My Sat Nav (GPS) above about 20mph always shows my speed to be higher than my cars speedometer, by about 5mph. It seemed to me that the car speedometer would be correct but a friend thinks the Sat Nav is correct. He claims cars speedos are deliberately set, by the manufacturer, to display a lower speed to avoid litigation, e.g. so if you got stopped for speeding by the police, you couldn’t sue the manufacturer.
I would think if anything it would be set to show your going faster. I’d rather my speedometer said I was going 60 when I was really going 55 then the other way around.
Secondly, I would trust the speedometer over the NAV.
A simple way to tell is set yourself on cruise control, and count mile markers. If after an hour at 60 mph you’ve hit 55 mile markers, the GPS is correct.
I have a Garmin nuvi and experience something similar. On interstates I have compared my car’s trip odometer (main odo does not show tenths of a mile) to mile markers and found it to be very accurate, while the GPS is off by maybe 5%. My GPS shows miles to the hundredths.
That’s for distance. I assume that the speedometer and odometer are working from the same input, but maybe not.
Isn’t this what the OP is saying is the case?
As far as I know, all car manufactures calibrate their speedometers to show a value higher than in reality. All cars I’ve ever driven that I’ve actually tested for this, have shown this to be true. (Typically 8-10 kmh higher, which corresponds nicely to the 5 mph the OP is quiting.)
I’ve tested this by running the car at exactly the posted speed limit (typically 80 kmh) past speed information boards, which are quite often set up alongside highways in norway. They display your actual speed as you pass a marker. When I run past one with the needle of the speedometer as accurately on 80 kmh as I can, I tend to get a feedback of traveling in the 70-73/74 kmh area.
ETA: One should always check this oneself of course, before one rely on pushing the limit and still avoid a fine 
That’s unusual - normally the satnav will show a lower speed because, as mentioned, the cars speedo is designed to read a little high. IIRC the tolerance spec is 0% to +10% (at least in the UK), so assuming a symmetrical gaussian distribution in the batch spread the manufacturers would be wise to set the nominal reading to +5%. It’s not just the accuracy of the speedo system that affects this, the effective diameter of the roadwheels will also play a part. If the tyres are a bit worn or under-inflated, then the speedo will read a little high.
I’d be inclined to believe the satnav, bearing in mind the speed reading it displays is a rolling average so you need to be cruising at a constant speed for a few seconds for the reading to be accurate.
I run a couple of GPS devices together (one camera locator, one satnav), and they always agree with each other. When the GPSs show 70 mph, the speedo will say 74.
No, it’s the opposite of what the OP said.
Talk about overtesting. How about set your speed, measure your time to go one mile and then calculate your speed. Whichever one is right is right.
Changing your tire diameter or changing the gear ratio of your car can affect the speedometer readout. I have a tuner for my car and when I change to my snow tires in the winter (16" wheels from 17" wheels) I need to input the tire circumference.
My GPS unit tends to read about 10% lower than the car speedo - that is, when the sat-nav says 70mph, the speedo reads about 78mph. As others have said, I assume the car speedo is deliberately set to over-read. The OP’s case is unusual.
Instantanious speed readouts from GPS units are derived from measuring Doppler shifts of the satellite transmission frequency. This is nowhere near as accurate as a speed estimate made from calculating distance and time from GPS position fixes, but is much faster, with more frequent updates and lower latency, and thus more useful. Vertical motion can upset the speed estimates, as well as their being generally more noisy.
Never mind – answered above.
I think the sat nav will be more accurate, provided you are travelling in a straight line at a constant speed. A sat nav measures knows where you are, not which way you are pointing (it determines direction of travel by checking the last position against the latest one). It calculates speed by seeing how far you travel in a time period. I don’t know how sophisticated this algorithm would be. Changes in elevation wouldn’t be a problem, that can be measured by checking the position, but I don’t know if it would try to do any curve fitting when travelling around corners. If not, it would underestimate your speed when travelling around a corner, as the path travelled would be longer than apparent simply by measuring positons.
On the other hand, the speedometer should be a more precise intrument, and should exhibit less lag when you change your speed. However, speedos are typically calibrated to show a higher than actual speed reading, for legal reasons.
