

Well, it's not really about getting max power asap per se, it's more the drivability that results in having this ability waiting in the wings. If a prod of the throttle at 3000rpm gets you 0.5bar in short order, it'll feel very responsive My traction isn't the best compared to some people's reports on here, I can chirp the back tyres in 3rd on a warm day, probably because I only run 265's on the rear. That may be slightly better next summer as I've now got more compliant CW suspension on though.

So anyway, to summarise, I figure that the GTX3582R will give more power for the same boost due to being in a more efficient region, and will also spool up slightly faster (3 to 4%). Mating this with a divided exhaust housing at 0.68AR on my divided manifold would be a hoot

Just need the money now

-Ian

The only time ive had problems with turbo stall/surge is if i went to small on the hotside, Ryan got over it on my car by limiting boost lower in the rev range, It dont always show up on the dyno (rolling road) as the car wont be loaded as hard, thats one good thing about hub dyno's, they replicate road conditions better with longer runs.

 

I changed to a bigger ar and it was fine after that, I just looked in your garage and it says .72, id change to the next size up, i suspect the turbo Ian is looking at would be fine as that will be a .82ar, cant say for sure though.

Good point re the dyno. Your turbo will spin up a lot earlier in the rev range in 6th with a heavy load on compared to 4th gear on a power run.

The surge you got would be because the turbo was spooling up too fast and trying to deliver more air than the engine could ingest. Increasing the hotside AR would slow the spool and "cure" it, but limiting boost electronically is a much better solution, as you'd keep the small AR benefit to turbo spool without trashing the turbo and drivability with surge

Garrett provide a "turbine map" for the three hotside options they offer, but I've got no idea how to interpret them at the moment I think the 0.82 AR can flow 5lb/min more exhaust than the 0.63AR can. Their site just says it can affect top end power as the back pressure increases at higher exhaust gas flow rates. As I don't want any more power I'd be willing to stick ot a 0.63AR for best low down response.

-Ian

Sounds very interesting. Curious about what you said regarding the map of the GT in that it runs out of puff. I can't remember the members name (mrgee or something) but it was the guy with the 2JZ 3 series BMW that had the 35r on and running around 620hp iirc. I have no idea what AR he was running on the hot side though so it's possibly a much higher AR that's allowed him to up the power at a cost of spool. It was definitely gassing as he was up at 2.0 or 2.1bar which, for me, is way past what that tubby should be pushing.

I've heard reports of 700hp from the GTX but not actually seen anything concrete to back it up, just claims through browsing forums. If that's even remotely the case, and the spool is indeed slightly improved (which I see no reason to doubt, technology moves on afterall) then it sounds like a fantastic core. With the smaller exhaust AR and running a sensible boost pressure (around what you have mentioned) it certainly makes food for thought.

Have you thought about the GTX30r? Again according to reports this turbo can hit around the 600hp mark quite safely. I'm not sure what the boost levels would be here though, most likely up past the 1.7bar mark. It was said to flow enough to match the power levels of the original GT35r, not sure what the setup was that brought about these figures, but might be worth a little research since you are being quite thorough so far

Good point re the dyno. Your turbo will spin up a lot earlier in the rev range in 6th with a heavy load on compared to 4th gear on a power run.

 

The surge you got would be because the turbo was spooling up too fast and trying to deliver more air than the engine could ingest. Increasing the hotside AR would slow the spool and "cure" it, but limiting boost electronically is a much better solution, as you'd keep the small AR benefit to turbo spool without trashing the turbo and drivability with surge

 

Garrett provide a "turbine map" for the three hotside options they offer, but I've got no idea how to interpret them at the moment I think the 0.82 AR can flow 5lb/min more exhaust than the 0.63AR can. Their site just says it can affect top end power as the back pressure increases at higher exhaust gas flow rates. As I don't want any more power I'd be willing to stick ot a 0.63AR for best low down response.

 

-Ian

And i was getting that surge on a 76mm turbo with a .81 ar, Id love to understand more of how it all works, CFM's etc is way over my head if im honest.

It does suprise me you are talking of a divided .68, Is it a 3rd of the volume you lose on a divided turbo? thats tiny, id have thought that would surge like hell on a 3.0L.

