Jon Petty wrote:
> I have a project where I want to provide some additional fuel to a vehicle
> for  improved throttle enrichment. I need some help with a TPS (throttle
> position sensor) rate of change (ROC) function.
[snip]
> Sampling frequency is  a significant issue.  I guess I probably want to
> average a few readings then wait a period of time, then resample. The time
> between measurements will greatly affect the rate of change value. If it is
> quick, like 10 msec, it will be very difficult to measure because the TPS is
> a human activated event. If sample time is long like 1 second, the event is
> over before the system can react. I believe that 100 msec between averaged
> samples will work.

Jon, even that human muscle reaction can reach 100ms in a repetition
cycle, your TPS would need more than 10 samples per second for a deep
analysis, since you will have several variable points in between each
100ms.

I understand that you want not only ROC (rate of change) parameters, but
also a curve of the full-throttle percentage on time, so you can
generate further computer calculations.

I would guess that a min of 50Hz (20ms) would be a good sample
frequency, while a 20Hz (50ms) would be an acceptable but not rich in
details.

The simple "Sample & Store" routine is enough for a later transfer to a
PC for further analysis.
Actualy, there are several PC shareware programs that can provide deep
analysis on your collected data, try
http://www.zdnet.com/pcmag/downloads/index.html for a search.

Wagner.