I have been told that car speedos are a couple of miles an hour faster reading than the car is traveling. At the numerous radar speed warnings in my area this proves to be true.
I have been assuming that was caused by problems with the map database, not the GPS.
Professional GPS engineer here.
Right.
Not sure that I agree. A typical consumer receiver has a doppler-derived velocity estimate accurate to better than 1mph, given reasonable SNR and GDOP (measures of satellite signal strength and constellation geometry, respectively). Certainly in my experience doppler-derived velocity is way less noisy than differentiated position.
Over a long distance/time interval in a straight line, distance divided by time is probably more accurate though. I think it would have to be at least half a mile, which is hard to achieve in practice.
I have a GPS that I use in two different vehicles.
In one vehicle, the MPH is dead-on. In the other, the GPS reads 5 mph faster than the speedometer.
mmm
nobody’s used cables for years. Most if not all cars now just use the ABS sensors- typically on the non-drive wheels- for this. There’s really nothing to “wear out.” At least nothing that would affect accuracy.
I’d like to see a cite for what in this thread has been only anecdotal hearsay (i.e. “I’ve been told”) that automobile manufacturers deliberately set speedometers inaccurately.
However, for further anecdotal evidence: both of my cars (a Ford and a Subaru) have been 99% accurate whenever I’ve passed roadside radar-based speed displays.
My nav unit shows about the same distance my trip meter. (There is no speed readout on my nav unit, which is built into the car.)
I’ve been keeping a spreadsheet since I bought the car from a friend in 2008, to keep track of mileage. In over 300 entries, the nav unit almost always reports MPG optimistically compared to my calculations. I use the trip meter for the number of miles, and the amount of fuel on my pump receipt for the number of gallons. As of yesterday, the average discrepancy between the mpg stated by the nav unit and my calculations is 3.53% greater for the nav unit. (Average calculated mileage is 47.56.) I try to refuel at the same pump, though this is obviously impossible. Generally I refuel at one of two pumps, and usually at the same time of day. I let the pump shut itself off, rather than giving it a couple extra ‘clicks’.
You’ll probably get an even better reading by setting it to a slower speed as it get’s near the top. This way the pump isn’t being shut off by turbulence in the filler neck.
I’ve noticed that my calculated mileage vs what my car says is typically off by about 2MPG. I finally did the math and realized that at 10 gallons, being off by as little as a half a gallon can change my mileage calculations by over 2MPG. For the sake of not wanting to do the math each time, I’ve chosen to trust the car over myself.
I think the GPS is actually supposed to be off so you won’t miss your turn while you drive. I have missed many turns because I drive fast and the GPS could not keep up
Ding! Winner.
I don’t recall the last car I worked on that had a speedo cable. My 1979 Volvo maybe.
Systems on current cars send a signal (usually from the ABS unit) which is most likely the average speed of the wheels (2 or 4 depending on the design) to the cluster where it translates to the speed readout and odo display.
Anyway, I would guess that most cars with electronic speedos come from the factory in the 2%-5% accuracy range.
There are a couple of things that can have an effect. Tire size as has been mentioned, but also tire growth. Brand A and Brand B tires might have the same circumference when the car is parked, but when you spin them up to 70 mph, one of them might have gotten 1" bigger than the other.
I tested mine today. My real GPS unit is so old, I don’t even know if it works, so I used the GPS in my phone. The results were that it showed my speed to be 1-2 mph lower than my speedometer, but I know for a fact that my speedometer is slow by 2-3 mph due to oversized tires. It was using 10 satellites the entire time.
It could have been the phone, but I would expect the GPS on the phone to be as accurate as a car’s nav unit. Neither are meant to be survey grade GPS devices.
I t hink you’re wrong on 2 out of 3 things here:
- Lots of cars do not have ABS, so they can’t use the ABS sensors.
- Most cars track the speed of the drive wheels, not the non-drive wheels.
Both speedometer systems and GPS have inherent inaccuracies (pretty well covered in the above posts) that prevent either one from consistently indicating a vehicle’s true road speed. Now, either CAN show true road speed, and sometimes they do (may be often or rare depending on variables), but there’s no simple answer as to which is more accurate.
In current production cars, I think you would be hard pressed to find a non ABS car. If the car does not have ABS, a simple electronic pickup in the final drive will very adequate.
I’ve had multiple GPS’s in my Saturn and the GPS’s are consistent to each other and within 1 mph of the car at highway speeds (car reads high). The car speedometer is very close to the radar signs used to alert drivers of their speed. I’d put my money on the GPS and would guess the error rate at around 3 % given the altitude fluctuations displayed. Not sure if an altitude error rate is comparable with speed errors.
The manual for my GPS (and I think for most Garmin units) claims speed accuracy of 0.1 MPH.
These guys claim that an accuracy of 0.5 MPH is probably more realistic.
I just turned in a leased Pontiac G6 (2007 model year) earlier this year that did not have ABS. It wasn’t the bottom-tier “value edition” either, it was the base model with the v6 engine package (which added several thousand dollars worth of options, but not ABS).
Indeed it was - me was tired 
I recall helping a friend with an old truck he bought in the late 70’s. Those were the days, it was 10 years old and raised in a salted road environment, you could see daylight out the bottom of the doors. We drove 90 miles at 70 mph steady, then when I did the time vs. distance calculation and looked at the odometer, determined we had actually done 55mph. I suspect the previous owner among other things had traded the wheels for a much smaller set to cash in with someone.
Let’s say a tire is 2 feet outside diameter. An inch less (say, smaller rim) is 1/2 inch or 1 in 24 less. (4%); 2 inches - 8%. Underinflation of the tire to lose an inch radius would give a similar 8% reading error. 5mph in 60mph is 1/12 or 8%. Hmmm…
Oh yeah, the guy also went through a whole tube og grease just lubricating teh frame points. After a few hundred miles the driveshaft fell off, as the only point he missed was the hanger bearing in the middle of the crankshaft. Ah yes, $200 jalopies, those were the days.
Someone may have (also) swapped out the stock differential for one with a different gear ratio to get better acceleration and/or a lower speed range. Assuming the speedo was driven off of the transmission output shaft, this would result in a lower wheel speed for a given indicated speed.