Re: How You Measure [was Dyno down the road...] <longish>

[SMILLER@xxxxxxxxxxxxxx]

I would never question Jim Conforti's expertise in this area, and I really
didn't want to get into anything that might cause Mark K to start a new
flame war, but I must partly sorta kinda defend Mark on this one...

>Date: Tue, 10 Nov 1998 14:31:55 -0700
>From: Jim Conforti <lndshrk@domain.elided>
>Subject: Dyno down the road...
>
>Mark,
[snip]
>Recording the time it takes to get from Speed X to Y in Gear Z
>only tells you the INTEGRAL of the curve with respect to rpms.
>
>(Area under the curve for the Calculus-Impaired)
>
>You'd need to have an accurate accelerometer continually logging
>data to show the actual torque curve itself..
>
>And you still have the problems of varying wind loads, road grades
>and operator error..
>
>[and snip again]
>
>Jim Conforti

When making engine modifications, there are 2 questions that one might
ask:

1.  How much power did I gain witht the mod?
B.  How much faster does the car go because of the mod?

Jim's comments, above, address mainly item 1.  Having an accurate set
of before and after dyno graphs can tell you loads about how much
power you gained at all RPMs.  Mark's method answers only item B. 
Timed acceleration runs tell you how much faster the car goes as a
result of said mods.  Both are valuable pieces of information.  Plus,
freeway runs save you the cost of the dyno rental.

In both cases, care must be taken to account for, and correct wherever
possible, the environmental variables that can affect the results.  On a
dyno, each run should be corrected for temperature, barometric
pressure and maybe even humidity?  (Sorry, I've never actually done
dyno runs, but this seems logical.)  On local freeway runs, you would
need to use the same stretch of freeway for each test, and also account
for the same variables as the dyno, plus wind speed and direction, and
maybe precipitation (like you would be driving 100 mph in the rain, right). 
Oh, and don't forget to factor in the fuel, cargo and passenger load,
ideally it would be the same for all runs.

Then there is the question:  What is the intended use of the data?  If
you're selling a product, you had better properly account for all the
variables.  Customers want to see the dyno sheets.  If you just want to
know about your own car for your own satisfaction, you could probably
get away with just estimating the effects of the variables.  But, I would
be careful how I presented such data to my friends and the general
public.

All things being relative, there's one other aspect of this to consider:  If
you're just doing this for your own satisfaction, and your car is (oh, say
for example) 0.2 seconds faster 60 - 100, and you didn't properly correct
for all the variables, then you must consider that your results are within
the range where the environmental factors may have made the
difference in the performance runs.  On the other hand, if your mod
made the car 3.5 seconds faster 60 - 100, and you just kinda estimated
that the weather was the same for both runs, then you've probably got a
worthwhile modification, regardless of how it would look on the dyno -
assuming, of course, that you didn't mortgate your mansion to buy the
mods!

So to recap (hold on, almost done now!), both methods of measurement
have valid applications.  The audience (that's the other 4000 of you out
there reading this) needs to consider what was done and why it was
done.  Then you can decide if it is something you want to try for
yourself.

Phew, done now.

Scott Miller
Golden Gate Chapter
BMW CCA #44977

------------------------------