Speedo Inaccuracies 2014 Yaris

[Mick F]

I was out today and took my Garmin Montana.  It's a hand-held device which has multiple uses.  I mainly use it when cycling.

https://buy.garmin.com/en-GB/GB/p/523640

I did three trials at an indicated 40mph, 50mph and 70mph.  I tried to do a 30mph but it wasn't very easy due to traffic etc.  I didn't want to do 30mph on the open road because I would have held the traffic up.

I did a total of 52.92 miles as recorded on my Montana, and the car said 52.9 miles.  Spot on, I reckon.  However, the speedo isn't very good at all.

At an indicated 40mph, my Montana recorded 36.8mph = 3.2mph low.

At 50mph, it was 45.6mph = 4.4mph low.

At 70mph on the dual carriageway, the Montana read 65.2mph = 4.8mph low.

Attached are the graphs I extracted from the raw GPX file using an activity program called Ascent.  The blue bits are the elevation profiles.

Hope people find this interesting.

Mick.

[FROSTYBALLS]

So assuming the Montana is near to the actual speed you were travelling, the speedometer is within legal limits - ie reading 110% of the actual speed, and for speeds between 25 and 70mph, 110% plus 6.25mph.

[Mick F]

Oh yes, I realise it's within legal limits.  I wasn't saying it wasn't, but it was surprising to me that if they can get the distance bang on, they should be able to manage better than circa 7% low.

The legal limits were there from the early days of rotating Bowden cables and a speedo head and magnets moving a pointer.  These days it's all electronic and fed from the onboard computer ............... just like the dead on accurate odometer.

Never checked the speedo of a car before, so maybe - just maybe - that car speedo accuracy varies from car to car. If I'm doing spot on (according to my speedo) a constant speed of 40mph in a 40mph limit, but I'm actually doing 36.8mph, it could be the car behind who's speedo is more accurate than mine would think I'm dawdling. 

Next test for me, is to find what the speedo says at true speeds.  Looks like the all important 30mph will be more like 32 or 33.  Considering the police don't prosecute for a couple of MPH over the 30mph, I could actually aim at a reading of 35mph.

Mick.

[Heidfirst]

Legally they are not allowed to read high so they allow a tolerance to keep them in the clear, 6-7% low is perfectly normal for a Toyota in my experience. Other manufacturers similarly will read low by varying amounts according to manufacturer

This figure will also vary slightly as your tyres wear down if you have an accurate enough method of measuring.

[Heidfirst]

2 minutes ago, Mick F said:

Never checked the speedo of a car before, so maybe - just maybe - that car speedo accuracy varies from car to car. If I'm doing spot on (according to my speedo) a constant speed of 40mph in a 40mph limit, but I'm actually doing 36.8mph, it could be the car behind who's speedo is more accurate than mine would think I'm dawdling. 

Next test for me, is to find what the speedo says at true speeds.  Looks like the all important 30mph will be more like 32 or 33.  Considering the police don't prosecute for a couple of MPH over the 30mph, I could actually aim at a reading of 35mph.

Trucks tend to have more accurate speedos hence why sometimes it may look like they are breaching the limit.

& yes, knowing the speedo inaccuracy I incorporate it in my driving ...

[Mick F]

As the tyres wear, the odometer would be wrong .............. but it isn't.

How can trucks have more accurate speedos?  If they can, why can't cars?  Perhaps there's accuracy out there, but it isn't with a Toyota.

I really do wish I'd checked previous cars we've owned.

Mick.

[monkeydave]

yeah mine does 37 when displaying 40 when going past the speed signs that tell you your speed

[FROSTYBALLS]

At the end of the day there is no pressure or requirement for manufacturers to install more accurate speedometers. Those currently fitted meet legislation, be that of UNECE, the EU or the UK (whose requirements differ from the EU), which is all that manufacturers want.

[Heidfirst]

1 hour ago, Mick F said:

As the tyres wear, the odometer would be wrong .............. but it isn't.

The difference is minimal (something like 1%) but it will exist. Even increasing/dropping tyre pressures by e.g. 5 psi will vary the circumference of the tyre & again  be measurable if your equipment is accurate enough.

How can trucks have more accurate speedos?  If they can, why can't cars?  Perhaps there's accuracy out there, but it isn't with a Toyota.

