A bit of history (part 22) – Turbos, turbos, everywhere
- ainsworthashley
- Nov 15, 2017
- 7 min read
Scoot back in time to October 2013, and one of the first changes I had decided to make as a result of having to pull the whole build apart again to fix my mystery noise…the turbo. It took me the best part of 12 months to decide which way to jump, and felt like one of the biggest (and hardest) decisions to reach a conclusion on.
I had been using a standard Garret GT3076r with a 0.63AR exhaust housing (the smallest option), the idea being to make decent torque without a crazy high boost threshold…the 76mm compressor wheel being capable of flowing plenty of power, and the smaller turbine housing in theory meaning less exhaust gas flow was needed to spin up the turbo and get it on song.
That said I had to admit I had been somewhat disappointed with the results, the build had made peak power & torque at just 1.4bar of boost, with any additional pressure simply generating excessive heat and reducing power outputs. The setup did put down ~400hp/400ftlbs at this pressure, which is very respectable, but the engine somehow felt strangled, with the boost threshold still seeming a bit too high despite the smaller turbine housing.
The general consensus seemed to be that the turbine housing was simply too small for the build…
Additionally I was concerned that the exhaust was overly restrictive as it tapered down to 2.5 inch at the back end in order to mate to the HKS Hi Power section, and then there were also doubts over how much of a bottleneck the Radtech charge cooler core was creating…!
Prior to diagnosing the mystery noise I had simply planned to swap in the mid-sized 0.82AR turbine housing, but the little voice at the back of my mind kept questioning the exhaust and charge cooler setup as well…and I had to admit I was missing the punch you get with higher boost and torque, I just wasn’t convinced the GT3076r could provide this.
So I decided to use the full engine rebuild as an opportunity to address some of these doubts, and get things “right”…more on the charge cooler and exhaust later, but for now let’s focus on the turbo itself….
Now anyone who has ever given any real in depth thought to spec’ing a turbo charge will understand how difficult it can be…there are so many different makes and sizes, with plenty of overlapping sizes in between to confuse things! You can get stuck into efficiency maps, and lots of theoretical mathematics, but then it generally seems to come down to past experience and looking at the results other people have achieved with similar builds.
In an ideal world you would fit pressure sensors to your exhaust manifold, bang on a turbo and measure pressure ratios and use this data to inform your decision…unfortunately, however, in the real world turbo dealers do not offer a “try before you buy” service, so unless you have so much money that you can simply buy 3 or 4 different turbos and try them all, you have to try to “guesstimate” which unit you want first time.
The following really is overly detailed, and probably not particularly interesting to most car enthusiasts out there, but I wanted to share my thought processes with regard to how I chose which turbo to buy, warts and all! Since writing most of this back in 2013 I have learnt a lot more, but I haven’t gone back and changed what I wrote at the time…so yes, it is overly simplified (and downright wrong) in places, but it does give an insight into how I got to where I am!
Precision have a range of turbos, and I’d heard some really good things. Garret had recently introduced the GTX range, and of course there are the Forced Performance billet HTA compressor wheels to think about. I excluded the Borg Warner range from my search for some reason, I think it was down to the size of the units, and the apparent mandatory internal waste gate design (having already acquired a Tial wastegate).
So how to choose? Well I got in contact with the dealers and manufacturers and discussed my options a LOT with them…!
I was nervous of the Precision range as they don’t use water cooling on the ball bearing units (the apparent better option), which would be fine in a front engine layout, but with the engine in the back you don’t have the benefit of all that air flowing through the engine bay to help cool things down. I contacted Precision, and, well let’s just say they were less than helpful! I think they told me something along the lines of “we are a manufacturer and not in a position to answer specific questions or offer any form of advice on the use of their products”….very helpful, and a great way to generate sales!
Garrett and Forced Performance, on the other hand, bent over backwards to answer my questions, and gave me as much time as I wanted talking through my options, and the pros and cons of various setups.
So on this basis I narrowed my choices down to the following:
GT3076 (Garrett)
GT3076HTA (Forced Performance)
GTX3076 (Garrett) / GT3079HTA (Forced Performance)
GTX3576 (Garrett) / GT3579HTA (Forced Performance)
GT3582HTA (Forced Performance)
GTX3582 (Garrett)
GT3076
This was my starting point, and only on the list as a point of reference: in theory with my new setup and a larger exhaust housing this could make over 500hp supposedly with a nice early spool up, and excellent transient response.
