Speedo Inaccuracies 2014 Yaris

[Mike J.]

28 minutes ago, ratbat said:

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

Previously, I asked if someone without navigation (and thus without GPS) could run the app - result: <tumbleweed>. 

If they could, then it would be ECU speed. Have you any links to say why it is GPS speed?

[Steve Whits]

My Auris has the most inaccurate speedo I've ever had (verified by my dealer at the last service) but because it's within the 10% limit I'm stuck with it.  The Yaris is almost spot on. Weird!

[FROSTYBALLS]

 

21 minutes ago, Steve Whits said:

My Auris has the most inaccurate speedo I've ever had (verified by my dealer at the last service) but because it's within the 10% limit I'm stuck with it. 

For the UK, which differs from the EU, the legal requirement is that a speedometer should show no more than 110% of the true speed, and for speeds between 25mph and 70mph, no more than 110% of the true speed plus 6.25mph.

So although the speedometer is almost certainly designed to comply with the EU requirement (ie to show no more than 110% of the true speed), legally in the UK speedometer accuracy has a broader requirement. 

[Steve Whits]

On 9/21/2018 at 7:50 PM, FROSTYBALLS said:

 

For the UK, which differs from the EU, the legal requirement is that a speedometer should show no more than 110% of the true speed, and for speeds between 25mph and 70mph, no more than 110% of the true speed plus 6.25mph.

So although the speedometer is almost certainly designed to comply with the EU requirement (ie to show no more than 110% of the true speed), legally in the UK speedometer accuracy has a broader requirement. 

Very frustrating though 😉

[jonquirk]

The workshop manual for the mk1 Yaris has an interesting little chart which shows the acceptable readings at various speeds:

20mph: 20-23mph

40mph: 41-44mph

60mph: 63-67mph

80mph: 84-88mph and so on. I wondered why the speedo reading differed from both my GPS and the signals available via the OBD-II port. It is annoying that Toyota appear to make the speedo deliberately over read when the correct figure is available at the diagnostic port.  

[FROSTYBALLS]

On 9/21/2018 at 7:50 PM, FROSTYBALLS said:

For the UK, which differs from the EU, the legal requirement is that a speedometer should show no more than 110% of the true speed, and for speeds between 25mph and 70mph, no more than 110% of the true speed plus 6.25mph.

Within the EU and UK, it is illegal for speedometers to show less than the true speed (under read). Legally, manufacturers have to ensure speedometers over read within the limits quoted above, and this applies to all manufacturers who sell within the the EU and UK, not just Toyota. 

What is shown by GPS and the OBD port is irrelevant.

[Mick F]

Many years ago, my mate had a Mk1 Cortina Estate.  He had fitted a replacement differential from a Mk1 Cortina but from a different model. (or was it a Corsair?)

He and his wife, and me and mine, went on a camping holiday in Northern Scotland.  We drove the same distance, but when we got home, our odometers were way out.  We then did a series of experiments driving at pre-arranged constant speeds in convoy and made notes.

Can't remember the specific numbers involved, but when I was doing 30mph in my Hillman Hunter, he was doing 25mph or something like that.  His car was seriously under-reading! 🙄

Mick.

[jonquirk]

I have always believed that speedometers were permitted to have up to +10%/-0% error which seemed reasonable for an analogue device. It also makes sense that speedometers should not under read to avoid drivers going faster than they think they are. I find it hard to believe that regulations would mandate that speedometers must over read, rather than saying they must not under read. After all, accurate, calibrated instruments are fitted to police cars; expense is the main reason they are not universally fitted.

There are factors beyond the manufacturer’s control that could lead to a speedo under reading, such as overinflated tyres on the drive axle, or non standard tyres being fitted. This could lead to them choosing to build in a margin of error that would keep the driver legal and prevent the manufacturer being drawn into a speeding prosecution.

Back to the Yaris: we have seemingly accurate figures coming from the speed sensor, and visible at the diagnostic port but a speed display configured to over read within the legal limit. Someone stated that it is possible to dig into the software settings on a VW and correct the speedo error, which makes me wonder if it is a legal requirement that they must over read, rather than not under read?

