CarTalk.com Best of Deals Car Reviews Repair Shops Cars A-Z Radio Show

When is 40=42?

While driving my '02 Honda 40 mph on the way home the other night, I was passed by a newer Honda, also going 40 mph. I could tell he was going 40 by the giant glowing numbers on his dashboard. I increased my speed to match his, and my low-tech speedometer (you know, the big red stick thing that points to the numbers) said I was going 42. Even in a car 8 years old, I was expecting something better than a 5% inaccuracy. Anything worth following up on here?

How do you know it wasn’t his that was inaccurate? Or maybe both were inaccurate and you were both doing 43 MPH.

If you have different sized tires than the car shipped with, this can cause this. Anyway, your problem isn’t uncommon. Most speedos are off by a mile or two. If you look at the dash on a cop car, that’s why those will say “certified calibration”

Wasn’t there an issue with older Hondas where they had to extend the warranty because the odometers were rolling too fast, meaning the speedometers also read too high?

You get a digital stopwatch and do 60 MPH. When you hit the mile markers at exactly 60 seconds, you know your speedometer is accurate. It’s the only way.

No, that was Nissan, IIRC.

Honda did get in trouble because of setting there odometers to read high within the maximum legal percentage so the warranty would run out faster. Honda did have to compensate its customers.

If you have ever use a trundle wheel to measure distance, you’ll notice they never use pneumatic tires. Differences in air pressure and tread wear, just two factors, can easily account for differences in registered speed in cars. When we go to range finder technology, expect them to be more accurate. Until then, deal with it. Unless you’re willing to give up the comfort of pneumatic tires for now, make sure your pressure and tread wear measures are up to recommendation, and let the police worry about the actual speed of the other guy…it ain’t the car “low tech” that’s that far off.

For what purpose would you follow up? There are shops that check speedometers. You can have yours checked if you want. I’d bet that if you ran an average you’d find a majority of cars running a few over or under. I’d worry about it mostly if you’re worried about speed limits & tickets. Otherwise, I don’t think its worth worrying about.

Just a follow up…a difference of .3" inch deflection in radius which is little, due to air pressure or load is a difference of 2.3% in accuracy. You’re nearly half way to your so called speedometer inaccuracy. It’s impossible to judge relative accuracy of car speedos unless you set up “laboratory conditions” and eliminate as many factors as possible…
and like “pleasedodgevan2” suggest to get an estimate (and that’s all it will be) and only care about your own. Personally, mine are always withing a mile or two of my GPS unit estimates.

A GPS with speed readout on a flat road will do the trick, too.

How old are your tires??? How old were the tires on the other vehicle???

The tire difference…plus the inaccuracy of the speedo…could easily be 5%.

Actually that would only tell you it’s accurate at 60mph. Probably pretty accurate, but it could still be off somewhat at other speeds.

This is a classic case of random error versus systematic error. The speedometer is almost perfectly precise, but it can obviously skew a certain percentage fast or slow. The GPS will be more accurate, but is only precise to a circle with a radius of 20 feet or so, and so the instant speed reading (which is the measure of time over distance between two position fixes) is not very accurate, especially at low speeds.

However, over time, all the random error should average out, so if you can set an odometer on the GPS and compare it to the car odometer it should show you an accurate number. But another consequence of the random readings is that when you’re stopped, it records the distance between the random error as distance travelled, so if you do a lot of stop-and-go driving this won’t work either, so you have to take a nice long road trip to get an accurate reading.

Like I mentioned briefly above, speedometers skew off a certain percentage. I can’t think of a mechanism that would cause it to be off more or less at different speeds, so if you can do a time over distance measurement at any constant rate, you can extrapolate the error at that speed to any other speed.

I’ve had a couple cars (well, actually usually trucks since they tend to end up with oddball tire sizes) where the speedometers were way off, and I just printed off a little table that had indicated speeds in one column and actual speeds in the other. (Also handy for my '92 Isuzu which for some reason doesn’t have metric number on the speedo!)

5% is perfectly normal. And acceptable. I base that on having driven numerous cars past those radar cop trailers that told me how fast I was going. Typically it ahows one going about 2 mph faster than one actually is.