Speedo Error

[pkust]

Recently I purchased a '97 Boxter with a tiptronic and the optional 17" wheels M396 option, while I was driving home I checked the digital speedo with a GPS unit and found that the reading on the car was 4mph higher than the GPS which I have found to be dead on in my wifes Audi A4. On a stetch of highway with the Odo check signs I found that the odo was correct over a 5 mile stretch, has anyone else had any experience with this problem? Any advice on what to do about it other than setting the cruise 4mph higher.

[fastboydave]

Nothing new my friend, mine is too! It has been discussed on here a number of times, so maybe do a search?? Incidentally, i mentioned this to my OPC and they say there is no way of 'calibrating' them! They basically said that as it was within 10%, dont worry about it! Cr*p hey???.......

[savowood]

Mine is 3 m.p.h. off. Actually, when checked with an accelerometer, it was 2.6 to be exact. I have a larger tire on the car than stock. My fronts are 235/40-17 Toyo RA1 instead of the stock Michelin 205/45-17 rubber that came with the car.

I also first noted the difference when I got my sat nav (TomTom) and was curious about the actual difference. Most "officials" will tell you the GPS system as used by commercial entities is not accurate to better than 3 meters or so (I think that's the right number), and better than that is only for military use. They can't guarantee it is accurate. However, even mil-spec sat nav isn't going to be as accurate as an accelerometer.

Also interesting is the digital and analog speedometers on my car are exactly 1 m.p.h. off from each other. The digital is 2.6 and the analog is 1.6.

-Michael

[rick3000]

That is normal. I have measured my speed with my GPS, and with one of those police drive-by speed targets, and my car is off by about 3mph. It is no big deal, just remember that your actually going x mph over, so you don't end up speeding.

[White987S]

Again, very common. Mine too is off by 2-3 MPH when checked agains my Garmin Nuvi 680. True for my Porsche and Ford Explorer.

[pkust]

Again, very common. Mine too is off by 2-3 MPH when checked agains my Garmin Nuvi 680. True for my Porsche and Ford Explorer.

Thanks for the info to all of you, being a newbie to the Boxster this is one of the little things I hadn't heard about before.

[ebk5368]

I posted this a while back but it explains why the error is there and why fixing it is hard.

************************************************************************

I spoke with a friend familiar with the instrument cluster system.

The cluster controller receives a speed signal, typically CAN data from another vehicle controller, and then converts this via an algorithm into both the digital display value and the pointer position. It appears to be a linear function but it doesn't have to be; for example the graphic scale could be expanded on the lower range, of middle range, or upper range, depending on the manufacturer requirements. In the case of my S, such graphical trickery could be useful as the scale covers 175 mph in a 180 degree sweep and is pretty much useless as a ticket avoidance device. The difference between 35 and 40 is pretty hard to see. And also note that the display and pointer position are hopefully matched to avoid user complaints.

To cover any manufacturing tolerance in the dial graphics, and to insure the legal tolerance range is met (remember that tire model, wear, and inflation all have influence), an offset in the algorithm function can be introduced by the system designers, the goal being a skewed tolerance to make sure you don't get a ticket for going 64 when your speedometer says 60. The instrument controller algorithm is not adjustable by mortals. The odometer may or may not have the same issue as it may be driven by a different algorithm and there is no skewed tolerance concern. Having said that I've checked my odometer on long distance trips and it seems to be plus about 4 miles every 100.

It could be possible to trick the system with different tire size, or by finding the source of the speed signal and assuming it counts pulses from a toothed wheel, add or subtract pulses, but this would only fix a specific speed error at one point because such a 'fix' would introduce a percentage of value error and the speedo would only be accurate at the one point where both the curves cross. For example, if as built it is always ~~4mph high. If I increase the tire size by 6.25%, then at 64mph the indicated speed and actual speed would be equal. However, at 120mph indicated, actual speed would be greater, and at 30mph indicated, actual speed would be lower.

An interesting subject but not an easy fix apparently.