Difference in indicated speed on SatNav / speedo

skibum

Just got my new toy, one of the settings show's current speed. While trying it out I was doing an indicated 120 on my speedo and the gps was showing approx 112.
Just how accurate are gps systems regarding current speed??

unnameduser

Speedos usually show a higher speed that you are actually doing (given that tyre pressures etc are correct)

I would imagine that GPS calculates speed over small intervals.

IMO real speed would be somewhere in between both your readings

MercMad

Yeah the car speedo definitely will overestimate, but the GPS whilst more accurate will not update instantly, it may update every 10 seconds or something so if you drove over a period of 2 minutes it should be very accurate IMO.

Alun

Opinions vary, but you're talking 1/10ths of a km/hr. Thing is, that despite the inherent innacuracies in the calculated GPS position, these inaccuracies (due to upper atmospheric conditions etc.) don't really vary significantly over short distances, so calculating speed by comparing two positions taken very closely together cancels out any error.

On my car (Focus C-Max) the difference between GPS speed and the speed displayed on the speedo is about 5%, although interestingly enough, if I set my car's display into diagnostic mode where I can view my speed digitally it agrees almost exactly (to within 1% or so) with the GPS display.

Alun

MercMad wrote:

Yeah the car speedo definitely will overestimate, but the GPS whilst more accurate will not update instantly, it may update every 10 seconds or something so if you drove over a period of 2 minutes it should be very accurate IMO.

Speed on most GPS's is calculated and displayed every second.

v10

skibum wrote:

...Just how accurate are gps systems regarding current speed??

Yeah I was curious about this too .. I've had my portable GPS unit in 3 different cars and it is over by almost the same amount on every car.

I've heard people say the GPS is more accurate but I'm starting to doubt it.

If there was a bigger variation between the cars I might agree the cars are reading wrong but the fact is that if I drive at 120kph(GPS Speed) in any of the 3 cars they all show somewhere between 111kph and 113kph. I just find it hard to believe that all the cars are reading wrong by almost exactly the same amount !

Alun

v10 wrote:

If there was a bigger variation between the cars I might agree the cars are reading wrong but the fact is that if I drive at 120kph(GPS Speed) in any of the 3 cars they all show somewhere between 111kph and 113kph. I just find it hard to believe that all the cars are reading wrong by almost exactly the same amount !

Well, the days are long gone when speedometers were purely electromechanical devices with cables and induction discs. These days we're talking electronic sensors and intelligent instrument clusters with software in them, so they can basically do anything with the raw data that's presented to them before it's displayed to the user.

Take my example above. It appears that the instrument cluster is actually capable of measuring the road speed quite accurately (I believe my GPS!), but somewhere in the chain before it's presented to the driver on the analogue display an 'error' of 5% creeps in. Personally I don't think that's due to any inherent inaccuracy in the speedometer mechanism itself (these things are usually controlled by stepper motors these days), but has been deliberately programmed into the instrument cluster firmware to comply with the EU regulations concerning over/under reading of speedometers, but I can't prove that, of course.

EDIT: Here's an interesting article on what goes into designing a modern cars' instrument cluster ...

http://www.techonline.com/community/ed_resource/feature_article/21190

dubstub

Alun wrote:

On my car (Focus C-Max) the difference between GPS speed and the speed displayed on the speedo is about 5%, although interestingly enough, if I set my car's display into diagnostic mode where I can view my speed digitally it agrees almost exactly (to within 1% or so) with the GPS display.

I'm renting a C-Max this weekend and would be interested in seeing this difference. How do you put it into diagnostic mode?

Alun

dubstub wrote:

I'm renting a C-Max this weekend and would be interested in seeing this difference. How do you put it into diagnostic mode?

Hold down the button on the end of the indicator stalk whilst switching on the ignition and keep it held in until the display reads TEST (a few seconds). You can then cycle through various tests and displays by pressing the button repeatedly until it displays ROAD SPEED KM/H. Then just start the engine and it will continue to display the speed.

EDIT: That's assuming it's got the trip computer, if not press the trip reset button instead.

dubstub

Thanks, will check that out. Will be on Italian motorways so maybe it will make the lunatic speeds seem much more reasonable!

Eoin

Alun wrote:

Well, the days are long gone when speedometers were purely electromechanical devices with cables and induction discs. These days we're talking electronic sensors and intelligent instrument clusters with software in them, so they can basically do anything with the raw data that's presented to them before it's displayed to the user.

Unless you have the digital speedo (like the yaris?), you still have a skinny plastic needle that could probably be misaligned quite easily, regardless of how up-to-date the mechanics behind it are. That's why I was pleasantly surprised to find that my GPS and speedo show exactly the same speed.

