In Part One of our interview with Peter Luxon and David Inall of APS we discussed drivelines, exhaust and intake development - now we'll move onto intercooling and engine management...
"We do all of our intercooler development on the racetrack," says Peter. "We might do a little bit on country roads, but the majority is on the racetrack where we data-log the cars."
"It has become much too hard to test on the street," confirms David Inall. "We used to do a lot more on-road development, but with tickets... Our intercooler heat soak testing and everything else we now do on the racetrack - we just hire one for the day. It's much safer to replicate certain driving conditions on a racetrack."
And is a single type of intercooler core used in all of your kits?
"No. We specify the internal and external finning of each of our intercoolers. It is not just a case of here's the sort of core we use; we spend a sh load of hours on development," says Peter.
"With the XR6 Turbo we've had nine different cores built to our specification for testing. We have very clear ideas of what pressure drop we're prepared to accept and the charge-air temperature we need. On top of that, there are many other considerations - I mean, you've still got to have optimised timing and air-fuel mixtures at any given point.
"In no circumstances will we accept a pressure drop through an intercooler of 2 psi or more at the rated power. The intercoolers in our kits have a pressure drop from about 0.25 psi but, in a worse case scenario in an engine running 750 horsepower, it might be up about 1.8 psi.
"We always try to cool charge-air temperature as close to ambient temperature as we can achieve. Typically, at their rated power level, our kits keep charge-air temps to within about 8 or 10 degrees of ambient."
David adds, "From a power perspective, there is actually a charge-air temperature that you don't want to go below and that varies a bit from car to car. So there is such a thing as over-intercooling - but that doesn't apply to air-to-air intercoolers, because you've got an ambient condition."
And what are APS's views on water sprays and water injection?
"The problem with a water spray," says Peter, "is you can only hold a certain size reservoir in your car. Even although we map our water spray systems (where fitted) to operate only in certain conditions, we find the biggest problem is reservoir volume.
"In terms of efficiency, we've seen full-load charge-air temperature drops of around 8 or 9 degrees but, sorry, I can't remember the base charge temperature. It wasn't as dramatic as we thought it might have been, though. That's one reason why we try to build the most efficient intercooler foremost - we don't have to rely on running a spray.
"Back in the early days of our WRX kits we included a water injection system. One of the problems, though, was you'd get contaminated water in the reservoir and that would block up the jet. People found they could also run it out of water pretty easily.
"I really think that unless you have a sophisticated system where, say, you pull out 10-degrees of ignition timing and half the boost when the water tank runs empty, it's a very dangerous approach. Twenty-five years of experience has taught me that it generally ends up in tears - a good intercooler is far more reliable and consistent."
David continues: "Many years ago I remember Saab offered water injection as a dealer option, which I think - in this country - added about $2500 retail. It was very expensive because it did all the things that Peter just mentioned; it had a little auxiliary computer that took note from various inputs and played with the boost when the system ran out of water. It had all of the safeguards that you would expect from a car manufacturer. But it was a $2500 water injection system - and that was ten years ago.
"A more basic form of water injection might be fine for a person mucking around at home, but when you're in production engineering you follow a fundamentally different philosophy," he adds.
Why is it that the current GDB-series Subaru WRX so much more difficult to get power out of?
David says: "That is loosely a function of its ECU self-learning - but, really, I would call it a function of 15 years of technology. The design of the '00 Rex computer goes back to the mid 1980s - it's fundamentally the same sort of thing as an early Liberty. For the '01, though, the WRX was to become a world market car and a whole lot more R&D went into it. There is no doubt the '01 computer is not just one year more advanced than the previous model's - it's about a decade and a half in engineering terms."
Peter adds: "And as far as I know, there was no closed-loop boost control in up to '00 WRXs. You put a new exhaust on a later model WRX and backpressure goes down, but boost doesn't go up and - surprise, surprise - it doesn't make quite as much power as the pre-'00s."
And now onto ECU mods - APS is the Australian Unichip importer.
"Chipping cars is a very controversial area," says Peter. "You'll find that some standard cars make an unbelievable gain when you fit a UniChip, but other cars - where the engine management is so close to perfect to begin with - you'll only get a very small gain. Most late-model Porsches are that close to the money - they hold stoichiometric to over about 95 percent throttle opening."
