Anyone good friends with a cop?

[Dr. J]

Well I finally got a chance to to some experimentation with the speed issue.

I have 3 (easy) ways to measure the MINI speed:
- MINI OBDC (which is suspect)
- GPS (which isn't all that precise)
- RPM calculation (using my ScanGaugeII and some calculations).

On my way home today I hooked up the GPS, set the CC at a few setpoints (50 - 75 increments of 5) and recorded the values. The results are attached in chart form.

Basically, if you trust the GPS (the resolution on the GPS is 1 MpH, btw), the MINI OBDC overestimates the true speed by roughly 5% (running a trendline through 0, since when the car is stopped they MUST both be 0!). Back-calculating from the RPM figures from the ScanGaugeII (the RPM jumped around a bit, bu a variation of 20 RPM @ 3000 RPM is less than 1% variation), the MINI OBDC overestimates actual speeds by roughly 1.5%.

So the question now is - can the OBDC RPM value be trusted? After all, although I've convinced myself that the odometer is reading correctly, the displayed speed is definately off, and if the RPM more or less agrees with the speed, then it must be off as well, so what is there left to trust?

It would be nice if I had access to a track where I could do time trials and physically measure the speed of the vehicle, but that obviously is not happening.

[rlw]

So, would it be possible to obtain a more accurate speedometer-displayed speed by fitting tires that are slightly larger in diameter?

[AndyV]

I believe in the UK that the accuracy of the speedo has to be between +10% and 0%, so that it always over-reads, hence why manufactuers aim for 5%. It seems they are quite good at getting it at 5%.