[chbot] Data filter

[Marshland Engineering]

I think I have found the solution which is to average over a bigger sample
range. I did not want to do this originally as I would loose data either side
of my averaging. If I averaged over say 30 samples, I would loose 15 samples
at the beginning and the end of the data. If I only had 120 samples to start,
30 is quite a lot. 

The solution was to start collecting data both before t-start and after t-end.
I originally had the software start and stop recording against RPM values. 
Now I start recording before I get to my minimum value and recored a further
second or 2 after the max. This should work. 

Cheers Wallace.