GT3076HTA
A significant upgrade to the GT3076m with the HTA compressor wheel fitted- this should easily make 500hp and spool up earlier than the GT3076, but I figured it was too small for the savage kick in the butt I was looking for.
GTX3076 / GT3079HTA
I’ve grouped these two together because they appear to be pretty much the same turbo. When Garrett introduced their new GTX range they have been slightly misleading with the terminology: the exducer diameter of the compressor wheels actually being larger than the “76” implies. The HTA compensates for this by using a larger compressor wheel of a modified design, and so to all extents and purposes I would expect both units to perform similarly.
In their stats Garrett claim the GTX3076 is capable of a huge 640hp(!), however real world testing (thanks to the Scooby and Evo boys) seems to suggest the “30” size turbine wheel won’t flow that much power, and results generally seem to show a 30-40hp increase over the GT3076…so basically not really worth the cost of upgrading over my GT3076.
GTX3576 / GT3579HTA
Again, (rightly or wrongly!), I have put these two together in terms of expected performance.
The idea here is that by coupling the compressor wheel from the GTX3076 with a “35” size turbine wheel, you remove the restriction of the 3076, and can actually flow enough gas to make the bigger power. By keeping the smaller compressor wheel, you help keep a reasonable spool and transient response.
I’ve “heard” it described as spooling like a GT3076, but flowing like a GT3582(!), although in reality the manufacturers have advised to expect an increase in boost threshold of 300-500rpm by stepping up from the 30 to the 35 wheel.
GT3582HTA
This is a true 82mm compressor wheel, and represents an upgrade to the GT3582 in terms of both spool and power delivery, and I’ve seen impressive figures from this unit on another MR2 making over 550hp.
This would certainly pack the punch I’m looking for, but after talking to the other MR2 owner about how ferocious boost kicks in, I couldn’t help but wonder if this was just a bit too laggy for controlled road and track driving? Well- for my driving skills anyway!
GTX3582
Finally we have the new Garrett GTX3582 with a claimed top end of 750hp!
Before I understood that the new GTX compressor wheels were actually larger than the naming convention suggests, I had assumed this would be similar to the GT3582HTA, and I’ll admit I was very keen on this unit…improved spool and a top end over 600hp…
However further research put the peak power of this unit close to >9k rpm using my engine, which is just a bit too far for my liking, and I had to conclude this was just too big for my engine given how I wanted it to operate.
So putting this altogether the real choice came down to three units for me:
GTX3076 (Garrett) / GT3079HTA (Forced Performance)
GTX3576 (Garrett) / GT3579HTA (Forced Performance)
GT3582HTA (Forced Performance)
And I have to say- all credit to Owen Developments and Forced Performance, who both massively impressed me with the levels of support they were happy to offer me. They both spent as much time as I wanted chatting details through, and taking the time to understand what I was trying to do, and helping advise the best way their products could help me achieve it.
I went to them specifying an absolute top end of 600hp, but fully expecting to make less than this, whilst maintaining a car that was driveable on road and track: so boost threshold and transient response had to be preserved.
Both companies concluded that the 3076 was too small for these pwer output levels (if I wanted any form of exhaust pipe anyway!), and that the 3582 would compromise driveability. The GTX3576 / GT3579HTA consistently came out as the best solution: it can flow huge power, but should retain the driveability I’m after to boot it round the twisty stuff.
So the decision was made, and it was a couple of days before a brand new M-spec turbo arrived from Owen Developments….happy days indeed!
Reflecting back on this decision process, I think I was largely correct in my thinking- and the flow data I’ve since seen for the 3S really backs up this up. The 3S head weakness is really in the exhaust port’s ability to flow the exhaust gases out of the cylinders…which helps explain why the top end turbo quoted power levels are often unobtainable as the turbine housing bottlenecks, and flow through the system is further restricted.
By matching a larger turbine wheel to a smaller sized compressor wheel, this inherent design weakness isn’t exacerbated, potentially allowing higher output levels to be achieved without necessarily sending boost threshold levels through the roof. I’m sure the same turbo may not work as well on other design engines, but it does seem to work well with the Toyota 3S head design.
Anyhow- apologies for the overly detailed post, I know it has been useful to a few folk over the years since I originally wrote it, so it seemed only right to re-publish it in its entirety! Next up we move onto some of the new supporting modifications to help the new turbo to do its job properly…
Comments