Nerd-O-Scope: Honeywell Garrett’s G-Series Turbochargers

by Khiem Dinh


The new Honeywell Garrett G-Series turbo line is the most exciting product launch since the infamous Disco Potato. Everyone knows the GT-Series turbos were getting very long in the tooth. The GTX launch was an upgrade with updated compressor aero and new families of v-band, T3 divided, and v-band divided turbine housings. The GTX Gen2 launch was a further improvement in compressor flow along with new compressor housing castings. But overall, the foundation of the turbos had stayed the same for two decades. Enter, the G-Series, which is a clean sheet design to address and upgrade every aspect of the GT/GTX series turbos. This is a true motorsports grade turbo for the masses. 


The updated parts which make the power are the compressor wheel, turbine wheel, and ball bearing. I’m 99.995% positive the ball bearing is full ceramic with metal cages. You can in fact see the metal cages here. 


Looking closely at the compressor wheel, you can see all the lines from the tool paths used to cut the wheel from forged billet on a 5-axis CNC. Sometimes, you’ll see the blades have a smooth finish instead of the many tool path lines. This is due to the wheel being machined by a process called flank milling, which uses the side of the cylindrical bit to do the machining. Flank milling is cheaper and faster, but the aero design of the blade can be compromised as the tool has to come in directly from the side. This G-Series compressor wheel is point milled where just the tip of the tool is used to cut the shapes. It takes longer and therefore costs more, but the blade aero geometry does not have to be compromised in performance.


The new G-Series compressor aerodynamics are really, really impressive. The map on the left is a GTX Gen2 60mm diameter compressor wheel while the map is the right is the G-Series also in 60mm diameter. In going from GTX Gen1 to Gen2, peak efficiency of the compressor was compromised to gain more flow. The G-Series aerodynamics gained back that peak efficiency AND gained flow! Peak efficiency is 80%, which is pretty much as high as you’ll see short of Formula 1 turbos which use special and very expensive features. Keep in mind the GTX2860R Gen1 only had a peak efficiency of 77%. Also keep in mind that the smaller the compressor wheel, the lower the peak efficiency will be. This is due to the clearance between the wheel and housing being a larger percentage of the inducer area. Larger compressor wheels do run slightly larger clearances, but that clearance is a smaller percentage of the inducer area. Therefore, larger wheels tend to have higher peak efficiency; this goes for turbine wheels too. As another reference, this new G-Series 60mm compressor flows more and has higher efficiency than the old school GT3071R. 


The left map is GTX Gen2 67mm diameter compressor and the right map is the G-Series 67mm compressor. Peak efficiency increased by 7pts (!!!) and flow increased 20%! The efficiency increase translates into more power by reducing turbine power and therefore turbine back pressure. 


Page 1 of 4 Next Page
Bookmark and Share
Monday, November 06, 2017 4:03 AM
It was about time garrett did something to match BW.

I am in the market for a new turbo, and the g25-550 is right what i need, but I am having mixed feelings. I currently run a gt2560r.

First, i cannot get a .70AR in a T25 flange. It means i need a new elbow, downpipe, header and hotpipe too since the size changed. So, costs. And it matters, since switching to an EFR would cause the same problem.

And yet a gtx2867r would be a drop in replacement, and it would work with the .64 AR (although it will make less power and have worse response).

So, cost.

Currently, an internally gated gtx2867r gen 2 with its turbine housing but without the wastegate actuator costs around 200$ more than an efr6758.
Add another 100$ on top of the gts2867r with the turbine housing and you get ... a naked g25 550. Add 260$ for a turbine housing, and then you still need a wastegate actuator.

So is it really worth the expense ? The cost effective upgrade is the gtx2867r. The "bin your exhaust and hotpipe" options are an EFR7163 (i'd be too close to the 6758 limits for my taste) or a g25 550. Not really sure they can compare, the 660 is closer to the 7163 than the 550 is ... Still, i do not need 660hp (heck 500 will be well enough)
Monday, November 06, 2017 8:46 AM
@Crousti, if your car is just a street car... I'd say just do the easy GTX2867R upgrade. If your car is a track car, I'd upgrade the whole setup. The updated materials are much better suited for track use.
Dan DeRosia
Dan DeRosialink
Monday, November 06, 2017 9:43 AM
Do you have any guesstimates on inertia reduction of this method (reducing OD with steel-density materials) vs the EFR method of going to lighter exotic materials? I know Garrett has experience with Mar-M 247 from the motorsports turbos and that may have affected the calculus (and obviously there's other benefits) but I was thinking it would be interesting to know how similar the G-series turbines are to the EFR turbines for a given mass flow.

