1997 Mazda MX-5 Miata: Turbo Conversion With Garrett And TiAL Sport
April 18, 2013
When it comes to sorting out ancillary hardware like a turbo, wastegate and associated plumbing, an engine on the stand is worth, like, nine in the bush. That's how that saying goes, right? No? Well, it should.
One of the handful of reasons that Project Miata's new, stronger Keegan Engineering-built heart still awaits installation can be seen plain as day above. There are other, far less impressive excuses too, but I'll spare you the tedium of chronicling them here. Besides, there's a happy spinny thing to discuss.
Getting to Phase Two in Project Miata's power plant has been an incremental process. Clearly, we're approaching critical mass. Recall our goal for Phase Two of maximizing the powerband on 91-octane pump gas, an objective that requires flow, and this means turbo. No single piece of hardware influences engine performance more than does the choice of turbo. And corking up Keegan's long-rod 1.9-liter VVT BP with the wrong turbo would be counterproductive and just plain sad.
We consulted the brains at Garrett By Honeywell, ran some numbers, and decided on a turbo in their new GTX series, the GTX2863R. The GTX series replaces the compressor wheels of their existing ball bearing GT turbos with completely new, higher-flowing wheel designs.
Look closely above and you can also see a ported shroud (those four little struts in the housing right near the wheel), which widens its map.
The GTX wheels are machined from billets of forged aluminum. Regardless of manufacturer, billet compressor wheels are employed in aftermarket performance turbos for manufacturing reasons. Billet wheels are easier and more cost effective to produce in low volume than castings. In this case, the 63mm GTX compressor wheel is paired to the proven Garrett GT28 turbine in a TiAL Sport 0.86 a/r stainless steel turbine housing (more on this in a later installment).
Turbo matching, the analytical process of selecting a turbo for an engine, is something of an inexact exact science. It's a process that relies on math, experience and assumptions. No assumption influences a match greater than that of volumetric efficiency (VE). However, VE depends on backpressure, and you don't know how sensitive an engine's VE will be to backpressure until you run it or unless you have access to sophisticated engine modeling software like GT Power. Tricky. Then there's boost. How much boost you can run depends on detonation sensitivity, which also depends on backpressure, which is influenced by the turbo match. Round and round it goes.
This is when experience sometimes known as making a blind guess comes into play. We forged ahead under the assumption that the Keegan Engineering longblock will be a deep breather, what with its head work, +1mm Supertech valves and set of prototype cams by Trackspeed Engineering. The rev limit will be set to about 7,600 rpm, several hundred higher than stock. Keegan also built this engine with a tight squish distance, which will reduce detonation sensitivity.
Furthermore, we have a very capable ECU in the wings, the Vi-PEC V88, which has advanced knock correction and individual cylinder trims.
The above factors collectively biased the flow demands toward the larger 63mm compressor wheel over the runner-up turbo, the GTX2860R. In our case the engine's calculated lug line perfectly bisects the GTX2863R's huge 78% (!) peak efficiency island (use your mind's eye). This means the compressor will be operating at the highest efficiency possible
High compressor efficiency does two favorable things. The first is that it lowers the discharge temperature, so there's less heat load for the intercooler to deal with. But more significantly, high compressor efficiency means there's less work required by the turbine to drive the compressor, which in turn results in reduced exhaust manifold pressure, which improves VE. Same effect applies for turbine efficiency.
At the extreme of its flow capability the GTX2863R's compressor can deliver some 42 pounds per minute of airflow, which the Internet will tell you means that "bro, this is a 420-hp turbo." In reality you wouldn't want to run any compressor at the ragged outer edge of its map because the efficiency is at its poorest (~60%...which is still better than many Roots blowers at their best!) way out there, and you also risk overspeeding the turbo. So don't.
Also, it's ball bearing, which improves transient response and is more tolerant of extreme loading. Note that the ball bearing cartridge in Garrett turbos doesn't require as much oil flow as do conventional journal bearings. Typically, a restrictor is needed at oil inlet port of Garrett ball bearing turbos. Too much oil can overwhelm the seals located inboard of the wheels, resulting in a smoke screen, and can introduce enough windage in the ball bearing to negate its inherently superior spinny-ness.
TiAL Sport makes a nice steel -4 AN fitting with integrated restrictor for this. Measure your oil pressure at the turbo and select your oil inlet restrictor accordingly (the post-restrictor pressure should be minimum 10 psi at idle and minimum 30 psi at peak torque).
Next up: TiAL Sport turbine housing and wastegate.
Jason Kavanagh, Engineering Editor