Alun

eoin_s wrote:

Unless you have the digital speedo (like the yaris?), you still have a skinny plastic needle that could probably be misaligned quite easily, regardless of how up-to-date the mechanics behind it are. That's why I was pleasantly surprised to find that my GPS and speedo show exactly the same speed.

Well, it's not that "skinny" and I don't see what it being made of plastic has got to do with it to be honest.

In any case, when I'm doing exactly 0 km/h, the needle indicates 0 km/h as it should (there's no mecahnical end stop like on the old mechanical speedos), and the whole shebang is controlled by a precise stepper motor, all of which kind of detroys the 'misalignment' theory IMO.

My wife has a Yaris, so if I ever have time I'll do the same check on her car.

leeroybrown

Unless the speed reading is calibrated regularly or measured using a reliable external factor then there is bound to be a degree of error and variance.

Alun

Found this on a VW forum, which kind of reinforces the idea that the error is deliberately built in to the cluster firmware ...

Quote, originally posted by nasos »
I checked recently my speed with a GPS and discovered that my car has a permanent error indicating the speed in analogic speedometer. Climatronic and VAG-COM indications match the true GPS speed?

Nasos:

We raised this same question in the Phaeton forum a few months ago, and actually discussed it with the engineering people from the Phaeton assembly plant in Dresden.

They explained to us that under European law, the speedometer must be set up so that it never - under any circumstances - reads less than the actual speed of the vehicle. It is, however, permitted that the speedometer over-reads somewhat.

In the case of your car, the analog speedometer needle probably has a deliberate error built into it, to allow for the possibility of someone installing different size tires (larger diameter tires) onto the car. This is the case with the Phaeton, which has the same characteristics you describe: The digital speeds shown in the ABS and engine controller MVBs (which is where your Climatronic display is getting the information from) are generally dead-on accurate, but the speed shown by the pointer on the round dial is overstated.

The Phaeton engineers told us that they calibrate each car (via an adaptation adjustment) so that the digital speeds are exactly correct for the tire size fitted at the factory. This ensures that the odometer readings are accurate, and the ABS and ESP systems are using accurate speeds. The needle always reads about 7% higher than these 'calibrated digital speeds', though, and there is no possible way to correct the needle only, meaning, to match the needle speed up with the digital speed.

You could, in theory, adjust the correction factor for the digital speeds so that the needle speed was accurate, but if you did that, the digital speeds would under-read by the same amount that your needle speed currently over-reads.

Hope this makes sense.

Michael

Eoin

Alun wrote:

Well, it's not that "skinny" and I don't see what it being made of plastic has got to do with it to be honest.

In any case, when I'm doing exactly 0 km/h, the needle indicates 0 km/h as it should (there's no mecahnical end stop like on the old mechanical speedos), and the whole shebang is controlled by a precise stepper motor, all of which kind of detroys the 'misalignment' theory IMO.

My wife has a Yaris, so if I ever have time I'll do the same check on her car.

It wasn't a "theory" to be "destroyed", just a casual comment. It was just that the speedo struck me as something that could be quite easily misaligned. But clearly not.

MercMad

Interesting !!

One further thing is that there would be a distinction between built in GPS units, which are most likley more accurate than the prtable type I was refering to !

v10

MercMad wrote:

Interesting !!

One further thing is that there would be a distinction between built in GPS units, which are most likley more accurate than the prtable type I was refering to !

Can't see how they'd be more accurate .. they all get the information from the same place

Eoin

v10 wrote:

Can't see how they'd be more accurate .. they all get the information from the same place

I wonder if the built in units have better antennas, and if they receive a signal on a more regular basis - could have a bearing on their accuracy?

Alun

leeroybrown wrote:

Unless the speed reading is calibrated regularly or measured using a reliable external factor then there is bound to be a degree of error and variance.

Well, in the case of a modern car with what is essentially a completely digital system of measuring speed (even though it's ultimately displayed on an analogue display) the only really significant variable in the whole equation is the rolling radius of the tyres on the driven wheels, but I'm not sure that is going to account for the magnitudes of errors we're seeing here. Of course, if you fit after-market wheels and/or tyres of a different size than what the speedo is calibrated for you'll mess everything up, but that isn't what we're talking about here.