It seems very much like Garrett is focusing on flow per exterior package size, which alone may be seeing me trying one of these on my DD.
Monday, November 06, 2017 11:36 AM
Hard to say exactly... TiAl has about half the density of Inconel, so approximately have the mass for the same diameter wheel as a ball-park estimate. But the wheel geometries are so different, there's maybe +/- 15% error in estimating mass difference given same size wheels. EFR turbine maps aren't formatted the same as the Garrett maps, you'd have to play around in MatchBot to compare turbo sizes based on turbine flow.
Dan DeRosia
Dan DeRosialink
Monday, November 06, 2017 12:04 PM
Yeah - what I'm thinking is TiAl is also going to be less strong than Inconel/Mar-M so the sections might be thicker, even ignoring that the EFR wheels look to be full back rather than scalloped. I have a hunch that the G-series is pushing more flow per turbine size, but as you said, I might go play on matchbot to see if that pans out. It'd be a really interesting marketing coup if Garrett could claim similar inertia per flow to the EFRs through optimization.

Cool stuff, regardless, and the sweet spot being at a moderate PR makes me happy. I'd assume these are just the first of a new line... that map shape and mention of the GTX3584RS makes me think that some of these newer turbos might be a decent option for a ported 13B
Monday, November 06, 2017 1:31 PM
Are these rebuildable? I believe EFR's are not... this would be of great value to the rotary community, since they like to shoot seals through the turbine sometimes hah
Monday, November 06, 2017 1:44 PM
whats the cost target of these new turbos? are they gonna be about the same as the out going GTX Gen 2's or is there gonna be a premium on that?
Monday, November 06, 2017 2:47 PM
Likely not rebuildable. Turbos this size need to be balanced on a high speed VSR (vibration sorting rig) machine. Also, the way the turbine shaft wheel assembly into the ball bearing, the ball bearing could be damaged during disassembly. My guess is Garrett will have the same CHRA replacement program as they do with the GT/GTX turbos. That said, the Mar-M turbine wheel material should be more robust than inconel to foreign object damage.

Full-race.com had some pricing up on their site. I'd expect these to pull a premium (so close to race wars.... hahaahhaahahahah, sorry). But there is extra cost in the stainless steel turbine housing and mar-m turbine wheel materials, so it's fair to expect a higher price.
Tuesday, November 07, 2017 7:24 AM

Well if cost was not an issue i would not hesitate and just go g25 550. The problem is cost. a 2867r switch would be around 2K€.

A g25 550 setup is at the very least twice that. I'd rather put that 2K€ money as half of a quaife or ppg dog box internal conversion, since the stock box will not hold the power for long. Not to mention i do not want to go blind. If i have to be a guinea pig, i am not paying big money for that.

It is mostly a trackday car though.

I actually based my comments on price on the prices seen on full-race, which as you can see does not include any option (turbine housing is one of them, usually add 260$ for a standard one ...)

Tuesday, November 07, 2017 10:24 AM

Hmmm..... in your case, I'd probably do the transmission first as that will for sure fail. Then run your current turbo setup, assuming it still works, until it dies and then do the full upgrade to G25-550 setup. It will take longer to save up the money, but it will be have longer durability. Everything is a wear item on track cars. Heck, WRC cars swap out turbos like oil changes. I feel the higher temp capable SS turbine housing and mar-m turbine wheel of the G-series will pay for themselves in longer life and you'll get the bonus of better performance.
Tuesday, November 07, 2017 5:03 PM
based on the price on Full-Race's website, they're more than BG's EFR turbos, so that means they need to out perform the EFR. and based on what I've seen out of the Gen2's, I'm pretty sure we'll see that.
Wednesday, November 08, 2017 12:21 AM
Assuming that the G25-550 can achieve the maximum stated 550hp potential using the .92 turbine A/R, to what extent would peak power potential be limited by going down to the .49 A/R?
Wednesday, November 08, 2017 6:00 AM
I can see the underlying question ... no, it is not worth going for a .49 A/R just because it is T25 flanged. Not on anything larger than a 1.5L engine basically.

The backpressure we get from a .64A/R on CA18DETs on a gt2560r to gtx2867r starts to choke the engine around 5000, and that is with 1.4+bar boost. That 550hp would need ~2bar or more boost. You do not want to even try doing that on a 1.8-2L engine and a .49 A/R. It just is too small.

I'd venture saying it would probably be maxed around 300 - 350crank hp, but it would really have no lag. A gt(x ?)2554r would basically be the same, so again, no point in getting the expensive 550 with aht A/R.

Wednesday, November 08, 2017 8:03 AM
In the 2.0L-2.5L engine size range, I think the 0.72 A/R is going to be the sweet spot. I'm just estimating as I haven't seen the full turbine maps with efficiency curves. I see no reason at all to do the GT25-550 with the 0.92A/R; 0.92A/R implies you want maximum power and therefore would be matched to the GT25-660. The place where I see the 0.92A/R being used is on big V8 engines, where you have 3.0L+ per turbo. For the 0.49 A/R, I see that being good on 1.6L and lower. So small displacement 3 and 4 cylinder engines or as a pair on a 3.0L 6-cylinder.
Post Comment Login or register to post a comment.

MotoIQ Proudly Presents Our Partners:

© 2017 MotoIQ.com