Todd's Acura TL

Download the G-Tech Pro data files (GTP files).
Download GTech PASS (GTP file viewer).

Initial Runs: 09/15/2004

Todd joined the G-Tech testing club on Wednesday, after giving his 1999 Acura TL a spin (no, not literally.) Given that his Acura is is a FWD car with good power and traction, I expected a pretty straight-forward test and simple analysis. However, instead the data seems to be rather inconsistent with itself and outside references.

I will go into more detail as we go about the inconsistencies. Since I cannot pick one or even multiple factors that I can try to isolate, the best I can offer is to completely average all three acceleration runs together and only deal with the aggregate data. I would like to do another set of runs with a more careful eye, and snag Rob's OBD2 data logger as a backup. I don't like mysteries!

Acceleration:

Todd performed three acceleration runs and one power run, but the acceleration results varied widely. The first and third runs were not very close at 16.7 and 17.0 seconds, but the second was even farther apart at 16.1 seconds. Having a total difference of nearly one second on the quarter mile is huge, and should never occur unless there was a significant problem on at least one run (like when Chad tried to vaporize his tires over the course of 50 feet...), or when there is some bad data in the mix. Since Todd's second run was about the pace I expected of the TL, I figured that when I looked into the data I could find out why the runs varied so much.

Unfortunately, I have a whole lotta nothing on that front. The first thing to examine is the launch, but the launch RPMs graph shows that each launch was within 100 RPM of the next. Even though the RPMs were the same, when looking at the launch Gs there is a fairly significant difference between the three launches. The acceleration edge continues evenly throughout the second run, even though the shift points are all identical and occur at equal intervals.

So, since I cannot eliminate any one run at present, averaging all three gives an unadjusted aggregate of 16.3 seconds @ 85mph.

Power:

The TL's rated output is 225bhp @ 5500rpm, and 216lb-ft of torque at 5000rpm. The three acceleration runs show a massive power disparity between the runs. Given that the second run is more than 20bhp more across all RPM ranges, normally I would throw out that run as garbage. When you compare them against the power run we performed, you can see that both acceleration runs and the power run agree pretty closely. The second acceleration run, however, is way off the map.

Before the second run the G-Tech fell off the windshield since I had apparently not fastened it well enough. At first my theory was that since I had to re-attach it, the self-leveling was off for that run and therefore gave this mirror, but higher, power and performance result. However, we did perform a quick calibration after re-attaching the G-Tech, unless for some reason it didn't take properly.

The only problem with that theory is that the average power output across the three "good" runs is 129bhp @ 5600rpm. After adding in a 15bhp aero drag and 20-25% driveline loss for the auto tranny, the adjusted power output is 173-180bhp -- way far off the rated 225 bhp. Since Todd's car has only 60,000 miles and was running smoothly and responsively, I can't imagine it was down that much power. Of course, power is calculated from the acceleration measured and the weight entered, but our test weight on the scale was 3990 pounds, and matches the curb weight of 3461 pounds plus two people and fuel.

So, until we can re-test with the G-Tech and Rob's OBD2 data logger, I can't trust the measured power output. My best estimates based on the averaged acceleration runs would be 200bhp @ 5500rpm.

Technique:

Unfortunately, with the inconsistent data I can't lay out what technique changes to make to bring better results. Since the TL has a high torque peak (5000rpm) and power peak (5500rpm) in general I would recommend brake-torquing to as high of an RPM as possible without spinning the tires wildly. In practice Todd got it to a bit above 2000rpm, but it may be possible to raise the RPMs further, which should significantly drop the quarter mile times.

Other Info:

Obviously, I'm very disappointed with the lack of clear data. While I'm confident that we could get to the bottom of it with another G-Tech run augmented by OBD2 logging, I can't offer a lot with these runs. I apologize; I feel like I let something slip but I just can't see where the problem(s) lie.