I will continue to be guided by my GPS and do what the car shows as 32 in a 30, 43 in a 40, 54 in a 50 in order to make progress where possible, sure in the knowledge that I will not be speeding, especially given the way the police usually only stop those exceeding the posted speed limit by 10%+2mph, 35 in a 30 at which speed my car would be showing 38mph.

[bathtub tom]

1 hour ago, jonquirk said:

given the way the police usually only stop those exceeding the posted speed limit by 10%+2mph

Colleague had to attend a speed awareness course for doing 31 in a 30 limit down Bournemouth way. Perhaps he failed the attitude test?

[jonquirk]

The only way to be certain of not getting stopped for speeding is not to exceed the posted limit. The 10%+2mph is only a guideline. While not exceeding the limit it would be nice to be as close to it as possible if conditions are suitable. That means having accurate speedometers. If I do an indicated 70 I am actually doing 65mph and spend most of my time in lane one on the motorway. Without that knowledge I might be tempted to spend longer in lane two if there are lorries in lane one, thinking that I have a 14mph speed differential, rather than the actual difference of 9mph (lorry tachographs, unlike car speedometers, are calibrated).

[Heidfirst]

15 minutes ago, jonquirk said:

The only way to be certain of not getting stopped for speeding is not to exceed the posted limit. The 10%+2mph is only a guideline.

For enforcement as in legislation +1mph is illegal ...

The police in Scotland have never disclosed what their guidelines are, ACPO (now NPCC) guidance only applied to England & Wales. or not. Indeed, iirc they did trial a no tolerance policy for 6-months with a warning letter for 1st infringement (assuming that it wasn't wildly blatant) & ponts & a fine for 2nd etc. . 

[Mike J.]

2 hours ago, Mick F said:

Many years ago, my mate had a Mk1 Cortina Estate.  ……… Can't remember the specific numbers involved, but when I was doing 30mph in my Hillman Hunter, he was doing 25mph or something like that.  His car was seriously under-reading! 🙄

 

From my memory of those days, the speedo specifications were plus OR minus. This changed in the 1970's. Thus an old car (pre-1970's ??) can still under-read - just like you can ride in a car without wearing a seat belt (if never originally fitted).

Regarding speedo problems, a friend had a British Leyland car with the wrong axle ratio from build time (20% under reading, IIRC). He only found out when a police car paced him - he proved the speedo was reading 70 and so was let off. BL wouldn't reimburse the garage for the replacement costs ...

[FROSTYBALLS]

5 hours ago, jonquirk said:

After all, accurate, calibrated instruments are fitted to police cars; expense is the main reason they are not universally fitted.

Not all police cars have calibrated speedometers. For example West Yorkshire Police have calibrated speedometers in their Road Policing cars, and Divisional cars do not.

[YarisHybrid2016]

This will bend your brain: what if the error is deliberately non-linear?

There is no reason for the MEASURED SPEED not to be accurate. The inaccuracy in the speedo is deliberate, and controlled by the computer. As long as it is in the required legal limits, then it is fine.

As for GPS recorded speed - this is also subject to error and is NOT precise!

Quite frankly I couldn't care if at 50 MPH I'm doing 48 MPH... on a 100 mile journey (2 hours) the difference is 5 minutes... People worry about seconds in their driving but forget all the hours spent sat at traffic lights. Traffic volume has the single largest impact on journey times.

[Mick F]

I still haven't got back to this.

I will be coming home along the A30 through Devon early on Saturday morning.  If I can get my act together, I'll take a stopwatch and do an indicated 60mph and count the maker posts.  200m apart = eight per mile.  Do a steady indicated 60mph and note the time to do eight posts, and then do an indicated GPS 60mph and do eight posts and note the time.  I'll take a stopwatch as well as my Garmin Montana.

I'm taking Mrs Mick F to Bristol airport for her flight, then returning home.  I'll be by myself, so it's something to do for entertainment! 😀

Mick.

[Mick F]

Small set of experiments on the A30 this morning.

Steady 60mph as indicated on the speedo using cruise control, and kept half an eye on my Garmin Montana plus looking at the TomTom.  Both of the GPS units agreed on the true speed of 56mph.

Next, I increased the cruise control until the Montana and TomTom said 60mph and I saw that the speedo said 64mph.  Then increased again to 70mph as per the GPS and the speedo was saying 65mph.

