Slightly OT? Speedo accuracy vs GPS

I've often wondered about the accuracy of speedos in cars. People throw around a +-10% figure all the time although I've no idea whether really that figure is correct.

Anyway, I've got a 1998 S reg Pug 306, this appears to have a speedo (and odometer/revs) display which is controlled by the ECU (as demonstrated when I had an earth fault and it showed me doing 0-140mph in a few seconds when I was actually just crawling). As a result, would ECU controlled speedos be more or less accurate than a cable controlled one?

I've recently got a TomTom Navigator setup for my PDA including the TomTom Bluetooth GPS mkII unit. This all works well and can display my current speed.

At around 50+mph I've noticed that my speedo appears to consistently under read by about 5mph. Speedo says 50mph, TomTom says 45. Speedo says 70, TomTom says 65. As a result, I've now got in the habit of trusting TomTom and using that to judge my speed, consequently I'm going a little faster than I might have done in the past (speedo reading 55mph when going through 50mph roadworks, but TomTom saying 50).

Are GPS units that accurate? I know they're not terribly precise in a spot reading (few metres) but I would guess that with movement, especially at higher speeds, their 1+m accuracy becomes less important as it's the change in reading which is being used.

Am I being foolish to use my GPS to judge my speed, especially with regard to speed traps on motorways, especially if I might push it towards the 10% + 3mph principle?

Does a constant 5mph over reading by a speedo sound right? I don't notice it so much at lower speeds, although there is still a slight over reading.

One final thing, if my odometer say I've travelled 10 miles, and TomTom says I've travelled 10 miles then the car must be able to accurately measure the distance travelled. Assuming a reliable timer/clock in the ECU, why would the speedo under read, unless it was programmed to do that by the manufacturer? I've often wondered whether speedos are actually designed to under read, especially with ECU controlled speedos.

Any thoughts?

Thanks

D
Reply to
David Hearn
Loading thread data ...

Yes, by all accounts it's the GPS that is the accurate reading. Like yourself I now travel a little faster than the limit according to my speedo as I'm aware of it's inaccuracies.

That's fine but I'd stay away from very top of that band. Besides they don't

*have* to apply 10%+3mph.

Sounds about right, but at the top end. At 30mph my ground speed is 27mph,

55mph is 50mph, and 80mph is 74mph. I have to hit almost 90 on the speedo to get to 80mph.

Speedos are always slightly inaccurate so manufacturers do set them to slightly underread, as they are not permitted any margin for overreading.

Reply to
Mark Hewitt

The gps is most accurate on level roads with a clear sky view. I would trust the gps. Consistently every car speedo I have checked overreads according to either of my gps units, which both agree on speed.

10 mph over at a real 60 is quite common, although 5 - 7 is the most frequent, I have yet to find one that is accurate or underreads.

Locally we have one of those light up signs that warns if you are over thirty, the gps thirty is ok !!

There used to be a few light up signs that showed your speed, but the kids used to see who could get the highest reading.

mrcheerful

Reply to
mrcheerful

yes that's about what my GPS says normally, although you have to remember that GPS altitude measurments are not so accurate, I'm not sure how this effects the GPS speed reading if you are going down/up hill (or round a corner ?), I assume the gps is only going to be measuring horizontal speed between two readings rather than 'ground speed', if you see what I mean. I suppose it also depends on how many readings the GPS makes per minute, and how clever the algorithm is etc. I suspect they are quite capable though, I have measured speeds of >

500MPH in airplanes !
Reply to
mr p

Yes.

And it's -10% +0%.

In normal use they both will fall in the +4kph/-10% bracket. Exactly where depends on the manufacturer. In both cases, the accuracy of the final meter if analog in its cheapest form, may not be significantly better than 10%. If digital, it's going to depend on the settings.

In short - they don't spend any money to get it accurate. If they can shave 1p off the cost of the guages by having it read between 5% and 10% low, rather than 2.5% low, then over a million cars, they've just saved 10000 pounds.

Reply to
Ian Stirling

There are still some around here and they agree with my GPS. Funny because I think the idea of them backfires, you see the sign and think "oh I'm going slower than I thought, I'll speed up"

Reply to
Mark Hewitt

I suspect most may well ignore altitude completely. Going up and down a hill does affect the readings in my car, so only trust a reading on level ground. The best way to do it is not to rely on the sat nav all the time but to learn how the reading on your speedo relates to your true ground speed, then you won't need the GPS again.

Consumer GPS receivers typically report their position once per second.

Reply to
Mark Hewitt

Good point about ground speed, possibly ignoring changes in height. I'll bear that in mind.

D
Reply to
David Hearn

David Hearn ( snipped-for-privacy@NOswampieSPAM.org.uk) gurgled happily, sounding much like they were saying :

Yes.

Yes.

Yes.

Reply to
Adrian

David Hearn ( snipped-for-privacy@NOswampieSPAM.org.uk) gurgled happily, sounding much like they were saying :

I think you may benefit from reviewing basic trigonometry. It's incredibly unlikely to make any kind of realistic difference on any gradient that you'll encounter.

Reply to
Adrian

Unlikely yes, but still something to consider if you're already pushing things to the limit.

D
Reply to
David Hearn

Maths says no. Personal experience says it does.

Reply to
Mark Hewitt

Well, if you're going along a 1:5 gradient at 70MPH, then the gradient'll have a much greater impact on your speed than the 4% or so change in horizontal speed due to that gradient.

Reply to
Ian Stirling

No No No! The GPS uses doppler shift (of the signal from the satellites) to measure speed. It is nothing to do with the frequency of postion readings.

And it is Affect, not Effect! The former is a verb, the latter is a noun.

Tony

Reply to
Tony Brett

It is.

Nope.

Judging by the fact that I seem to always end up doing 45 through 50MPH signed roadworks, being held up by cars, and my speedo can only be 2% out and has calibration checks every two years, sounds about right.

Yep.

Nope.

Reply to
Conor

not using the frequency of taking readings, check.

Affect not effect, check.

GPS uses doppler to measure speed? well..... it varies.

It actually uses a combination of deltas between absolute positions, and an almost instantaneous speed reading obtained by averaging a few doppler measurements.

It's pretty hard to find out what your particular GPS receiver uses to calculate the speed it shows you. Is it 2-D or 3-D speed? along your motion vector or on the nominal surface? All doppler based, or weighted by longer term average, or all long term?

My TomTom in the car will quite happily be 10+mph out for seconds at a time on a long steady corner.

Reply to
PC Paul

Absolute bollocks. The calculation involves removing the relativistic effects due to satellite motion which also destroys any useful Doppler measurement.

It uses the time difference between (x1,y1,z1) and (x2,y2,z2) - which for absolute accuracy needs to be derived from AT LEAST 4 satellites.

Any less and the % error increases dramatically.

Slurp

Reply to
Slurp

There used to be one out on a B road on the Herts / Essex border - there was a fairly tight bend with, ISTR, a 30mph advisory on it and a flashing speed sign in the braking zone.... of course, when out on a ride with fellow bikers, there was in no way some sort of challenge to see who could take said bend at twice the advisory limit ;-)

Reply to
SteveH

Bzzt. Try again.

How many detailed technical links would you like to convince you?

Reply to
PC Paul

And another one.

Reply to
PC Paul

MotorsForum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.