There's a way to compare it with GPS either in the service menu, or if you have GPS in a phone device. Or... just drive by one of those radar speed limit signs that posts your speed vs. the limit. I intend to do that soon as I found the Gen II w/ non OE tires to read 2-4mph fast. Gen 1 was spot on within .5 mph @ 60mph.
I used my Blackberry Maps GPS today to compare speeds. At 27mph indicated, the blackberry shows 26mph. At 36mph indicated, BB shows 35. At 47mph, BB shows 45. At 62mph, BB shows 60mph. So like Gen II, we're looking at a built in 1-2+mph discrepancy which will only widen as tires wear. I'll have to check at 75mph indicated, but imagine we'll be close to 3mph high there. No wonder people "tailgate" when we think we're doing the posted speed limit.
At speedo indicated 65, GPS says 63. Not bad. They are allowed a much wider range of error than that. My Smart car is 4mph off from my GPS.
Are the gps readings 100% accurate? I thought they were a little off as well. I think the best way is to do 30-50 mph over the speed limit and wait to see what the cop writes on the ticket.
Mile post indicators help too - in my Gen II the odo rolled 1 mile before the mile post was reached, so there was true error. While some error is allowed, they're only allowed to error on the faster than actual side. Speedo cannot read slower than actual in a stock configuration.
Subtle. Hypothetically at 62mph doing actual 60, burning 1.0GPH (not really, but let's assume), you'd be getting actual 60mpg vs. displayed 62mpg.
Depending on the model most GPS units are accurate in spotting a specific spot to within 10 to 30 feet. However if that "spot" is moving (as in tracking mph's in a car) then the error in insignificant in comparison to the distance travelled. I can't remember the details just now but have documented it in Edmunds forums, but the TCH had an error in the spedo and an error in the odometer. It was one shuch that the speed was reading too high but the odometer was reading too low. Changing out the tire diameter to make one measurement more correct simply made the other worse. I'll be curious when I get my Prius to check this out. The odometer was about 3% off in the TCH and when I did manual calculations I took this into consideration and the FE was better than what one might have thought. Then again many people wanted to use their trip computer and it was typically overstating the FE by 1 mpg on average. I found that the adjusted manual calculation was about evenly in between the Computer reading and the unadjusted manual calculation. In a car like the Prius where the FE is significantly higher I will beinterested to see just how big the error is numerically.
Based on experience with 20th century cars, I don't presume that the odo and speedo have the same error. So I still use the pre-GPS method of counting seconds on a watch between mileposts while holding constant speed, and comparing odometer to mileposts over long distances. Not having received my Prius yet, I'm still accustomed to analog gauges. It is reasonable to get speedo calibration to better than one-half mph, and odo calibration to a fraction of a percent.
2 mph over-read at around 60 mph isn't much of an over-read. I understand that almost all cars will have that amount of over-read or more. I read somewhere on PriusChat that according to international requirements, car manufacturers are allowed to calibrate a speedometer to over-read by as much as 10%, but under no circumstances are they allowed to calibrate a speedometer to under-read by any amount.
If the speedometer and odomerter errors are not proportionally related and they were not in the TCH then the odometer error is the one of most significance to be tracking. My TCH was 3% off in the odometer reading with the car actually travelling more distance than it was accounting for with the odometer. That resulted in an actual FE 3% higher than manually calculated using miles and gallons used.
