More: HP / Torque calcs

[Roger Heflin]

On Mon, 16 Aug 1999, Dave Zug wrote:

> Well I have thie little graphing thingy putting lines on the screen, 
> easy enough.  It reads the scan software's datafile directly 
> too. 
> 
> 2 things to note: even at 8192 baud the number of samples per second 
> makes the graph pretty chunky. Also my RPM data seems to be wierd.. I 
> mean the readings can go like this (this is WOT in 2nd grar):
> 
> this is not actual data, but a representation
> 
> 2600
> 2750
> 2700
> 2775
> 2875
> 2925
> 3150
> 3150
> 3325
> 3425
> 3600
> 3575
> 3700
> 3825
> 
> 
> see it's like there was hard knocking and severe spark reduction but 
> there actually was none at all! my delta rpm graph is a 
> sawtooth! Might it be stale old values being pumped out by the ALDL 
> when its too busy to update them or something??? am I correct in 
> asuming the time intervals are EXACTLY equal between the readings? is 
> the time differenntial a steady value for 8192 baud? my stream has 56 
> or something bytes in it.. I can calculate the time ass-uming the 
> packets are heel to toe with no delay. has anyone calculated this?  
> blah blahblah.. {diarhea of the fingers deleted} ;-}
> 

I have ran the calcs.   And I have noticed much the same you have.

I believe to get even reasonable numbers you will need to add some
sort of interpolation, and smoothing algorithm to get reasonable
numbers.

Mine were getting pretty good hp numbers using the data for several
samples (the number of samples varied depending on the speed or rpm -I
don't remember which).  I assumed then that the hp was the hp for the
center position, but this method lacked some coverage and resolution.

The only real way I could see to do it, would be to increase the
sampling speed by either modifing the data stream from the chip.   And
beyond the rpm being rather flakey, only having the speed in units of
1 mph is rather limiting on the accuracy area too.

I did assume that the samples from the data file where equally spaced,
I don't know if they really are, my guess would be they are probably
fairly evenly spaced, and it really does not matter given that the mph
number is not accurate enough to computer hp reasonably.

			Roger