The speedometers in cars are notoriously inaccurate. This is something we have always known but with the ubiquity of GPS devices with rather more accurate (if lumpy) speed measuring this has become much more obvious. They are always calibrated to read high so that the inaccuracy will never cause you to inadvertently exceed the speed limit.

This is annoying. Like most drivers with some experience I don’t read the speedometer very often but judge it mentally, only checking the dial when changing speeds (or spotting a camera, but that is another matter). However, even then, we are judging our speed by a previously learned experience based on an inaccurate instrument. It is really noticeable when stuck behind someone doing exactly what they think is the speed limit when you know that they could be going 5 or more mph faster.

The law says that speedometers should be between -0% to +10%(+4 km/h) accurate. As they cannot achieve the -0% by design, they always over read a bit. They work by counting revolutions of the wheels and I thought that part of the inaccuracy was due to tyre wear so I did some calculations. I have 195/65R15 tyres; these are nominally 25″ diameter, i.e. about 807 revolutions per mile. If we assume that a new tyre has about 3/8″ of tread and when worn it is reduced to 1/8″ that is 1/2″ loss on the diameter and now 823 revolutions per mile. This is enough to cause the meter to read 2% fast; certainly significant but not as much as the discrepancy that I observe (nearer 6%). Changing your tyre or wheel size or the differential can make even bigger changes if you don’t get the speedometer re-calibrated.

It ought to be possible to make speedometers more accurate. They could use GPS technology—by setting my handheld device to sample quite frequently it becomes relatively smooth and if this were further smoothed by software and translated to an analogue device it could make a good and accurate meter. Alternatively it should be possible to adapt optical mouse technology. These work on any suitable rough surface, the road would be ideal, and the only thing that would need to be changed would be the range up to about 400 mm to the underside of the car rather than 4mm to the mouse mat and this would probably need a more powerful laser. Pointing the device at the tyre would reduce this necessary range. (I thought of this independently but I see that halfbakery has a similar idea—which proves how daft it is)

P.S. The ACPO formula for determining speed violations is currently 10% +2mph. Police car speedometers are calibrated -5% to +0% (the other way around to ordinary ones) so that they don’t think you are going faster than you really are. Camera speed traps are triggered at an even higher point than this but not enough to compensate for thinking you were in a different speed zone. Those annoying electronic flashing “SLOW DOWN” notices seem to be randomly set and very inaccurate, often reporting about the car three or more in front of you. They have no weight in law.

A House of Lords reply indicated that speedometers are actually manufactured to a -0% to +5%(+10 km/h) tolerance which over-reads still more, not meeting the legal requirement until 120 km/h (about 70mph). In theory when you think you are doing 30 you could be only doing 24 mph though I don’t think many are that bad.

Comments are closed.

^ Top