Take my car for example, the tyres are 205/55 R 16's and according to http://www.carbibles.com/tyre_bible.html these have a rolling radius of 315.95mm. Now, a difference in tread depth of 2.5mm is only going to account for a difference in radius (and therefore circumference and speed) of 0.8%. I've not been able to find any source to tell me what effect on rolling radius other factors, such as load or inflation have, but I'd be surprised if they were much more significant than that.

conor_mc

v10 wrote:

Can't see how they'd be more accurate .. they all get the information from the same place

I assume the same GPS principles apply on the ground as in the air, in which case many light aircraft pilots will disagree with you.

The reason is that built-in units usually have external aerials and therefore are more likely to receive sufficient signal strength from several satellites to allow a more accurate identification of position.

Hand-held units, in similar circumstances, may only be receiving from 2 or 3 of the several available sats, and so the margin of error in plotting position increases.

EDIT - post crossed with eoin_s.

Alun

eoin_s wrote:

I wonder if the built in units have better antennas, and if they receive a signal on a more regular basis - could have a bearing on their accuracy?

Built in units generally have external antennae and may therefore get better coverage, i.e. more satellites in view and therefore a more accurate position fix. That wouldn't necessarily equate to a more accurate speed measurement though.

v10

eoin_s wrote:

I wonder if the built in units have better antennas, and if they receive a signal on a more regular basis - could have a bearing on their accuracy?

I don't think so as these units get their information from several satelites simultaneously .. i.e. I can see (by pressing a button) on mine that is currently getting signal from 10 satelites. The others would be the same. In fact I've read elsewhere that the portable 3rd Party one are better than the in car ones.

The only thing I will say is that the one that use a bluetooth antenna seem to have a bit of lag in them. For example .. I have a Nokia 6680 also with Route 66 software and a bluetooth GPS receiver and if I use both this and my Portable GPS unit at the same time the info comes about a second later on the phone than the other unit.

Alun

v10 wrote:

I don't think so as these units get their information from several satelites simultaneously .. i.e. I can see (by pressing a button) on mine that is currently getting signal from 10 satelites.

Yes, but a hand held device inside a car can have it's internal antenna obscured by parts of the car, or muted by things like coated or heated windscreens. A roof mounted antenna can have a perfectly clear view of the sky.

v10

Alun wrote:

a hand held device inside a car can have it's internal antenna obscured

I guess this possible .. but then if this is true then I should see a difference if I held it just inside my open sunroof .. I must try this :rolleyes:

Alun

v10 wrote:

I guess this possible .. but then if this is true then I should see a difference if I held it just inside my open sunroof .. I must try this :rolleyes:

What's with the :rolleyes:? Just stating a fact.

Anyway, are you sure your unit is actually receiving a signal from 10 satellites inside your car? Often the display will show the current 'constellation', i.e. the satellites that the unit could theoretically be able to see if it had an unimpeded view of the sky. It only needs a usable signal from one satellite in order to get this information. It will then usually colour code each 'blob' on the display to indicate which satellite it is actually receiving a valid signal from. Depending on the date and time I'll sometimes have difficulty getting 10 satellites on my handheld GPS on top of a mountain, let alone inside a car in an urban environment.

v10

Alun wrote:

Depending on the date and time I'll sometimes have difficulty getting 10 satellites on my handheld GPS on top of a mountain, let alone inside a car in an urban environment.

To be honest it's usually between 6 and 8 but I have seen 10.

unkel

On a long empty stretch of M4 last night, I set my cruise to an indicated 130km/h. I watched my GPS for about 15 minutes and it was pretty stable in indicating 125km/h. I got the data from 7-9 satellites during this stretch. This is a 4% overstatement of speed

BTW, I've just calculated that with my tyres, when still above the legal minimal tread, the speedo would overstate speed by just under 2% compared to brand new tyres

Calculation: radius of 0.344m with full 8mm tread. When the tread is down to 2mm, the radius is reduced by 6mm to 0.338m. This decreases circumference from 2.16m to 2.12m, or 2%

My non standard tyres have nearly identical circumference (2.16m) as the standard ones (255/55 18" vs 235/60 16")

Legally speedos are required to read anywhere between 0% and 10% over the actual speed iirc

Alun wrote:

On my car (Focus C-Max) the difference between GPS speed and the speed displayed on the speedo is about 5%, although interestingly enough, if I set my car's display into diagnostic mode where I can view my speed digitally it agrees almost exactly (to within 1% or so) with the GPS display.

Wow, great find

BreadBoard

My GPS system shows that I'm doing about 10kph more than the car speedo...

unkel

BreadBoard wrote:

My GPS system shows that I'm doing about 10kph more than the car speedo...

Do you have large non standard wheels?

BreadBoard

unkel wrote:

Do you have large non standard wheels?

No they are standard. 16" Nissan alloys on a 00 Nissan Primera.

BreadBoard

I took a pic ;