I was going up to Preston today in a friends clio, according to her speedo she was doing 80mph, her sat nav (tomtom) was saying 72mph. Which one of these is most likely to be accurate?
It just seems so amazing the difference. I am guessing she must have been doing 75mph.
|
SFAIK Sat Nav speed is way more accurate than a car speedo which is only required to be within 10% but must not underead. However, car SatNavs do not know if you you are going up or down a hill and this can lead to very slight errors, not enough to bother you with a speed camera in the area, but very small errors.
|
|
Satnav is accurate as long as the road is straight and roughly level.
|
|
|
What AS says.
Pretty much every car I've owned in the past 10yrs has overestimated speed by 5-10% on the speedo. I've checked this against motorway marker posts that are known to be very accurate (every 100m), timing a constant 62mph (100kph) against 10km of M-way (whatever keeps you awake!).
Even had both my Sat Navs running at the same time (Navman and older Tomtom) - virtually identical speeds shown and consistent with measured distances as above.
Significant gradients may have an effect as AS suggests, but Pythagoras's theorem would mean even going up/down a 1 in 10 grade (pretty sttep at 60+ mph!) the forward velocity would only be under-reading by around 1%.
|
|
This is because the Law on speedo's specifies that they will not under-read 0% but will over-read and there is a spec but I can not find it right now. 70 mph actual, is often displayed at 74 to 77. Sat navs are basically accurate although some do not increment in 1 mph increments. If your car speedo is reading low check your tyre size and profile are compliant with the spec for the car. Regards Peter
|
Cheaper satnavs and camera locators can lose satellites and 'estimate' speeds, so cannot be relied upon all the time. But on a flat straight road in good weather conditions when they can pick up 5 - 6 satellites they will give a 100% accurate reading of your speed across the surface of the planet and expose your speedometer as an optimistic liar.
HJ
Edited by Honestjohn on 01/11/2008 at 09:48
|
Variation in height makes no difference to speed reading (no car is capable of changing height fast enough to make any variation). Straigtness of road does, and its always a fraction behind when changing speed. But straight, constant speed its 100% (or near enough) accurate.
|
Rattle
Thanks for asking the question - I have just found my snooper can be programed to show speed and I noticed it was about 4mph slower than speedo and was wondering which one was more accurate.
|
Variation in height makes no difference to speed reading (no car is capable of changing height fast enough to make any variation). <<
GPS can be used to measure altitude, but I presume on cars they simply measure changes in X-Y axes to work out distance travelled.
On a 1 in 10 gradient, for every mile you drive up/down the gradient, you only travel 0.995 miles along the horizontal axis. 0.5% inaccuracy is not zero, but it's pretty close.
On a 1 in 5 (think ski-slope) this would be 0.981, or approx 2% out.
Not zero, but fairly irrelevant unless there's a speed camera on a very steep slope.
|
|
Thanks this would certainly explain a few things. My ex likes to drive a bit fast and is happy plumeting along at 80-85mph in her 1.2 Clio, at least I now know she is really doing 70-75ish :
|
|
|
Certainly there will be no discernible error on a road of any gradient you are likely to try in a motor vehicle and while sober! Taking the extreme of a right angle triangle, sides 3x4x5. If you drove down the hypoteneuse, 5 units, a sat nav would think you had travelled along the 4 unit side, giving a 25% under-read error. Entirely hypothetical!
|
|
|
|
|
>>they will give a 100% accurate
There's no such thing as a completely accurate instrument. There is always some error.
What is more common is an instrument with an acceptable degree of error when considering its intended purpose.
|
NC
Sad thing about this is no one has told the average scamera partnership. When 'done' a few years ago, I was sent a 'calibration' of the meter. No indication of the test method, just a list of measurement points at 10 mile/hr intervals and a set of instrument readings, which (surprise, surprise) read exactly the same as the measurement point to the nearest whole number. No attempt to show a reading to an order of magnitude better than the measurement point. Meeting the manager of the partnership a while later I queried this and was told 'these instruments are totally accurate'. I tried to explain about measurement error and calibration curves for measurement instruments used in industry, but it was totally lost on him, as was anything to do with the angle of the beam compared to the direction of travel of the vehicle.
To get back to the original topic, my GPS shows my Vauxhall speedo to be reading 3 miles/hr fast at every speed betwen 30 and 70. It's an electronic speedo so that looks like a 'designed in' error to me.
JS
Edited by John S on 01/11/2008 at 14:34
|
>>No attempt to show a reading to an order of magnitude better than the measurement point.
Yes, I think that's the important information.
|
|
|
|
There's no such thing as a completely accurate instrument. There is always some error.
;-) What instrument do you use to prove that?
|
I believe in TomTom. At an indicated 80, my mums car shows 78, my misses car 77 and my car 73. Infact my car, to get the satnav to read 80 has to be nudging an indicated 90 - 80 isnt that fast but psychologically, I cant bear to stay at that indicated speed even though I know its not real!
I like an over-reading speedo, it makes it far easier to not creep over the limit.
My dads Impreza Turbo ( 96 P reg), first day he got it was discovered to have a hugely under-reading speedo - he took it for a blast down the bypass and it felt like it was going to take off at an indicated 50 mph - turned out when I subsequently followed him in my car that his speedo was reading nearly 50% lower than reality - of course Subaru changed it immediately with much kissing of backsides.
|
|
|
>>;-) What instrument do you use to prove that?
To demonstrate a measurement system, you would use many other instruments, each with their own calibrated and traceable errors.
Considering measuring speed by considering the doppler shift of reflected laser light, one might begin by measuring the noise contributions of the instrument - for one example what is the frequency error in the transmitted light. This measurement effectively boils down to the measurement of time, and one would need to trace back to the international standards - which of themsleves, always include and quote an error.
Only when all the possible noise contributions are considered can the realistic sensitivity of the instrument be estimated.
Phrased another way, an engineering measurement quoted without a traceable estimate of the error isn't useful and isn't a particularly scientific measurement.
"Mr Cruncher, your speed was measured at 46mph with a traceable measurement error of less than 0.1mph while travelling in a 40mph zone - you're nicked!" would be a more technically acceptable way of presenting the data.
|
"Mr Cruncher your speed was measured at 46mph with a traceable measurement error of less than 0.1mph while travelling in a 40mph zone - you're nicked!" would be a more technically acceptable way of presenting the data.
When I received a speeding ticket here in Germany, it said something along the lines of:
Measured speed = 43km/h
Amount deducted for possible measurement error = 3km/h (might have been 5, can't remember)
Amount deducted for possible instrument error = 10%
Actual measured speed = 36km/h -> therefore you're nicked (limit was 30)
|
|
|
|
|
|
|
|
|