"We can, however, make differences with timing," adds David, "if the customer is prepared to use a particular grade of fuel all the time. On the other hand, you can chip a 3.8-litre V6 Holden but it will only accept 36 degrees of timing - it won't accept 40. So if you've already got 36 degrees of timing and an optimal air-fuel ratio, it doesn't matter what you do to the management system - you won't pick up a kilowatt."
"The optimal air-fuel ratios and timing figures depend on what sort of engine it is - what one engine will accept, another will not," says David.
Peter continues, "I see mixtures in some engines that would absolutely disintegrate a WRX in three minutes. Certainly, the EJ20 is a very unique engine. If one tried to apply that same tune to another engine you could come undone pretty damn quickly.
"Of course, fuel quality and the operating conditions of the car are also very relevant. If we were in Dubai, we'd tune WRXs very differently to how we tune them here in Australia; people over there will sit on 240 km/h for half an hour. If you run an Australian tune over there, the pistons and conrods would probably end up all over the desert..."
"And even within Australia there are some significant variations - in Perth, for example, they don't have Shell Optimax," adds David. "That's why we supply the UniChip in our kits with our own maps and the approved dealer can fine tune where required."
"I don't believe there are any great scientific secrets in tuning," says Peter, "but the big thing is being able to do it live on an accurate dyno. When you can do it live you don't have to wear the car out going back and forth trying to tune it - you can hold it at a load and rpm site to make changes.
"I think if you've got real-time programming, a good technician and an accurate dynamometer you're going to get a good result - it doesn't really matter what brand [of management] it is, so long as it has a good number of sites.
"For any particular type of car there may be other real-time programming approaches - such as chip burning - but we're talking about a product [UniChip] that can be used on hundreds of cars. That gives us - and the customer - real flexibility."
"The UniChip can also be readily re-tuned by our distributors and that isn't the case with many other approaches to ECU modification," says David.
"Further, most retail workshops won't spend the amount of time necessary if in fact the tuning process is overly time consuming. They have constraints like workload and what the customer is prepared to pay - they are real concerns to a workshop operator. If it involves, for example, downloading maps and a number of phone calls you'll find some workshops simply won't do it."
So the UniChip can be programmed in real-time and is suitable for a variety of cars - what other advantages does it have over other approaches to engine management modification?
"As far as we know, the UniChip is the fastest product in terms of tuning accurately and quickly; no other system will let you adjust ignition and fuelling simultaneously on the dyno at the same site," says Peter.
"You can also make timing changes in just 0.1-degree increments and, of course, you retain all of the niceties of the factory system.
"I've never driven an aftermarket system that drives as well as the factory computer - they don't have anywhere near the processing speed of the factory unit. The big question is - why throw away all the car manufacturer's millions of dollars of design and development if you can go in and change only what you need to?"
We've now highlighted the UniChip's advantages - what are its disadvantages?
"Cost," says David Inall.
"There is no hard material cost in a reverse-engineering system, but obviously the UniChip is an electronic module in its own right. On the other hand, the UniChip can - right now - go onto the US-spec STi, whereas the generational change in its software has put everyone else back virtually to square one."
One thing that the EcuTek factory software reprogramming system, for example, can adjust is the amount of knock retard. That cannot be done with the current UniChip, so isn't that another disadvantage?
"That's an interesting idea," replies David, "but I've never seen it as an issue on WRXs. We don't set our cars up purely for maximum power. If somebody did run at optimum power AFRs and ignition timing in a WRX, I suppose I can understand why you would want a greater degree of knock retard available in a big hurry."
"I think if you've got the correct air-fuel ratios, timing and boost in the first place the knock sensor should not be so active," adds Peter. "And, look, if you want proper knock detection you'd take what's on a WRX and throw it in the bin in any event. If you want cylinder-specific knock protection go and look at a Porsche or BMW engine."
"One thing I'd add is that the UniChip doesn't currently have closed-loop boost control - but it will have very soon," says David.
And how far into the future will the UniChip be able to cope with new manufacturer developments - electronic throttle control, independent cam timing etc?
"Good question," replies Peter.
"There have been seven electronic engineers on it for the last few years and, yes, it will control electronic throttle, independent camshaft control, closed-loop boost and all of that stuff. We can't comment too much but the future of UniChip looks very good."
Air Power Systems
+61 3 9720 9170