How We Did It Testing is all in the numbers, and you need lots of numbers to really unders
The Testing
Testing was performed on a 1997 Eclipse GS-T with a front-mount intercooler, big injectors, cams, a popular hybrid turbo known as a Frank Stage 3, and a 3-inch exhaust. Two key points to remember when comparing these results are that this car exhibits minor boost creep and it has an internal wastegate.
We intentionally tested a car with boost creep because it's a common problem. In a later test, we'll use these same controllers with an external wastegate.
A Dynojet was used to simulate the load of accelerating on the street in third gear. Our two basic tests tracked boost when the throttle was slammed open at 2000 rpm and at 4500 rpm. The 2000-rpm test would give the controllers a chance to highlight any improvement in response, and illustrate their stability through the rev range. Tipping in from 4500 rpm challenges the controllers ability to suppress boost spiking as the turbo is already up to speed.
Interpreting The Results
When interpreting the data, don't dwell on the final boost level. We were shooting for 12 and 17 psi, but didn't spend too much effort landing right on those values. More important is the rate at which boost builds, and the controller's ability to prevent spikes and hold a constant boost level. The tests at 2000 and 4500 rpm are shown on separate charts.
The transition from building boost to holding it is a demanding event for the boost controller. The longer boost can be kept from the diaphragm, the quicker the boost will build. Wait too long and it will overshoot, open it early and response is soft. Even if you get the timing right, there's the problem of how much to open. Again, not enough and the boost keeps going up, too much and your gauge will bounce like a basketball.
Finally, remember these tests were done on a car with an internal wastegate that does exhibit boost creep. Some controllers specifically warn against use on engines like this, and not surprisingly, they didn't work so well.
Controller: A'pexi AVC-R
A'pexi suggests having the AVC-R installed at one of its authorized Power Excel Shops. This might be a good idea, as splicing into the car's ECU harness to intercept engine speed/injector pulse width, throttle position, speed, power and ground are required to get the most out of the unit. Wiring in a momentary switch, which is not provided, is also required to take advantage of the scramble mode. If you're comfortable digging through the dash, and making good clean electrical connections, you should have no trouble installing the unit yourself. A'pexi includes ECU pin-out diagrams for most popular Japanese makes.
A single solenoid regulates the signal going to the wastegate, and a separate MAP sensor can be placed wherever it's convenient. The quality of the components is first rate, with waterproof connectors used for all engine compartment junctions. About the most annoying thing in the world is not having the right vacuum "T" fitting when trying to install your new toy at 1 a.m. A'pexi comes through with a vast goodie-bag of tee fittings and hose clamps. Also of note are the vehicle-specific install diagrams for the twin-turbo Supra and RX-7, which interestingly recommend rerouting the turbos to run in parallel.
The control unit itself is handsome, with a huge vacuum fluorescent display that's easily viewed from all angles and auto-dims in low light conditions. Contrast is adjustable from a submenu. An adjustable bracket simplifies mounting on uneven surfaces.