> Since truck, bus and train speedos can be accurate why can't they do
> the same for cars? It can't be hard in these days of electronic
> everything to have some optical revolution counter on an axle so whats
> the problem? I drove past one of these flash you speed up signs the
> other day and while my speedo said 38 the sign said 34 so thats a good
> percentage out. And my tyres and new and inflated to the correct
> pressure so they're not the culprit. So why don't car manufacturers
> bother?

> Because it's not required? VW want to sell 8 million cars by 2011. If it
> cost them 10p to make a speedo read spot on rather than be inaccurate
> but legal then that's £800k saved just on the speedo...

> Because it's not required? VW want to sell 8 million cars by 2011. If it
> cost them 10p to make a speedo read spot on rather than be inaccurate
> but legal then that's

> Since truck, bus and train speedos can be accurate why can't they do
> the same for cars? It can't be hard in these days of electronic
> everything to have some optical revolution counter on an axle so whats
> the problem? I drove past one of these flash you speed up signs the
> other day and while my speedo said 38 the sign said 34 so thats a good
> percentage out. And my tyres and new and inflated to the correct
> pressure so they're not the culprit. So why don't car manufacturers
> bother?
>

> I don't see why it would cost more. Computer mice had accurate
> rotational measurement 20 years ago.

>
> If you have a digital speedo, then there's no reason why it can't be
> designed to be accurate to the nearest 1 mph (or maybe even better).
> However if the gauge is analogue, you've got the stiffness and linearity
> of the spring that controls the needle, which will vary from one gauge to
> another due to manufacturing tolerances. I suppose it would be possible to
> calibrate each speedo by supplying specimen inputs at various spot speeds
> and adjusting the software that counts the pulses so it gives an accurate
> reading.
>
> Do any modern cars still use mechanical communication (rotating cable in a
> sheath) between the wheels and the gauge, or do they all use Hall-effect
> sensors (like bicycle speedos) or light-and-rotating-disc sensors (like
> computer mice) these days...

>Cost.
>There can be a fair few percent difference between all the different
>wheel/tyre combinations on a particular car , especially when you take into
>account old and new tyres and over and under inflation.
>
>To try and match all of these would require individually calibrated speedos
>and periodic adjustment at a fair cost compared to "one size fits all" for
>less money

>>
>> If you have a digital speedo, then there's no reason why it can't be
>> designed to be accurate to the nearest 1 mph (or maybe even better).
>> However if the gauge is analogue, you've got the stiffness and
>> linearity of the spring that controls the needle, which will vary from
>> one gauge to another due to manufacturing tolerances. I suppose it
>> would be possible to calibrate each speedo by supplying specimen
>> inputs at various spot speeds and adjusting the software that counts
>> the pulses so it gives an accurate reading.
>>
>> Do any modern cars still use mechanical communication (rotating cable
>> in a sheath) between the wheels and the gauge, or do they all use
>> Hall-effect sensors (like bicycle speedos) or light-and-rotating-disc
>> sensors (like computer mice) these days?
>>
>> I checked my speedo against a sign that lights up if you are over
>> speed and it seemed to agree fairly well: the sign was triggered at
>> about 32 mph as shown by my speedo.

> ...

> Since truck, bus and train speedos can be accurate why can't they do
> the same for cars? It can't be hard in these days of electronic
> everything to have some optical revolution counter on an axle so whats
> the problem? I drove past one of these flash you speed up signs the
> other day and while my speedo said 38 the sign said 34 so thats a good
> percentage out. And my tyres and new and inflated to the correct
> pressure so they're not the culprit. So why don't car manufacturers
> bother?
>