[Jeff Frigo]

Found on the net, looks pretty good.

REAR WHEEL HP:
Now let's discuss "Wheel" HP. Additional losses occur between the flywheel and the tire, primarily from losses in the transmission and differential. The difference between Net and Wheel power should only be around 3% (stick shift) to 8% (automatic). But because most dynos utilize rapid acceleration to determine power, the inertial requirements push this up to universally accepted numbers of 15% to 20%. Here are the details:

TRANSMISSION AND DRIVESHAFT:
We typically find a 1.5% power loss in the 'gear section' of the trans and driveshaft (i.e., a stick shift or the non-converter section of an automatic). A torque converter's losses depend on the slip ratio, but up near the power peak well over stall speed for this argument, the slip ratios usually result in a power loss of around 4% to 5% for a stock converter. If you have a lock-up converter that is locked during the test, that's a 0% loss.

DIFFERENTIAL:
Also around 1.5% power loss.

INERTIAL REQUIREMENTS:
OK, so far we've accounted for a 3% (stick) to 8% (auto) power loss from "Net" to rear wheel. The other loss (that is incorrectly labeled 'inefficiency') is the inertial power requirements of accelerating all the rotating hardware in the system during the test; crank, flywheel, converter, even the two drive wheels.. etc. If the test could be done at a zero rpm per second accel (like on a hill, or like the factories test for Net!) then there would be 0% loss here. But the whole concept of an inertia dyno is to accelerate a known inertia (the drum) and calculate what power is required to do so. The guys at DynoJet simply do not add on the extra inertia of your wheels, axles, driveshaft, trans parts or crank-related hardware because they vary on every vehicle, and they don't want to mess with it... so they get a lower number. I've done lots of work in this area and turn all of these extra inertias into "pounds" and just add it to the vehicle weight with very good results. When testing on a DynoJet, of course, the gear ratios and all come into play, but as long as the test isn't in low gear, this 'extra mass' is commonly in the 7% to 15% range. This gives us the "15 to 20%" rule mentioned by others.

BOTTOM LINE:
Unless you want to do the specific inertial math, I tend to use 17.5% for automatic cars with stock converters and 13.5% for stick shift cars or automatics that were tested with the converter locked up.

Bill Watson