How accurate is your speedometer
 

I went to visit my sister this weekend and I was curious about how off my speedometer was. Those 'what's your speed' signs on the side of the road had been showing that my car was right on the money and so I counted how far off the speedometer was by watching the mile markers for 100 miles.

Over 100 miles of road the car got ahead by 2 tenths of a mile. So it is only .2% fast which I didn't really think any cars did except police speedos. My old car was 10% slow so when I was doing 77 it was reading 70.

Marvin used to confuse the hell out of those speed signs for some reason, coming towards one it would be jumping up and down by 10mph each way. However on some measured miles posted on one piece of road, he seemed fairly accurate, but there's only 2 miles worth.

Bear in mind that you're only within .2% until you wear your tires some or get new ones. From full tread to barely legal is a 2-3% difference.

I was comparing my wife's speedometer to the speed indication on her GPS.

Originally, the speedometer was about 1 MPH off from the GPS (highway speeds). Then I started to wonder if her air pressure was high engouh, and after I cranked the air pressure from ~30 to ~36, the speedometer matched the GPS exactly.

-Bob C.

Looking at the spare that's never been used and comparing it to the tires on the car it seems they are just a tad over halfway worn. they still have a lot of tread on them tho.

Mine's about 2% fast with new tires, to 4% fast with old worn tires.

mines is 5 mph over what i should be, if im going 15, im actually going 10 mph. the scanguage gives me an accutate speed cause i called a friend while were both on the highway, the scangauge matched what he said on his speedometer

Does it read 5mph over at every speed, or at highway speed? Speedometer inaccuracy is usually a percentage, so a speedometer that's 10% off would read 5mph off at 50 but 1mph off at 10 (and 8mph off at 80).

If it reads 5mph over at all times, it should be easy to fix by re-aligning the needle or the motor that moves it.

At highway speed my speedometer reading is within 1 MPH of my GPS. Usually it reads 1 MPH slower than my speed reported by the GPS.

-Jay

I thought all speedometers were supposed to error on the fast side by a certain margin?

On the highway I never really look at the speedometer. I look at my GPS for my speed.

-Jay

yea they can be off by a certian percentage, most cases they just program the buffer to compensate, as long as you know your speed is off by XX ammount at X speed your fine. just remember with the manual cable spedos that the odometer is prolly the most accurate while the spedometer can be off.

Keep in mind that the speedo and odometer are separate devices, and can have different degrees of calibration. I did a similar check-the-mile-markers test in my CRX and found my odometer to be accurate within 0.3-0.4%, but according to GPS readings, the speedo reads 6-7% fast. This seems to be a common artifact in Honda speedos, which has led some people speculate that they are intentionally calibrated to read fast. Why? We're in the US... If the speedo read slow, somebody would surely sue when they got a speeding ticket.

Cable drive speedos aren't driven directly but commonly have a "magnetic torque converter" type arrangement where magnets are spun to bias an iron plate or other magnets against a spring. Since magnets lose strength over time, or ferromagnetic materials may become magnetised under the influence of magnets, this means that a speedo in an older car may gradually read slower. When it was new it might have been a few percent high, but by the time it's 10 or 15 years old it might be reading closer to "dead on" or reading slower. It is however possible that the odometer could have a direct driven gear reduction drive off the speedo cable.

Also Chrysler cars of the 80s and 90s for example have about 10 possible final drive ratios, although only 4 were commonly used, and varying tire sizes, but only 3 speedometer drive pinions, 19, 20 and 21 teeth. The factory would put in the closest match for the final drive and tire combo, but as you might guess, this meant that speedo error could vary by model and options across the range.

Also one's measure of odometer accuracy may be affected by the number of curves in the road one is traveling. This is because it measures the speed of the differential, which will be turning as fast as the fastest wheel, so it always measures the "outside" of a curve rather than the line down the middle. So on a series of 10 hairpin bends, you might show 1/10 mile greater than actual distance. However one should realize that the position error on a GPS reading of a route like that will make the comparison questionable anyway. (i.e. you can be 10x the position error off, or greater if it doesn't update very fast)

Wrong. The differential spins at the average speed of the two wheels. If you jack up a car (autos in park, manuals in gear with the engine off) and spin one of the drive wheels, the other side will spin in the opposite direction. The differential is being held stationary by the transmission, so the movement of the two wheels needs to average out to zero.
A more dangerous example is to put the car in drive (still jacked up) at a low speed and stop one drive wheel. Don't try this if your car has a limited slip differential - the wheel won't stop. The other side will pick up speed and spin at double it's original speed. Again, this is because the differential spins at the average speed of the two wheels.

compared to my GPS, my speedo's off 10% and the odometers off 14%. my guess is the odometers more precise (mechanical) and the speedo magnetic spinny part (technical term) is 20 years old and tired. Reason it's so far is I've got a different differential and tires than the speedo gear in the transmission is set up for. (3.9 instead of 4.1 and 205/60r16 instead of 195/65r15)