General Whom to believe

[Eklipze3k]

OK, so on the way back from Dundee I was overtaking a few trucks on the A9, my speedo was saying I was doing 100, but my GPS was saying 85ish. Now I understand that the speedo becomes less accurate the further up it goes but after I spotted this I started checking it more often.

It seemed to correlate to the GPS unit up to about 35mph, by the time my speedo said I was doing 60 the GPS reckoned it was about 52 and at 80mph on the speedo, the GPS read that as almost exactly 70mph.

Which is liable to be more accurate? Should I split the difference, or always err on the side of caution and go by what the speedo says as it was always underestimate, giving me less chance of being caught speeding ...

 

[5abr3]

you'd most likely be safest going with the cautious side, theoretically tyre wear will affect your speedo's reading (though only slightly) whereas GPS is reading relative to geo-stationary objects, but still isn't dead on if I remember correctly

 

[GoBe]

Car speedos are designed to read high at speed. GPSs however if in error give a lower than actual speed.

 

[jimbro1000]

Ironically GPS becomes more accurate the faster you go just because the error in position reading becomes smaller in relation to the size of the change in position per computation cycle.

If you are standing still the GPS is only (as standard) accurate to within 30 feet and even then only if you are getting good reception. The american military version of the receiver is accurate to within a few inches and for some reason they don't like handing out the more accurate version. Some bright sparks at a UK university (I think it was Cambridge) came up with a refined algorithm that is as accurate as the military one and it has made its way into some receivers but I don't know which ones.

Some receivers also use two aerials (supposedly a meter apart) that significantly increase accuracy and also allow the receiver to work out which way you are facing without the need for forward motion.

The system I use is an oldish Garmin and it has proven to be very accurate, contrary to what the authorities would like us to believe. Some cars are however also accurate - my Sei reads the same as my GPS to within 1 or 2 mph, whereas my old X1/9 typically read over by about 27% (yes you did read that right). Not all cars have a linear increase in speed error though - the previous generation Renault Espace we had was 20% inaccurate up to about 20mph, then over-read by 4mph until you got to about 50mph and then started the error started going up again in proportion. The only conclusion there is that the error was deliberate.

 

[Italian Coke Can]

My GPS reads 101mph @ an indicated 120mph.

It's worse with low profile tyres.

At normal speeds it's reading 10 mph, but >70mph it goes off more.

However, most cars over read by 5mph anyhow.

 

[StoneNewt]

The thing is though the positional error is quasi-systematic so the effect on point to point speed is minimal, what has a far larger effect on normal sat-nav systems is the lack of elevation position & the huge sample period.

I could bore you with details but basically I have a high accuracy GPS speedo & trip which is always consistent with a mechanical speedo when calibrated properly. I've found that sat navs seem to read typically 1 or 2mph low & that can increase up to 8% on corners with steep inclines (ie reading 65mph at 70mph).

 

[levemircento]

the trucks would have only been doing a max of 60mph, they are governed to 56mph but usually nudge a tad over, you would only need to be doing 65 to pass them easily.

I dont think there are many car speedos of the mechanical type that are accurate

 

[blu73]

you running standard wheels and tyre size?? Mine reads almost exactly the same as the speedo at most speeds but i got 15's so speedo should be way off standard measurements..

 

[Eklipze3k]

you running standard wheels and tyre size?? Mine reads almost exactly the same as the speedo at most speeds but i got 15's so speedo should be way off standard measurements..

Yep, everything standard.

Thanks for all the info guys and gals, I guess it's always gonna be better to err on the side of caution when it comes to speed.

 

[turbo_jim]

The thing is though the positional error is quasi-systematic so the effect on point to point speed is minimal, what has a far larger effect on normal sat-nav systems is the lack of elevation position & the huge sample period.

I was under the impression that GPS receiving signal from 3 (or more) satellites would give you an altitude as well as an x-y position. In which case the sat-nav could calculate speed in whichever plane the vehicle is travelling. Would be interested to find out more if i'm wrong. I work for a digital mapping company, and our equipment always gives an altitude readout.

 

[StoneNewt]

I was under the impression that GPS receiving signal from 3 (or more) satellites would give you an altitude as well as an x-y position. In which case the sat-nav could calculate speed in whichever plane the vehicle is travelling. Would be interested to find out more if i'm wrong. I work for a digital mapping company, and our equipment always gives an altitude readout.

It's possible to extract altitude data with 3 or more locks but GPS systems don't actually extract or use this data from what I understand. Also I think for a reliable x, y, z location you need 5 or more locks. Remember that sat nav GPS systems use typically the cheapest setup that does the job.