I've just checked my speedometer against a police trailer at the side of the road which flashed up what it thought my speed was. My speedometer said 40 and the trailer said 36. Is it likely that the error of +11% will apply at all speeds, or will the percentage error vary throughout the speed range?
Who says that the trailer is 100% accurate.
As it is set to display only whole numbers, you may actually have been travelling at 36.49mph.
This puts your speedo accuracy within the 10% allowed.
As to whether this applies over the whole speed range, who knows.
I find that a satnav used on a road without a gradient gives a fairly good idea of actual speed that you can compare with your speedo.
A few months back I was on an HST class 43 (a modified one with electric doors) on a stretch of railway with a 125mph limit. I put my satnav out at it for a good five minutes it was reading a solid 125mph.
Sat navs are very accurate on straight flat bits of road the speed can't really be that wrong, the limit on sat navs is the software which controls the navigation.
Did a quick search for scientific papers on this to no avail so there may be an opportunity here for an Honest John test.
All we need is access to the long perfectly flat straight in Germany (J. May did the Veyron Test there), GPS, couple of cars, and some sort of approved speed detection device.
Fifth wheel? They used to check speedometers during road tests in the fifties and sixties with a sort of bicycle wheel device, carefully calibrated, attached to the car's back bumper. Used to publish a scale of speedometer and true speeds all the way up to the car's maximum too.
SFAIK the signal we get in our cars etc is in some way degraded from the accuracy avaliable to military who can put a weapon on a given location +/- about 6 ft. I think sat nav is used for things like the construction and alignment of bridges? One sees people walking around with odd back-packs with aerials stick up out of them. I think they are surveyors etc.
Yes the SATNAV that we have is a "Numbed Down" version of what is used for the milatery. where as you say the military have accuracy of 6ft (and maybe less), civilian satnavs will have a greater scope for inaccuracy, and Im lead to be beleive that this has been done on purpose.
Actually my maths was a bit out, its probably more like 79k. The reason I suspect that is at an idicated 70 my satnav tells me I am doing 65. European speedos always seem to over read. My dads Fiesta is the worst, an indicated 80 is actually 73.
What does that do to all the associated readings such as total miles travelled?
Speedos are deliberately inaccurate, so that no driver can claim he was 'doing 29' when police claim he was doing more. You can check your odometer with GPS, or against the 100-metre posts (or the recent half-kilometre signs) along the M-ways. My registered miles are over-read by 0.4% - much more 'accurate' than the speedo.
Optimist - my speedo over-reads by 3 mph right across the range. I happened to notice today that, as my computer read out of distance travelled changed to 430 the odometer read 430.2 which seems pretty accurate to me.
As far as I understand it, the Speedometer and Odometer in my car are both driven by the same rotating cable. This means that if the Speedometer is overestimating by +10% (actually allowed to be between 100% - 110% +6mph under uk regs), the Odometer will also increment by +10%. That is, for every 1000 miles travelled, the Odometer will increment by 1100 miles. This seems to be somewhat unfair in terms of the resale value of my car. This deliberate over estimation is also in the interests of manufacturers, as it will allow them to claim better mpg figures than is actually the case. Whilst appreciating the safety aspects of not under displaying actual vehicle speed, I would think that it should be possible for Speedometers to be accurate to within +2 or 3 mph. Out of interest, I have found that the Speedometer in my Dodge Caliber (manufactured under u.s. regulations) to be far more accurate. Come on Europe, time to catch up.
Yes the SATNAV that we have is a "Numbed Down" version of what is used
for the milatery.
You might not be able to determine your current position to the same accuracy, but will that affect a speed calculation, which takes an average of a number of readings? As long as the inaccuracy is less that a certain amount, say 20ft, then I don't see how that will make a significant difference.
Most (99%) GPS chipsets do not derive speed from geographic information as some of this forum believe. They calculate speed (more accurately, 3D velocity) by measuring Doppler shifts of the GPS D-Band carriers tramsitted by the satellites. These signals are affected by various scenarios, but in good reception conditions (i.e. not urban canyon) the speed from the receiver chipset can usually be considered accurate to approx 1kph. HOWEVER, bear in mind the speed signal is averaged (sometimes over a number of seconds) and is not instantaneous speed, it is also just the 2D component of a 3D velocity so changes in altitude (climbing/descending) will affect its accuracy. GPS is subject to signal dropout in which most chipsets switch to simple dead reckoning (speed is assumed to be fixed) . Manufactuers of end user products, e.g. TomTom add their own layer of averaging, and many "frigs", to make things fit and look reasonable, to cope with short term inaccuracies, loss or deteriation of signal. They also ROUND numbers down and up. So in short - dont rely on that GPS in a court of law - it could be +/- 2mph even in good conditions.
