As we inch ever closer to getting Project Miatabusa on a dyno, more and more unfinished business keeps revealing itself. Ok, the car starts and runs now, but it still has lots of disconnected hoses, gauges not working, and untested changes waiting to blow things up.
Neel Vasaveda of Apex Speed Technology is on board to tune the engine. I keep mentioning Neel as some guy who knows more about Hayabusa engines than we do, but his tuning expertise goes much deeper than that. Neel does engine management, wiring, and race support for a staggering variety of serious race teams, like the Jordan Suzuki AMA superbike team, The Porsches of Magnus Racing and The Racers Group (between them, 24 Hours of Daytona winners in 2009, 20011 and 2012), and Leslie Porterfield's Bonneville bikes. But this pro-racing stuff is a total snore compared to the oddball special projects he's always entangling himself with. Like replacing the pathetic '80s Italian dual ECUs on a Lamborghini LM002 with a Pectel SQ6, or making sure there's enough electronic wizardry in this V10 E30 M3 to match its mechanical awesometude, or cleaning up the mess that is a certain Hayabusa-powered Miata. Yeah, we're in good company here...
Experienced as he is, we're not gonna ask Neel to tune this thing by ear. Good tuning requires data. Most dynos have widebands you can stick up the tailpipe but that data is always suspect. You have no idea of the quality or history of the sensor, and the exhaust in the tailpipe took some non-negligible amount of time to get to the tailpipe. If the engine is revving quickly, you could be reading data that is a few hundred rpm out of date.
More importantly, the Hayabusa has different air/fuel trims for each cylinder, partially in an attempt to adjust for the inconsistent airflow caused by the unusual pairing of its 4-2-1 header. Since we replaced that header with a 4-into-1 header custom made by SPD, there's a really good chance fuelling is going to be wrong on at least two of those cylinders. The only way to know is to simultaneously sample the air/fuel ratio directly at each exhaust port.
To make this happen, we grabbed an AEM 4-Channel Wideband UEGO Controller. Wideband O2 sensors aren't just a simple self-contained sensor that spits out a voltage corresponding to air/fuel ratio. The sensor works in a feedback loop with some complex controller electronics. If you want the full scoop on how this works, dig way back in the MotoIQ archives to Mike Kojima's story on how widebands work, which was one of the first stories published on this site. Otherwise, just understand that the electronics are complex and expensive, and having one box that simultaneously controls four sensors is quite handy.
One cool feature of the AEM 4-Channel Controller we're not using yet is the backpressure compensation. If you read Kojima's story, you'll understand that accurate sensor readings rely on the sensor being near ambient pressure. if you have a lot of backpressure, the readings will drift. This is why you should never put a wideband before a turbo. But if you want individual cylinder data on a turbo car (and isn't that why you have a 4-channel?), you can install an exhaust pressure sensor and the UEGO controller will automatically compensate the readings. The box comes pre-wired for this feature, but doesn't come with the pressure sensor, since not everyone has a turbo (yet).
Even though the AEM 4-Channel UEGO controller can control up to 4 sensors, the box only comes with one. This way you're not paying extra for unnecessary sensors if you're tuning a Geo Metro. Keep this in mind when ordering!
The sensor comes with AEM's Stainless Tall Manifold Bung, shown above. The bung is extra-tall to hold the sensor farther out of the exhaust stream. The sensing element is still in the stinky stuff, but the recessed position is less disruptive to exhaust flow. The bung is also finned to help keep the sensor from overheating with high EGTs. Both these features are hugely important when installing sensors on the small, hot, flow-sensitive header primaries!
Before drilling big holes in the sexy SPD header, I wanted to test fit the bungs to make sure they were in positions that didn't block bolt access to the header flanges or air injection fittings, and left wrench room for installing the sensors themselves. I did this mockup with cyanocrylate glue (a.k.a. Super Glue). Just go to any hobby shop that deals with model airplanes and ask for CA glue. The Zip Kicker activator makes the glue cure instantly, so there's no need to sit and wait. Just put two drops of glue on the bung, put it where you want it, and spray the activator on it. There will be a little crackling sound, a toxic plume of stink, and sometimes even a little plume of smoke, and BAM, the sensor is locked in place. The glue is still weak enough to break off by hand, but the bung is held on well enough to confirm wrench clearances.
These sensor locations were deemed wrench safe, so I took the header, with bungs glued in place, to Mario Lozano at TSR Fabrication in Gardena, CA. Having the bungs glued in place annoyed Mario slightly, since he had to clean the glue residue, but there was no question where the bungs had to be.
(By the way, the air injection fittings shown here will cause some minor complication. If the air injection system is operating during tuning, the extra oxygen will throw off the air/fuel readings, so we'll have to be careful to plug the air injection's inlet every time we test.)