TECHNICAL EXCELLENCE

Speedometer/Odometer thread - is 7% reasonable? The answer...
Posted by Dennis Marer on Friday, 7 November 1997, at 4:26 p.m.

In fear of this becoming "Porsche Pete's Odometer Class Action Suit" board (and with hopes of getting back to some more interesting topics) let me post my back-of-the-envelope calculations on what I expect from my speedometer and odometer. My conclusions: a minimum variation of 7% is quite reasonable.

This is as scientific as I can get without doing actual experiments. Check my work - this was done over my lunch hour. If you find fault in my arguments, I'm sure you'll post corrections - thanks!

Let's say my wheels are 17", plus a couple inches of rubber on each side - approximately 3" (6" total, without actually walking to the parking lot). Let's also *assume* the wheel+tire combo is perfectly round and does not deform, and starts with 1/2" of tread (1" total). The diameter of the wheel would then be 24", so the entire circumference of the wheel is: C1 = (3.14159 * 24") = 75.39816".

Now 6 months later you've worn those babies down to the cords, leaving your precious tread at every stop light between here and Diluth. This would mean you're sliding all over the road, and your circumference is: C2 = (3.14159 * 23") = 72.25657".

The percentage change is (C1-C2)/C1 = just over 4%, meaning you can expect your odometer readings to 4% over the lifetime of the tire. Now consider the other contributing factors:

Tire pressure and air temperature contributing to deformation: if your tires are under-inflated or it's cold, your speed will be lower at the same RPMs. Over-inflated or hot, your speed will be higher at the same RPMs. My guess is that this factor is small, something less than 1%.

Integer division hardware rounding to the nearest "tick". I think from reading the technik manual there are 32 "ticks" per rotation. So at most you could be 1/64th rotation off (C1/64 = 1.17809625") which is another 1.5%.

Tire rotation is measured on one tire only, the left rear. If you make more right hand turns than left hand turns your odometer can be miles off in a single trip. Imagine the Indy 500 - the guy who is higher up on the track actually travels miles further than the guy who sticks low. In normal use this is probably insignificant, buy I just wanted to consider all the options.

Tire size may also vary by manufacturer, style, and rim size (16" vs. 17"). These differences may also be slight (around 1%?), but do contribute to the overall problem. The running total so far is somewhere around 7%!

Porsche has obviously recognized this fact, and biased their calibration so your actual speed doesn't exceed what the speedometer says. I'll make a wild guess based on 16" vs. 17" discussing their calibration is made at: maximum PSI on a moderately warm summer day in Zuffenhausen (70-75 degrees farenheit) with brand new Pirelli P0 tires on 16" rims in an unloaded car. But that's just a guess.

My conclusions - it's impractical for Porsche to use a tire-based scheme of measuring velocity and achieve accuracy better than 93%, given the variability of tread depth, tire deformation, the measurement technique, and tolerances in manufacturing. Those who need to know their speed with more accuracy have many choices:

1. Adjust your tire inflation to compensate for temperature and tread loss. Replace tires on a monthly basis to reduce effects due to tread loss. (Or is it just me who loses so much tread?)

2. Get a GPS unit to corrborate your actual velocity. Re-engineer the odometer and speedometer to incorporate this new data input source.

3. Design and patent a new (cost effective) scheme for measuring velocity, then sell this to major auto manufacturers for buku-bucks. Actually, I expect you'd have to sell each unit for about $10 each to make inroads into the market for the current, less acurrate units.

4. Don't worry about it. All cars have the same problem. In fact, these speedometer/odometer units are probably more accurate than their mechanical counterparts of 30 years ago. Just enjoy the car, and go on with life.

Darn. All those math classes are starting to pay off.

[ Home ] [ User INFO ] [ Boxster Board ] [ Archives ] [ Sponsors ] [ Museum ]
[ Events ] [ Polls ] [ Links ] [ Classifieds ] [ Boutique ] [ Pete's Page ]