About 6 months ago I was approached to see if we can do anything with the BRZ. At first, I said no, but I'd be glad to put it on and see what the other guys are doing (moto has been MX-5 only up until 2012). So we put the car on the dyno to see what the purchased tunes did...and sure enough this is what we found. BTW, this is NOT any tuner starting with a V [thanks for the heads up though John--I owe you!!!]
So here we some good gains 4-6k, and up top with the higher rev-limit. We also worked with another tuner to send back and forth datalogs via email (as our business model) and put on nearly a dozen runs--and yet no better (actually worse than stock). So...at this point having done the MX-5 for a few years this looked easy enough. Giant torque dip, 2.0l high compression motor...wonky cam phasing and AFRs on the stock map...Sure!
However, not so easy.
After a few weeks of testing and road tuning we got to this (on a different dyno)
So we're doing better, but still up top it is a bit lacking. The nice thing is it gave plenty of time to play with the RaceROM features, custom maps, and a whole bunch of other goodies in the software. The EcuTek guys invested a ton of time and money into this. Nothing like this has ever been available for tuners using the stock ECU.
Regardless, further testing ensued. It would be as though for every two steps forwards, we went a step back. Unfortunately you won't find Subaru's in-house methods for peak cylinder pressures or their calculations for intake tract length vs. piston cup size in any textbook; but fortunately patents are actually quite useful in finding out a particular strategy. And the crux of it is that unlike the any straight DI, or straight port injection, variable intake or exhaust cam car--all of these MUST absolutely be perfect in timing to get any HP gains. Before you would tune the ignition timing, lean out the AFR, remove torque limits and there you go, 10-20 hp. Well, Subaru/Toyota, thank you very much, the job has become harder. The stock car will learn up to a
very high ceiling. The stock maps are already fairly optimized. To make HP on this platform is no joke, I credit Visconti and Perrin for having real gains anywhere early on. Regardless, this is the last dyno from last weekend. We finally got the torque dip more of a "dip" rather than "pause". With custom mapping we also got the top end opened up a heck of a lot more. (note: this is the lower reading dyno again)
With all that said, dynos only tell half the story. Acceleration is about 1 second quicker 0-60, and with the latest RaceROM features you can record "longitudinal g's". This is REALLY cool, because you can actually see what the tunes are doing in real world conditions and with learning. You'll find high variability between runs based on temps on the dyno. If you have a straight road, good traction, low winds; repeatability on the street can actually be
better if you're taking care since you can cool off, take advantage of the CAI, not doing ECU resets etc.... The flat foot shifting, launch control, autoblip are really cool. Another neat feature is to "calm" the rev limiter so auto-x guys aren't losing 5mph in 2nd.
So this is how you can look (anyone with ProECU can do this with RaceROM v4) and see what you're getting g wise. Look like a torque curve maybe?
The 2d view-this can be used to analyze the differences the maps are having. You can also use software like virtual dyno to get HP ratings. This is more precise for instantaneous torque, whereas the VD types are a better indicator of area under the curve since they use timing. Neither are a substitute for a dyno, but are extremely useful for tuning if done proper.
And finally, tuned we have this logged G curve on the latest tune:
Much smoother. So that's it for now, we will resume finagling with it as it warms and maybe do some E85 testing. For now I'm fairly happy to say you can definitely make HP with this car on pump gas--though any broad gains across the board I'm a bit skeptical. This I found to be very easy to do with the dyno via gearing or learning. In the dip and up top there is clearly some room.