Because trucks cost a lot more, there is more pressure on/cost re. driving time & legally they must be fitted with tachometers which require regular calibration & by EU regulations a tachograph must be accurate to plus or minus 6 km/h so it's an area that truck manufacturers & owners will both pay more attention to. No car manufacturer is going to give you total accuracy in the UK but will all build in a + margin of error. If you were charged with a speeding offence, found guilty & the reason was an inaccurate speedo then they would leave themselves open to potential prosecution & law suit.

 

[Mick F]

Having slept on this, I really think that car manufacturers have (at least) two issues.

1. It may be well indeed that it keeps costs down.  I don't personally see this, as a zero cost tweak in the firmware, the speedo could be nearer ......... still low, but nearer.

2. This one is more likely I reckon.  Cars are seen as exciting fashionable things, they want them to be desirable, and most of all, they sell speed.  By having the speedo reading high, the car appears faster.  Why have a speedo that reads up to 140mph in a Yaris?  Because they was to appear exotic.

Mick.

[Mike J.]

18 hours ago, Mick F said:

.... but it was surprising to me that if they can get the distance bang on, they should be able to manage better than circa 7% low.

As mentioned before, our cars have an ECU based speed, and thus distance, which is 'spot on', and the deliberately over-reading speedo that is probably ECU controlled (I doubt we have a cable connection from the gearbox).

Also, as mentioned before, you can 'see' the ECU derived speed via the "Glass of Water" app that one can download and install on the car via USB. I just did a quick check using this app and cruise control and got these figures:

30 mph speedo was 28 mph app

40 mph speedo was 37/38 mph app

50 mph speedo was 47 mph app

60 mph speedo was 57 mph app.

 

Additionally, each flick of the cruise control stick changed the ECU speed by 1 mph.

 

[Mick F]

5 hours ago, Mike J. said:

30 mph speedo was 28 mph app

40 mph speedo was 37/38 mph app

50 mph speedo was 47 mph app

60 mph speedo was 57 mph app.

Not quite correct.

You're showing a standard 3mph below true.

My readings by accurate GPS, gave a circa 7% error, though the distance over 50odd miles was dead on.  40mph = 36.8  50mph = 45.6  70mph = 65.2

I too used cruise control.

There's more to this than ECU outputs.  I reckon that the it's the firmware of the speedo itself as well.  Funny that the distance is correct but the speed isn't.

Mick.

 

[Heidfirst]

15 minutes ago, Mick F said:

 Funny that the distance is correct but the speed isn't.

believe it or not that is not unusual (& not just in Toyotas) . Ex factory the speedos are deliberately set fast but the odometers are usually pretty much spot on.

[Mike J.]

27 minutes ago, Mick F said:

Not quite correct.

You're showing a standard 3mph below true.

I don't understand your comments. I believe the app is using ECU data and is fully accurate subject to tyre wear only - probably counts MG2 rotations. This ECU based speed is also seen via Torque Pro, but you need more equipment to 'see' it.

[Mick F]

2 minutes ago, Mike J. said:

I don't understand your comments. I believe the app is using ECU data and is fully accurate subject to tyre wear only - probably counts MG2 rotations. This ECU based speed is also seen via Torque Pro, but you need more equipment to 'see' it.

My comment is that you are showing a 3mph error which means that at at an indicated 70mph, the "true" speed would be 67mph.  Maybe your car at an indicated 100mph would be really be doing 97mph?

I'm saying that the true speed at an indicated 70mph is 65.2mph in my Yaris, despite what you say your ECU may be saying in your car.

Mick.

[Mike J.]

1 hour ago, Mick F said:

I'm saying that the true speed at an indicated 70mph is 65.2mph in my Yaris, despite what you say your ECU may be saying in your car.

I have used Torque Pro which shows that GPS and ECU speeds are very close. I would say that ECU has a greater chance of being correct than GPS especially in instantaneous speed mode. You showed that the ECU can work out distances correctly, so it must work out speed correctly - and this is shown by the app.

[Mick F]

My GPS device and the computer program together work out the true speed over the ground.  My Garmin Montana is accurate to 7ft or less for every data point.  It doesn't record instantaneous speed, but calculates it from each and every data point.  No GPS measures speed, just records lat/long position.

I cannot comment on the ECU speeds, only on the data that I can see.

I don't agree that an ECU that measures distance correctly must also measure speed correctly.  It may well do ........ but it's not a "must".