Sounds very interesting. Curious about what you said regarding the map of the GT in that it runs out of puff. I can't remember the members name (mrgee or something) but it was the guy with the 2JZ 3 series BMW that had the 35r on and running around 620hp iirc. I have no idea what AR he was running on the hot side though so it's possibly a much higher AR that's allowed him to up the power at a cost of spool. It was definitely gassing as he was up at 2.0 or 2.1bar which, for me, is way past what that tubby should be pushing.

The compressor map of the GT35R just about pokes into the 2bar territory but for a max of about 55lb/min, so great for a smaller engine but rubbish for a 3 litre one.

I think the way of working out what flow you need goes like this:

1) Pick your engine displacement and max revs

2) Pick your boost pressure

3) You can now work out your CFM

4) From the CFM and turbo efficiency, you can approximate power

I don't know where turbo size comes into it yet though

I've been working on running 21psi of boost which results in 813CFM or 66lb/min in our engines. This roughly equates to 550bhp +/- 10%, which is about right for a T67/GT35R sized turbo at 21psi.

So I dunno why you'd run one at 2bar and only see 620bhp unless you were mugging yourself and actually reducing power due to overheated air and lowering turbo life due to overspeeding it!

I've heard reports of 700hp from the GTX but not actually seen anything concrete to back it up, just claims through browsing forums. If that's even remotely the case, and the spool is indeed slightly improved (which I see no reason to doubt, technology moves on afterall) then it sounds like a fantastic core. With the smaller exhaust AR and running a sensible boost pressure (around what you have mentioned) it certainly makes food for thought.

There is a rule of thumb that says 150CFM = 100bhp, so you'd need 1050CFM for 700bhp, or 85lb/min. The GTX3582R doesn't even reach 80, it chokes at 75, so I'm unconvinced. I think it'd cheerfully hit 600+bhp at 25psi though.

Have you thought about the GTX30r? Again according to reports this turbo can hit around the 600hp mark quite safely. I'm not sure what the boost levels would be here though, most likely up past the 1.7bar mark. It was said to flow enough to match the power levels of the original GT35r, not sure what the setup was that brought about these figures, but might be worth a little research since you are being quite thorough so far

Funnily enough I was perusing the spec of that as you typed The GTX3076R would be right on the choke line for our needs if it flowed the same as the 3582, but if it needed a higher boost pressure to flow the same CFM it'd be OK. However, the surge line is a lot closer, so if it did spin up to this higher boost faster, chances are it'd surge under 3500rpm Also, the turbine speeds are pushing 130,000rpm at max revs, compared to 103,000rpm for the 3582. I just think it'd be slightly too small, but it'd be worth a try.

And i was getting that surge on a 76mm turbo with a .81 ar, Id love to understand more of how it all works, CFM's etc is way over my head if im honest.

 

It does suprise me you are talking of a divided .68, Is it a 3rd of the volume you lose on a divided turbo? thats tiny, id have thought that would surge like hell on a 3.0L.

Most of this is over my head too I suspect a lot still comes down to having to simply try a turbo in the real world. The fact the T67 turbos work just fine when they appear to be off the edge of their map makes me think I've missed something somewhere.

I've never heard of losing a significant amount of volume on a divided housing, although it must be reduced some. At a couple of hundred quid it's probably one of the cheapest parts to try out in a back to back test. It may be that it only impacts top end power at 7000rpm upwards, in which case I'm laughing

Having said all that, Garrett's catalogue rates the GTX3582R as 450-750bhp har har Perhaps my CFM -> BHP calc is slightly off, or it's US horsepower or something.

Having said all that, Garrett's catalogue rates the GTX3582R as 450-750bhp har har Perhaps my CFM -> BHP calc is slightly off, or it's US horsepower or something.

Yeah, as I said that's what I've read on the net. It's basically quoting those figures though rather than real world testing. Also, I'm assuming those figures are based on high volume engines too. GTX30 is over 600 iirc.

If I had the money to chance, I'd try a GTX3076R. If it delivered the top end power I'm used to and didn't surge down low, it's probably be one of the best street turbos you could get.

This thread may be of interest, with maps on page 4 & 5.

http://www.supraforums.com/forum/showthread.php?615951-Ideas-for-better-spool-from-my-GT35r-setup/page4

Good find, thanks David. Looks like the 3076 is a viable option