Then, back down to GPS 60mph and used a stopwatch to time the passing of the 200metre posts.  Eight per mile, and at GPS 60mph, it took exactly one minute per mile.  I repeated this test a few times to confirm.  Also, I reset the trip odometer and checked eight posts and it registered whole miles every eight posts.

Therefore, I can happily state that my speedo is reading high but the odometer is correct.  Also, that GPS speed is accurate.

Mick.

[FROSTYBALLS]

GPS gives an average speed across a number of points, so although it may be more accurate than a speedometer, it isn't wholly accurate. So really all you may have proved is that your odometer is pretty accurate. Speedometers read high - which people already know.

[Mike J.]

From my previous posts we already know that there is an ECU speed that generally matches GPS speeds and it must be the source of the odometer accuracy.

[Mick F]

1 hour ago, FROSTYBALLS said:

GPS gives an average speed across a number of points, so although it may be more accurate than a speedometer, it isn't wholly accurate. So really all you may have proved is that your odometer is pretty accurate. Speedometers read high - which people already know.

I don't agree.  GPS speed is accurate, and I proved it.  It may get an average speed, but it's accurate.  Speedometers read high, yes, we all know that. If the readout is in whole numbers or even tenths, it's accurate enough.

[Mick F]

1 hour ago, Mick F said:

 ................. Then increased again to 70mph as per the GPS and the speedo was saying 65mph.

Sorry.  Typo. 😪

It should have said 70mph on the GPS is 75mph on the speedo.

Mick.

[FROSTYBALLS]

2 hours ago, FROSTYBALLS said:

GPS gives an average speed across a number of points, so although it may be more accurate than a speedometer, it isn't wholly accurate. So really all you may have proved is that your odometer is pretty accurate. Speedometers read high - which people already know.

 

1 hour ago, Mick F said:

I don't agree.  GPS speed is accurate, and I proved it.  It may get an average speed, but it's accurate.  Speedometers read high, yes, we all know that. If the readout is in whole numbers or even tenths, it's accurate enough.

You can disagree all you like, but GPS isn't wholly accurate as it gives an average across a number of points for speed - fact ! The US government quote a horizontal error of 4 metres, and some devices can reduce this to 3 metres. 

[Mick F]

Ok.  Have it as you want.

One minute per mile is 60mph and my GPS devices and BOTH of them agreed with my (clockwork) stopwatch and repeated over over mile after mile with repeated testing.  I had nothing better to do with driving west on the A30 through Devon as the sun was rising.

200metre posts on the roads.  Eight of them equal 1600metres which equals 0.99419miles.  Do that in one minute, and you're doing 60mph or near as dammit.

Mick.

 

[Steve Whits]

On 9/23/2018 at 2:16 PM, FROSTYBALLS said:

Not all police cars have calibrated speedometers. For example West Yorkshire Police have calibrated speedometers in their Road Policing cars, and Divisional cars do not.

That is standard practice with all forces as only Road Police units are trained to use the on board systems to carry out "following checks". (Although some forces have integrated Road Police units with other units now but they will still be specialist vehicles)

[ratbat]

On 9/19/2018 at 6:04 PM, Mike J. said:

Previously, I asked if someone without navigation (and thus without GPS) could run the app - result: <tumbleweed>. 

If they could, then it would be ECU speed. Have you any links to say why it is GPS speed?

biggest clue is when you go through a tunnel and the water app stops with an error about losing GPS signal.

i would guess they use gps to monitor your driving, they wouldn't know what car its being used on, of if there would be accelerators available, so they use the GPS to monitor speed and direction changes.

Also, just because a car doesn't have satnav, it may still have GPS, just not one you can directly access

[Mike J.]

4 hours ago, ratbat said:

biggest clue is when you go through a tunnel and the water app stops with an error about losing GPS signal.

i would guess they use gps to monitor your driving, they wouldn't know what car its being used on, of if there would be accelerators available, so they use the GPS to monitor speed and direction changes.

Also, just because a car doesn't have satnav, it may still have GPS, just not one you can directly access

I assume you have confirmed the tunnel interaction?

Our cars have accelerometers and yaw sensors used by the electronic stability control - no need for GPS.

I doubt Toyota would fit 2 GPS systems in a car. Check out eBay for secondhand Touch and Go GPS modules - big items with their own antennas. I doubt you would need another antenna if the car had a fully integrated one by default.