« January 2006 | Main | October 2006 »
February 17, 2006
ECU Tuning - Too much information?
So I get early to the racetrack in my supercharged 2004 Acura RSX and end up the first car in Lane 6, waiting for the time trials to begin. Usually they do it by lane, but this time they did it across lanes. It was a cold day, and so my car wasn't really warmed up when I was suddenly told to stage for my first run.
I got off to a good start, but the car felt sluggish. There just wasn't any power even though the boost gauge blinked at me, signaling I had exceeded the 10 psi boost limit it is set for. The time slip said it all: a 15.2 second quarter mile at 93 miles per hour. That's a second and a half slower than usual.
Fortunatly, I had datalogged the run on my laptop, connected via USB to my Hondata K-Pro ECU. The log showed that I had run very high boost, up to almost 11 psi as opposed to my normal 8 psi or so. I had also shifted at 7,500rpm instead of the usual 8,600. That's because the engine just hadn't seemed to want to rev any higher. Air-fuel ratio seemed fine, though on the rich side at about 11.5. I had almost 40 knocks, which seemed very unusual as my car is tuned conservatively. So I looked for anything else that might have been wrong, and finally I found it: the Honda K20a engine had never gone into VTEC. I had run the entire race on the low cam.
What should happen is that under full throttle, the ECU issues a signal that then switches the VTEC solenoid, oil activates a switch, and the cam mechnism changes to the high speed cam, commonly known as the "VTEC" point. Mine didn't. There was no VTEC signal at all. So I pulled out of the lane and parked the car as I didn't want to risk damage. I then datalogged and found that gunning the engine still did not produce a VTEC signal. I checked the oil, which I had changed that same morning, to see if perhaps I was low. I wasn't. I also didn't see any physical damage that might have been done by the service station. I then checked a datalog from earlier in the day, and there was a VTEC signal.
I then reloaded the calibration into the ECU, thinking something may have been off with it. No dice. Since I use an older version of the Hondata K-Manager software, I decided to load the latest version and then upload the calibration again. Nada. I called Comptech in the hope that someone might still be there and may have a clue. No one answered. I asked a guy who also has a Comptech supercharger and who works at an Acura dealership. Everything looked fine to him.
Then I notice that the latest version of K-Manager lets you set the speed below which VTEC will not kick in. It is 6mph. So gunning the engine while parked would not have triggered the signal. I changed the setting to zero mph and upload again. Bingo! We have a VTEC signal. So I decide to race again (I did poorly).
I still don't know why there had been no signal and begin to suspect that it may have to do with engine temperature. I send an email to Derek at Hondata and he answers right away. Yes, VTEC needs an engine coolant temperature of 140F whereas my car had been at 136 degrees. Case solved. And no, he doesn't think revving that high on the low cam did any damage, an assertion later repeated by Comptech technicians.
Sooo... having the latest and greatest in ECU tuning and datalogging tools had allowed me to figure out what ailed my car. It didn't give me an immediate answer as the 140 degree cutoff wasn't documented anywhere. Without datalogging I wouldn't have known what was wrong and likely just would have tried again. The car would have been hot and VTEC would have been normal. Does that mean our new ECU tuning tools make us too reluctant, too scared of every variable that seems a bit off?
Posted by conradb212 at 2:07 PM