Garrett G-Series turbos
#46
EvoM Guru
iTrader: (1)
He only mad 812whp with twins.... So, generally, to calculate what a turbo might make as a twin setup, you double its rated output, and then subtract 150hp. Reversing that math, we get 481whp. Not sure how much boost that was, or what fuel he ran, or how well built that VQ is, but 481whp out of single turbo would be garbage. The video also doesn't show the graph for very long, but it doesn't look insanely special.
#47
He only mad 812whp with twins.... So, generally, to calculate what a turbo might make as a twin setup, you double its rated output, and then subtract 150hp. Reversing that math, we get 481whp. Not sure how much boost that was, or what fuel he ran, or how well built that VQ is, but 481whp out of single turbo would be garbage. The video also doesn't show the graph for very long, but it doesn't look insanely special.
Doing some ballpark math, a VQ puts out 300whp NA. So 812/300 = 2.7 compressor pressure ratio as an estimate. using 10whp per lb/min airflow estimate on a race engine with race gas means ~40-41lbs/min per turbo. Plotting on the compressor map, the turbos are sized about perfectly for having max efficiency at rated power. As an estimate of max power potential on this engine, probably 55lbs/min at ~3.75PR, so 1100whp potential on this setup as an estimate.
I have it on good authority AMS got ~620whp out of the G25-660 on their Evo turbo test rig. Completely maxed out, no intake restriction, no exhaust restriction. It's probably more realistically a 550whp turbo which is on-par with a GTX3071R Gen2.
The EFR7163 is the natural competitor and the compressor also flows 60 lbs/min but only has a max efficiency of 74%. Doing some more ballpark estimation for rotational inertia which is a function of mass x diameter^2. (54/63)*(54^2/63^2) = 0.63. EFR wheel is half density, so.... 0.63 is close to 0.50. compressor side, 67^3/71^3 = 0.84. So using rough math, g25-660 is just a bit higher rotational inertia than efr 7163. The MarM247 turbine wheel of the G-series is much more robust than the TiAl of the EFR. For those who don't know, MarM is the material used in the Garrett Motorsports turbos, such as the ones abused to hell in WRC with all the anti-lag. Keep in mind improved compressor efficiency and turbine efficiency reduce lag. So a turbo may have higher mass and inertia but if it has higher efficiency, then can have the same transient response (or better) than a turbo with lower mass and inertia.
Last edited by spdracerut; Apr 16, 2018 at 11:57 PM.
The following users liked this post:
Construct (May 8, 2020)
#49
Evolved Member
Interesting to note that comparison overlay. Garrett (red) is spinning 15k RPM faster at the end of the efficiency (around 60 lb/min) yet is more efficient than the EFR. Usually speed = friction = heat = efficiency loss.
#50
Evolving Member
Join Date: Jun 2004
Location: Australia
Posts: 371
Likes: 0
Received 0 Likes
on
0 Posts
Provided Garrett produces more turbine housings in T2 T3 T4 , and with integral gate options , I think they may have the legs on the Borg .
Turbos get overspun and if they don't fly apart its a bonus .
I'd like to see what G series they come up with to compete with the 7670 .
Turbos get overspun and if they don't fly apart its a bonus .
I'd like to see what G series they come up with to compete with the 7670 .
#51
Provided Garrett produces more turbine housings in T2 T3 T4 , and with integral gate options , I think they may have the legs on the Borg .
Turbos get overspun and if they don't fly apart its a bonus .
I'd like to see what G series they come up with to compete with the 7670 .
Turbos get overspun and if they don't fly apart its a bonus .
I'd like to see what G series they come up with to compete with the 7670 .
#52
He only mad 812whp with twins.... So, generally, to calculate what a turbo might make as a twin setup, you double its rated output, and then subtract 150hp. Reversing that math, we get 481whp. Not sure how much boost that was, or what fuel he ran, or how well built that VQ is, but 481whp out of single turbo would be garbage. The video also doesn't show the graph for very long, but it doesn't look insanely special.
#53
Evolved Member
iTrader: (1)
The G25-660 is actually a pretty good competitor to the 7670. I expect them to be withing 20whp of each other. The 7670 is really like a 550whp turbo with the GTX3576R Gen1 being like a 600whp turbo. The G25-660 should do or be really damn close to 550whp. The turbine wheel efficiency of the 7670 is really poor which is why it's held back on power. I'm actually surprised people buy the 7670 with the availability of the 7163. The 7163 makes a little less power, but should more than make up for it in spool and transient response.
