The speedometer in my last car was out by 9% (indicated 109 / actual 100), so I checked the speedometer on my 2017 RDX.

Less than 1% out, so basically spot on.

Indicated / Actual (KPH)
60 / 60
80 / 79
100 / 99
120 / 119

GPS checks done on a straight and level road with the cruise set.

 

Awesome

 

I've always wondered how accurate those gps systems are, when it comes to speed. I mean, it's shooting a beam out to satellites in space, and back to this 100-400 dollar device. It seems there would be room for error. Hell, the gps systems we use at work for surveying and AutoCAD stuff are only good up to, I think, 10 cm... that's out by four inches. Using equipment that costs over 10k.

If im not mistaken, an auto manufacturer has wiggle room with speedometers. They have to be accurate to within a few percent. Being 9 km/h out seems like it would be way out of tolerance. 3-4 km/h? Sure. But 9 km/h?

 

Though, there may be other reasons why your speedometer was out by so much. If you ever had different sized tires, or different sized wheels, that can account for a difference in itself. Hell, comparing brand new tires to used tires can cause a difference in the reading. I'm not surprised your new RDX appears to be spot on... you're likely on brand new tires, on a brand new car, with likely zero issues. The question remains- what was your old car and did you change anything on it?

 

That's not the way GPS works. It doesn't 'shoot a beam', it's a receive only system. The receiver get signals from several satellites and calculates the position based on differences in the timing of the received signals.

I doubt that the RDX speedo is based on GPS. Most automotive speedos read around 5% fast. If your car speedo is right on it's mostly just a lucky coincidence. It will change as your tires wear and when you replace them, among other factors.

 

New car and speedometer was checked within a few weeks of buying it. Therefore new tires as well.

I ran a standards calibration lab for over 15 years so measurements are not new to me. And under the conditions I outlined, a GPS speed indication is far more than accurate enough. There are three (I believe) time bases on each GPS satellite and each one is better than 1 second in 75,000 years and that is a conservative number.

 

Who said it was?