48

Via word-of-mouth, I have twice heard claims that speedometers are deliberately designed to read high, so manufacturers would not be subject to litigation for speeding.

When I first heard this claim, I thought it was simply an excuse to justify speeding. However, I've noticed that this claim is even on Wikipedia (ref.), albeit unreferenced:

Vehicle manufacturers usually calibrate speedometers to read high by an amount equal to the average error, to ensure that their speedometers never indicate a lower speed than the actual speed of the vehicle, to ensure they are not liable for drivers violating speed limits.

So my question is, simply, is this true?

Question: Do manufacturers deliberately calibrate speedometers to read high?

There seems to be plenty of such claims online, but I didn't find any trustworthy evidence to support the claim.

Moreover, this claim seems inconsistent with government-enforced accuracies on speedometers (i.e., if speedometers were deliberately designed to be incorrect, then they would be less likely to meet government standards).

  • 5
    Tire wear springs to mind first, a not insignificant bias. On a tire with 25 inch overall diameter, 1/8 inch of tread loss means a 1/4 inch change in diameter. This means @50 mph, your speedo will predict you are traveling at 50.5 mph even for a small loss of tread. Drive on bald tires and the bias may be several times that large. – user3344 Aug 8 '12 at 11:35
  • 3
    But I think more important is the "weatherman" bias. (The weatherman tends to predict rain if there is even any chance of rain. Thus people may be happy when it is supposed to rain but does not. But if you predict it will be nice, and a tornado comes though, they are rarely happy. So weather forecasts tend to be biased.) You prefer the speedo predicts high, which lessens the chance of a ticket if it is in error. The manufacturer also should prefer errors on the high side, as that lessens the odds of any potential lawsuit. – user3344 Aug 8 '12 at 11:46
  • 2
    Another good point is, ALL measurement devices will be in error. Errors will often naturally follow a Gaussian-like distribution, thus having tails on either side. But suppose you simply remove all devices that predict low? Now you have a skewed distribution, created not by design, but by restriction where the error is biased to always predict high. Even if you remove only those devices that are below -1xsigma, this still skews the distribution to be biased high. – user3344 Aug 8 '12 at 11:55
  • Here is a source that claims actual speed vs shown speed : click2how.com/your-speedometer-wrong-can-drive-faster - If true that shows up to a 10% margin of error meaning your warranty will wear out 10% earlier than it actually should since the odometer goes off the speedometer. – Chad Aug 8 '12 at 16:06
  • @Chad, any sources that the odometer goes off the speedometer? – ardent Aug 8 '12 at 16:56
13

I don't think speedometers are deliberately set to read high, but you are less likely to get a speeding ticket that way than if they read too low. If they read high, you can guarantee to be below the speed limit, even when your speedometer indicates you're slightly over.

EDIT The first sentence above conflicts with what follows below: if the manufacturer is not permitted to apply a speedo that reads low, that can only be done by making sure it reads high. The only other possibility is to make it read exactly right, which is obviously impossible.

In Australia the indicated speed must not be lower, but can be up to 10% + 4km/hr higher than the actual speed. This applies when the car is tested as prescribed (correctly inflated standard tyres, unloaded, etc). The testing procedure and accuracy specifications are here (Motor Vehicle standards act 1989, Australian design rule 18/03 (.pdf)):

5.3. The speed indicated shall not be less than the true speed of the vehicle. At the test speeds specified in paragraph 5.2.5. above, there shall be the following relationship between the speed displayed (V1) and the true speed (V2).

0 ≤ (V1 - V2) ≤ 0.1 V2 + 4 km/h

Because of this, police in Victoria only let you go about 2km/hr over the speed limit before they book you. As far as I know, New South Wales still allows you 10 km/hr.

More info about speeding laws in Australia:

http://www.trafficlaw.com.au/speedos.html

  • This answer (along with Tom77's answer) is pretty clear-cut. I'll accept this one, since I'm Australian. Although, the choice was mostly arbitrary. – Douglas S. Stones Sep 10 '12 at 22:33
56

In the EU, speedometers are not allowed to display a speed that is lower than the speed the vehicle is traveling. Manufacturers calibrate speedometers to comply with this regulation as they would not be allowed to sell their cars if they did not.

Reference - EU Regulations for Speedometers (75/443/EEC)

The speed indicated must never be less than the true speed. At the speeds specified for the test in 4.3.5 above and between these speeds, there shall be the following relationship between the speed indicated on the dial of the speedometer (V1) and the true speed (V2):

formula

  • 8
    The US DOT has a similar requirement [fmcsa.dot.gov/rules-regulations/administration/fmcsr/… that the measured value @ 50mph be within +/- 5mph (80km/h +/- 8km/h). If I were a speedometer manufacturer I know which way I'd calibrate my errors... – KutuluMike Aug 8 '12 at 14:42
  • 1
    Speedo manufacturers could be held liable if the dial read low and you, trusting it, were caught speeding, so because there will always be some error, the manufacturers calibrate to make the lower error bar higher than true speed. – Rory Alsop Aug 8 '12 at 14:49
  • Thanks! That's pretty solid evidence to show the claim is true in the EU. Although it leaves open the question as to what happens outside the EU. – Douglas S. Stones Aug 8 '12 at 22:46
  • The only evidence i have is that i recently emigrated from europe to outside europe. My car used to read about 4mph too fast, now it reads about 5kph too fast. So i suspect the same situation exists. – Sirex Aug 9 '12 at 4:12
-3

@Tom77's answer is correct, in that the speedometer must be designed to already read a little high to avoid ever reading lower than actual speed. However, that rule only applies to the speedometer on the instrument cluster.

Often there is a second speedometer based on the car's GPS unit that is part of the navigation system. Some systems can display it, e.g. the standard Hyundai/Kia navigation system can show current speed, and it displays the number read from the GPS receiver.

The GPS receiver can be both too high and too low. The speed calculation is based on a straight line between the current and previous measurement, typically made at 1 second intervals although 0.1 second intervals are also possible on most modern receivers. As such corners are not accounted for, and the GPS system is not perfect either (accuracy is typically better than +/-5m but can fluctuate, especially in a moving car).

Some of the information contained in this post requires additional references. Please edit to add citations to reliable sources that support the assertions made here. Unsourced material may be disputed or deleted.

  • 1
    There are two ways a GPS can compute speed. One is by calculating from position fixes, as you state, and the other is by measuring the Doppler shift. Doppler shift measurements are extremely accurate. – Mark Feb 15 at 1:57
  • 1
    My Garmin GPS shows current speed and it matches EXACTLY my digital speedometer in my 2012 Honda at all speeds.This has been the case for all of my recent Hondas except when I changed to wheels and tires that don't exactly match the factory spec. My Suzuki motorcycle (dead stock) on the other hand is off by about 5 MPH at 70 MPH (it reads high). It has since day 1. – Tim Nevins Feb 15 at 14:27

You must log in to answer this question.

protected by Community Mar 12 '14 at 15:23

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

Not the answer you're looking for? Browse other questions tagged .