Whatever the issue here, it's a question of accurate GPS positioning versus accurate speed from the ECU, but whichever is correct, the readout on the speedo is wrong.  My issue with your ECU information is that the error is linear.  3mph at all speeds.  That's a very odd situation for errors, because a percentage error would be the norm.

I say it's circa negative 7% error .................. not a flat negative 3mph across the board.

We'll agree to differ.

Mick.

[Mick F]

PS:

Just thinking.  If a mile takes 60 seconds at 60mph, I wonder how long my Yaris would take to do a mile with my GPS reading 60mph?

I say that the distance is correct but the ECU speed is incorrect.  You say that an indicated 70mph would be 67mph (minus 3mph), when I say 70mph would be 65.2mph (minus 7%) therefore we should have some arbitration here regarding a stopwatch? 

Give me a few days, and I'll check this out.

Mick.

[Mike J.]

31 minutes ago, Mick F said:

Whatever the issue here, it's a question of accurate GPS positioning versus accurate speed from the ECU, but whichever is correct, the readout on the speedo is wrong.  My issue with your ECU information is that the error is linear.  3mph at all speeds.  That's a very odd situation for errors, because a percentage error would be the norm.

To me, if the errors were a fixed percentage, then the speedo would be cable driven. As it isn't (99% certain speedo cables disappeared years ago), it is software driven.

[Heidfirst]

I don't know if you can do it on a Toyota with access to the full software but on some VWs you can change the % error for the speedo via software. e.g. https://www.my-gti.com/140/volkswagen-speedometer-error-correction

[Mike J.]

On ‎23‎/‎01‎/‎2018 at 12:14 PM, Mike J. said:

Also, as mentioned before, you can 'see' the ECU derived speed via the "Glass of Water" app that one can download and install on the car via USB. I just did a quick check using this app and cruise control and got these figures:

30 mph speedo was 28 mph app

40 mph speedo was 37/38 mph app

50 mph speedo was 47 mph app

60 mph speedo was 57 mph app.

I have finally got around to the 70mph figure:

70 mph speedo was 66 mph app.

[Rostronic]

All speedometers read fast. There is a limited number of accurate speedometers available 'aftermarket' which are fitted to police cars but are very expensive. If you want accuracy use GPS/SatNav it's the safest way with the speed warning alarm turned on, gives a beep at 3 mph over the limit.

[Mike J.]

35 minutes ago, Rostronic said:

All speedometers read fast.

We know that. This thread is trying to determine the inaccuracy. My take is that the speedo is deliberately over-reading as the ECU knows the precise speed - count rotation of MG2 and use final drive ratio and tyre size maths - but tells the speedo to do an over-read of around 3-4 mph. Note that the mileage seems to be very accurate, as it would be due to the ECU's info. When I used Torque Pro, the phone GPS and ECU speeds were very close.

[Mick F]

I've yet to do a measured mile at speedo indicated 60mph.  I'll do the same again at GPS 60mph.  Give me some time, and I'll do it.  My Garmin Montana has a stopwatch app but I'll probably take a real stopwatch.

Some years ago, I was involved in discussions on a cycling forum regarding GPS distance accuracy, and I proved mine was correct by a measured mile on the A30 West of Okehampton ............. it's down hill for quite some miles! 

200metre marker posts = five per Km = eight per mile = 60secs per mile.

https://www.google.co.uk/maps/@50.7044714,-4.0901424,3a,37.5y,221.09h,72.52t/data=!3m6!1e1!3m4!1sFyShC5Lsq2c7jQuoV7W9Bw!2e0!7i13312!8i6656

Mick.

[ratbat]

On 1/23/2018 at 12:14 PM, Mike J. said:

As mentioned before, our cars have an ECU based speed, and thus distance, which is 'spot on', and the deliberately over-reading speedo that is probably ECU controlled (I doubt we have a cable connection from the gearbox).

Also, as mentioned before, you can 'see' the ECU derived speed via the "Glass of Water" app that one can download and install on the car via USB. I just did a quick check using this app and cruise control and got these figures:

30 mph speedo was 28 mph app

40 mph speedo was 37/38 mph app

50 mph speedo was 47 mph app

60 mph speedo was 57 mph app.

 

Additionally, each flick of the cruise control stick changed the ECU speed by 1 mph.

 

Bit late to the party, but the water app uses GPS for speed not the ECU.