DaveK

I think my navigation system is supposed to be accurate to 10m - and considering when it tells me to turn, it's definitely far more accurate than 100m.

robbed666

GPS velocity accurracy is quoted typically by a lot of commercial GPS units as +/- 0.1MPH. BUT, this is an average, and remember to take several readings over the longest flat STRAIGHT (bends will upset the result) you can find. GPS accuracy improved a great deal when the US turned S.A (Selective availabiliy, an error introduced into the position data to reduce accuracy. Used to stop 'hostiles' from using system for targeting) off. DGPS made 'Selective Availabilty' obsolete any way. I have a Gtech Pro, works OK, uses accellerometers. It can calculate 0-62, standing 1/4, and if you input vehicle weight, BHP. I have 5 'speedos' in my 964 2 GPS, and 3 driving off the speed endcoder. I'm running 18' speedlines. One of the speedos is on based around a PIC MCU. Ths is typically within 0.5MPH when checked against both GPS units as detailed above. I can recalibrate for any wheel tyre combo, and adjust for tyre expansion as speed and/or temp increases. UK Construction and use regs require speedos to be +10% -0%. i.e they can never indicate a speed slower than you're actually travelling, but changing wheel tyre combos does throw this out, but as Christer said, not as much as you think. You need use rolling radius in your calcs not just measured 'R and PI'.