So, if speedometers are calibrated to be high, how about the unattended radar signs on the side of the road? We have a permanent one near our house, and several temporary ones. I’ve driven past them in three different vehicles in the last year, and they always read precisely what my speedometer reads.
Either your speedometer is calibrated correctly (some are, different manufacturors use different tolerances) or the radar sign’s calibration is off by a very similar amount to the speedo in your car.
Agree that actual measurement by taking the time over a fixed distance is best (unless you have access to a calibrated radar gun). I’m assuming (and I think this is a good assumption) that interstate highway mile markers are surveyed and accurate enough for this test.
Doing it over an hour has an advantage in that time measurement error is reduced (a second or two error in when you stop and start the stopwatch isn’t a big deal when the total time is 3600 seconds) and the disadvantage that it is very hard (and probably impossible in traffic) to hold your speed exactly constant for that distance, even with cruise control.
Doing it over 1 mile and it becomes easier to hold your speed constant, but you need to be pretty good with starting and stopping the clock.
Decent compromise is to do this over 5 or 10 miles, depending on what roads and traffic you have available.
My guess is that the GPS is accurate, assuming that it is getting a signal from an adequate number of satellites. If the route has trees, buildings, hills, etc. that block reception, it may be less accurate.
Another thing to consider is the gradation of the speedometer - most that I’ve seen have markings in either 2 or 5 mph increments, though a few have 1 mph markings. If yours has a mark every 5 mph (e.g., 55-60-65-70 mph), you are simply guessing if you think you can read it as say, 68 mph.
A good quality GPS unit is, IMHO, always going to be more accurate than a mechanical device like a speedometer in your car. The speedometer in your car is rarely going to be dead-on because of the number of variables.[ul]
[li]Mechanical wear- besides the cable, speedometers have LOTS of moving parts, which will all wear over time, decreasing accuracy[/li][li]Manufacturing tolerances - All speedos don’t leave the factory working perfectly.[/li][li]Intentionally incorrect speedos - as stated earlier, manufacturers tend to build in a little fudge factor. My Element speeds shows 63, while two different GPS devices show 60, and a measured mile confirms 60.[/li][li]And, the big one- TIRES. Tires are a HUGE variable in two ways. (a) Tire manufacturers are inconsistent about circumference, even in the same size tire. A 215/75-16 BFG tire shows a different circumference than the same size tire from GoodYear. That’s error right there. And then (b) a worn tire has less tread, and therefor less circumference than a new tire. Error right there too.[/li][/ul]
Here’s a clipping from Wiki on GPS performance versus mechanical speedos,
GPS devices are capable of showing speed readings based on how far the receiver has moved since the last measurement. As the GPS is an independent system, its speed calculations are not subject to the same sources of error as the vehicle’s speedometer. Instead, the GPS’s positional accuracy, and therefore the accuracy of its calculated speed, is dependent on the satellite signal quality at the time. Speed calculations will be more accurate at higher speeds, when the ratio of positional error to positional change is lower. The GPS software may also use a moving average calculation to reduce error.
As mentioned in the satnav article, GPS data has been used to overturn a speeding ticket; the GPS logs showed the defendant traveling below the speed limit when they were ticketed. That the data came from a GPS device was likely less important than the fact that it was logged; logs from the vehicle’s speedometer could likely have been used instead, had they existed.*
Interesting conversation… I have always trusted my GPS speed over my speedometer for the reasons leftfield6 mentioned about. I’d add tire pressure to his/her point about tires.
Although now that I’ve read this conversation, I guess I will have to re-evaluate whether I can trust my GPS. I know where some marks used by police helicopters are on the interstate near my home. Maybe I can get my wife to time us next time I go by there.
GPS isn’t necessarily always more accurate. They don’t factor altitude into the equation to determine speed and, even if they did, the altitude readings aren’t always that accurate on consumer grade GPS units anyway, so if you’re going up and down hills, it will be off.
Unless you’re sitting still and letting it average, at any given time, your GPS coordinates may only be accurate to X meters, with X being nearly anything.
In theory, it should be more accurate, but how often does the little symbol representing your vehicle stray off the road, cut inside a curve, go straight for a few seconds after you made a turn or go anywhere else you didn’t go?
What speed is shown does not indicate the winner of the race.
My speed is what the LEO said it was. Doesn’t matter what I say.