Motor car speedos by law have to be in the range 0% to +10%. So invariably they always indicate a faster speed than actual. If a manufacturer calibrates a speedo to be +5% , then be aware there are other factors that affect measurement accuracy. There is an effect between the tyes and the road often called "creep" and "slip". These are largely unquantifiable variables which can typically give +/-2% variations depending on road surface, going up/down hills etc etc. Of course rolling radius of a tyre is changing constantly too, not to mention wear, which is a longer term issue.
Odometers and speedometers are usually independently calibrated on cars (there is no 0-10% requirement by law on odometers) - an innacurate speedo doesnt mean the odometer is the same.
As I have said before on this forum I owned a Primera and that showed 77/78 mph on speedo whilst tom tom showed 70 mph. Changed to C5 and speedo and sat nav are usually within 1 mph. All those years in Primera thinking I was breaking speed limit!!!!!!
Dont assume your GPS is always right! -
kiss (keep it simple)
Another way to calibrate your speedo or odometer is to use the 100m posts alongside the hard shoulder. Count the time taken to pass 16 posts ( 1600m is very close to 1 mile). If you are doing 60mph it takes one minute ( at a constant speed of course). A bit of simple arithmetic is all that is required if your time is different or you want to check other speeds.
My GPS is spot on, the only car I have had which was accurate was a Saab 99.
Its is easy to spot those that are driving to their sat navs during an average speed check as they are generally doing 10mph faster than the rest!! In the days before sat nav I would always cruise motorways (50K a year) at an indicated 80mph to allow for the 10% ish error and knowing the extra few mph of true speed would'nt merit a tug.
Dont assume your GPS is always right! -
Just for a giggle last week I installed a little app called 'Speed' on my iPhone. This merely does what I presume satnavs do: query the phone's GPS for speed and display the result on the screen. Two observations:
1. On a clear M40 on Friday, it roughly confirmed what a stopwatch and the 1600m test had told me ages ago - that a true 70 reads as about 75 on the dial.
2. There was a surprising degree of wobble and lag on the reading. With the car's speed fixed by the cruise control and the needle unmoving, the GPS reading would vary by up to 2mph. It also seemed to change with the curvature and gradient of the road, and to take a couple of seconds to reflect any true change in speed.
A third observation: all that GPSing ate the iPhone's battery in barely an hour.
The Saab comment is interesting. I had a 1996 900 and a 1998 9-3 and both claimed to have very accurate speedometers. This was supposedly because they used data from the ABS sensors, but I'd expect that to be true of pretty well any car from that period onward. I never tested the claim for myself, so does anyone here have any observations to back it up?
Dont assume your GPS is always right! -
kiss (keep it simple)
The Saab was from 1975 so no fancy electronics. Incidentally the speedo only went to 110mph I think, which was just about enough. It made it easy to read too because there was a number every 10mph. Even the "sporty" fuel injection EMS version stopped at 110. Could it be one of the few cars which could go faster than the speedo? My Mazda6 goes to 150 so it doesn't go much above half way most of the time.
I've only just read this topic, so a few points which I'd have made earlier if I'd been looking:
The Spanish test taximeters apparently using a rolling road
The UK calibrate HGV tachographs in a similar way.
If the speedo is inaccurate by something approaching 10%, what does that do to all the associated readings such as total miles travelled, trip miles travelled, miles to refuel and mpg?
Quite a lot, I should think. Indication of miles travelled would be higher than actual distance travelled by the car, mpg figure would appear to be better than it actually was.
A cynic would say that this "feature" is a deliberate attempt by manufacturers to squeeze services slightly closer together, thereby increasing the number of oil changes etc carried out by dealers along with all the ancilliary work and profit opportunities that entails. For example, if a company retains its fleet vehicles for 3 years then some of them are going to have one service more than they "need" because of the spurious higher mileage reading. Extrapolate that across the entire UK dealer serviced vehicle parc and that's a lot of extra income...
Could it be one of the few cars which could go faster than the speedo?
My 86/C FIAT Panda 1000S speedo once ran to 4mph off the top of the dial. Alright, the top was only 100mph but still.
Also saw the same effect once on an 89F BMW 328i with a 150mph speedo and a very long downhill run-up.
Original Mini with, single central speedo , downhill, with following wind etc., would go past the end of the scale 90 and reach the ignition warning light with valves bouncing at about 6,000 rpm, or if you had shimmed them up or fitted double springs then you might reach the F on the fuel gauge - Poor little A series!