The EFR7670 compressor flows more (4 lb/min) than the EFR7163, and has the bigger turbine housing - try making the same peak numbers with a .80 A/R housing on the EFR7163 as you would with the .92 or 1.05 A/R on the EFR7670. Each turbo has its place.
If the G25-660 is a 550whp turbo, it might spool faster, but can't make more power than the GTX3576R. I think it will be right on par with the EFR7670 IMO.
#54
Rating a turbo for whp is difficult, but head-to-head, the GTX3576R is right on par with the EFR. The EFR will out-spool it badly, and respond quicker.
The EFR7670 compressor flows more (4 lb/min) than the EFR7163, and has the bigger turbine housing - try making the same peak numbers with a .80 A/R housing on the EFR7163 as you would with the .92 or 1.05 A/R on the EFR7670. Each turbo has its place.
If the G25-660 is a 550whp turbo, it might spool faster, but can't make more power than the GTX3576R. I think it will be right on par with the EFR7670 IMO.
The EFR7670 compressor flows more (4 lb/min) than the EFR7163, and has the bigger turbine housing - try making the same peak numbers with a .80 A/R housing on the EFR7163 as you would with the .92 or 1.05 A/R on the EFR7670. Each turbo has its place.
If the G25-660 is a 550whp turbo, it might spool faster, but can't make more power than the GTX3576R. I think it will be right on par with the EFR7670 IMO.
This shows a GTX3582R vs GTX3576R, both in 0.82 A/R housings on a Subaru. I would actually say this result is a bit extraordinary as I typically expect the 3576 to spool ~500rpm faster than the 3582 in the same A/R housing on a typical 2.0-2.5L engine. I should note this dyno is Church's in Los Angeles which puts out pretty much crank hp numbers.
The 7670 is really on par, peak power-wise, with a Gen1 GTX3076R or Gen2 GTX3071R. By my estimations, the G25-660 will make the same power as the Gen2 GTX3071R but have about the same rotational inertia as the EFR7163.
#55
Evolving Member
If ya'll haven't driven something with a twin-scroll EFR on it, you're missing out. The way those things respond on the street is exceptional. Comparing dyno graphs between an EFR and another turbo is silly IMHO, they'll never show the response difference.
Almost had the G25-660 car up and running this week, but turns out we had a crank that wasn't properly machined for the flywheel. I'll get those numbers for you guys as soon as I can.
Almost had the G25-660 car up and running this week, but turns out we had a crank that wasn't properly machined for the flywheel. I'll get those numbers for you guys as soon as I can.
#56
Comparing dyno charts is still valid as when it comes to certain events like drag racing, power is what matters. Also, anti-lag features can be used to eliminate the differences in rotational inertia on a track car. On a road course where engines spend a lot of time at higher rpms, reduced rotational inertia becomes less of an impact. Same with PDK/dual clutch transmissions. All that said, the goal is to always improve transient response which has traditionally been through reduced inertia and improved compressor and turbine efficiencies. With e-turbos coming soon, reduced inertia will become less important with an emphasis likely transitioning to maximum efficiency.
#58
Evolved Member
Nice! Its the 660, right? Which hotside? Look forward to seeing how it goes, hoping you have a setup to make the most of it I saw a clip show up of an EVO 9 on youtube a couple of weeks ago claiming to have a G25-660 but there was (like seems common for dyno pull vids) no numbers, no pic of a plot or anything but it did seem to stand up well.
#59
Nice! Its the 660, right? Which hotside? Look forward to seeing how it goes, hoping you have a setup to make the most of it I saw a clip show up of an EVO 9 on youtube a couple of weeks ago claiming to have a G25-660 but there was (like seems common for dyno pull vids) no numbers, no pic of a plot or anything but it did seem to stand up well.
Thanks,
yes, it's the g25-660 with 0.92 A/R hot side. Unfortunately, at the moment, the engine can't handle the full potential of this turbo since that the internals are almost stock. A 2.0l forged long rod will be come in future; for the moment will be see about the spool and response for a 450 HP configuration. I also saw that video on youtube and i asked some info, but without response.
#60
Thanks,
yes, it's the g25-660 with 0.92 A/R hot side. Unfortunately, at the moment, the engine can't handle the full potential of this turbo since that the internals are almost stock. A 2.0l forged long rod will be come in future; for the moment will be see about the spool and response for a 450 HP configuration. I also saw that video on youtube and i asked some info, but without response.
yes, it's the g25-660 with 0.92 A/R hot side. Unfortunately, at the moment, the engine can't handle the full potential of this turbo since that the internals are almost stock. A 2.0l forged long rod will be come in future; for the moment will be see about the spool and response for a 450 HP configuration. I also saw that video on youtube and i asked some info, but without response.