drive train losses

Huw Powell audi at
Thu Aug 26 21:40:28 EDT 2004

> My point is this - the dyno measures what comes out at the drive wheels 
> - and then adding an arbitrary 20 or 25 or 30% "for losses" is simply 
> incorrect, and does not do justice to the effort, expense and skill of 
> the person who modified the engine.

What would be more useful, besides the good old seat of the pants test, 
would simply be fairly standard, duplicable tests, especially if run at 
various stages of the project - especially before starting (since that 
gives a baseline of what was getting to the wheels "stock," rather then 
the factory figure).

It strikes me that there are three, of varying usefulness:

1. Dynamometer.  advantage: consistent measurement, factors out air 
resistance and vehicle weight by measuring car standing still. 
disadvantage: expensive to run, hard to find awd setups, and heat soak 
factors will be different when sitting still (or do they run fans on the 

2. G-tech type measurement.  advantage: all the tuner-types probably 
have ready access to one to play with, so lots of measurements can be 
taken.  Hopefully, if you get the vehicle weight and air resistance 
figures right (and gear ratios, and tire size... right?), the numbers 
will be consistent and repeatable.  disadvantage: with poor weight and 
Cd numbers, the calculations won't give accurate results.

3. Drag strip time slips.  advantage: it's the raw unvarnished truth - 
if you get a 13.0, there is no debate about it, the car did it. 
disadvantage: driver technique and clutch/tire slippage issues make it 
harder to compare apples to apples.  Although, they do give you your 
reaction time, don't they?

Anyway, I guess my point is that playing the numbers backwards using 
"rule of thumb" percentages is just cheesy and who cares?  If you dyno 
300 hp, or turn in that 13.0, you know you have something good.  If you 
improve the numbers, they're better, whatever might be happening at the 

By the way, I suspect that the actual "drive line loss" formula, which 
of course will vary from car to car, is a complex second or third degree 
(or worse) formulation - ie, a constant, plus a factor linear to power, 
and perhaps ones that are proportional to power squared (or cubed, or a 
power below one, like a square root...).  And it probably varies by rpm, 
too, so whether you get max hp at 5500 or 6500 will change things.

Huw Powell

More information about the quattro mailing list