Mechanical efficiency vs. calibration
A dyno is a tool, not a trophy. It measures wheel torque under a particular load with particular assumptions on that day. Change the assumptions, get a different number. Useful for tuning deltas; unreliable as a universal scoreboard.
What a dyno actually reports
What accelerates the car is tractive effort at the tire integrated through gearing, shift control, and traction—not a single peak on a graph. Two cars with the same peak rwhp can perform very differently if one:
How an engineer uses a dyno (and wins on pavement)
The dyno doesn’t make you quicker. Engineering does.
A dyno is a tool, not a trophy. It measures wheel torque under a particular load with particular assumptions on that day. Change the assumptions, get a different number. Useful for tuning deltas; unreliable as a universal scoreboard.
What a dyno actually reports
- Wheel torque → horsepower is computed:
HP = Torque (lbft) * RPM / 5252. That’s math, not magic.
- Device + settings matter: inertia vs. eddy-current dynos, ramp rate, load control, smoothing level, tire pressure, strap force, fan placement/CFM, and gear choice (1:1 is best: TR-6060 4th, 8HP90 6th) all move the result.
- Weather/standard changes the printout: SAE J1349 vs. STD/J607 corrections can differ several percent under the same pull. Same car, same day, different correction = different “power.”
- Drivetrain loss isn’t a fixed %: it varies with gear, temperature, fluid shear, tire deformation, converter/lockup, and bearing/friction state. Back-calculating “crank hp” from a wheel number is guesswork.
What accelerates the car is tractive effort at the tire integrated through gearing, shift control, and traction—not a single peak on a graph. Two cars with the same peak rwhp can perform very differently if one:
- holds converter/TCM lock and shift timing cleanly,
- keeps IAT2 under control so it doesn’t pull timing,
- maintains lambda and fuel pressure at WOT,
- and actually puts power to pavement without spin or wheel hop.
How an engineer uses a dyno (and wins on pavement)
- Baseline on one dyno, same settings. Change one variable at a time. Log everything: boost/MAP, IAT2, spark advance, lambda, fuel pressure, KR, trans slip/lock status. Compare deltas, not egos.
- Validate off the rollers. Use data that represents work done over distance/time: trap speed, 60–130 GPS, repeat hot-lap pulls. If it doesn’t repeat, it isn’t real.
- Tune for control, not theatrics. Don’t “make a number” by zeroing torque management or pushing timing into knock. The fast car is the one that repeats with margin when heat-soaked.
Bottom line
The dyno doesn’t make you quicker. Engineering does.
