OT: Car speedometer accuracy

[Dave Plowman (News)]

It's not. Depends on the car maker. Some do fit accurate units. Same as always.

[dennis]

Tyres slip where they contact the road, they have too so you can turn corners. This is the main cause for wear.

[news]

Chronometric?

I had one on the motorbike and whilst I understood the principle of using an escapement to trip after a time interval and reset the distance spindle I could never see how it actually worked.

AJH

[Dave W]

Why can't it happen? You've snipped my agreement that one revolution of the tyre must equal one revolution of the wheel. It's just that various parts of the tyre are going at different speeds.

[Dave Plowman (News)]

Yes. I'd guess they were the earliest type. But still in use quite late on with 'sporty' cars. And of course by the police until electronics made things easier. They were notable for moving in steps. Perhaps that's why some now use a stepping motor. Nostalgia.

[dennis]

Most steppers will do a sweep of the dial just after you turn the ignition on. They don't need to but its trendy. You can turn it off in the settings menu on some cars.

[Dave Plowman (News)]

Ah. I thought the cars I'd seen do that used a sort of plasma display. So no physical needle.

My stepper motor idle valve does chunter back to the stop then count up to the correct position at switch on. Very quickly, of course.

[Johnny B Good]

Well, the "Chronometric"(tm) speedo heads *were* actually based on mechanical clockworks, not a magnet or aluminium disk in sight. They literally measured a distance over about a quarter of a second gating interval derived from a balance wheel escapement mechanism.

The speedo cable feed drove a clutch disk which in turn rotated another disk against a restoring spring which would return it back to its reference starting point. This disk had a peg which in turn drove the indicator needle around the speedo dial against a detent which would hold the reading during the next measuring cycle. The detent would be released briefly at the end of the subsequent measuring cycle before locking the reading so that on deceleration, the needle could drop back to the slower speed measurement. During acceleration, the needle would simply be nudged up to the higher reading every quarter of a second or so.

These could be calibrated very closely to the actual road speed by adding or removing clip on shims to trim the mass of the balance wheel and were guaranteed to be linear (give or take a detent notch - 1mph I believe).

However, the real issue with speedo accuracy derived by measuring the prop shaft or road wheel revolutions has more to do with the differences between new and end of life tyres (where, fortuitously, the effect of wear on the tread blocks which results in a reduced tyre diameter is countered by the reduced give or 'slip' in the now reduced profile tread blocks), the effects of "Tyre Growth" at high speeds due to centripetal force and the variations in 'slip' due to variations in wind resistance and slope of the road.

Whilst the basic spinning magnet and aluminium disk speedo can't be so readily calibrated to the same accuracy as the classic "Chronometric"(tm) speedo heads, with care, it is possible to approach this level of accuracy very closely but, when all the other sources of variations in the rest of a system dependant on sensing speed from the prop shaft or road wheel(s) are taken into account, there's not a lot of point in doing so. It is sufficient that a given model of speedo head intended for any one model of road vehicle remains consistently matched to the rest of its siblings coming off the production line.

Modern satnavs can generate quite precise average speed over ground readings which, in conjuction with maintaining a steady cruise speed derived from the vehicle's own speedometer, approximates very closely to an actual road speed for comparison against the speedometer from which you can derive your own personal calibration curve by which to assess true speed versus the speedometer reading.

For any number of reasons, it should come as no surprise to find a modern car speedo is likely to be some 5 to 10 percent optimistic of the actual road speed it purports to be measuring. All you're doing when 'calibrating' the speedo from your satnav's average speed readings, is determining the magnitude of this error for future reference, bearing in mind that it's probably best to assume the speedo reading as 'accurate' for steep downhill stretches of high speed sections, especially when enjoying the speed enhancing effect of a strong tailwind into the bargain.

[stvlcnc43]

You're looking at it from the wrong side: radius is the crux, the tyre NOT touching the ground can be of any shape, it can be a tank track. The tyre i s merely a cushion between the solid portion of the wheel and the ground. I f the cushion were on the ground and the wheel was just steel, the result w ould be exactly the same.

[Fredxxx]

There's me thinking that the travel in a single rotation of a wheel was determined by its circumference!

Have I been wrong all this time?

[Harry Bloomfield]

Dave Plowman (News) brought next idea :

They drive enough steps towards the zero, to be certain they have reached the stop pin, then advance from there - a known reference point, driven by the pulses.

[NY]

You are correct, but only for a solid wheel. If the wheel is compressible (as the inflated tyre makes it, even if only slightly) then what TNP said is correct: the radius of the part that touches the road is the only bit that matters. Given that the wheel itself is circular and that (presumably) the tyre compresses uniformly as it rotates (ie that the radius to the point that touches the road doesn't vary as the tyre rotates), a tyre with a slight "flat" on it where it touches the road is equivalent to a solid wheel of the same radius as the bit that touches the road - which will be slightly smaller than the radius of the uncompressed tyre.

I wonder how much a tyre typically compresses and therefore how much smaller the effective radius is than the radius of the uncompressed tyre if the wheel were raised off the ground.

[Vir Campestris]

There are posts every 100 metres along motorways. Get your passenger to count 16, and ignore the odd 9m.

The satnav will be bang on on a straight level road. It falls out of the technology - it has your position, and access to several atomic clocks.

Andy

[Harry Bloomfield]

Johnny B Good brought next idea :

Thanks for that explanation, I did wonder how they worked.

[Dave Plowman (News)]

Yes. Characterised by a stepping movement. However, eddy current ones are pretty consistent. So if properly calibrated in the first place can be perfectly accurate enough. Certainly far better than a 10% error at 30 mph.

Right. It's one of the few things cars wise I've not owned or had apart.

As has been said, tyre wear can only result in the speedo over-reading even more than the law allows. But never under read. Tyre growth is going to be minimal at the sorts of speed legal in the UK.

[The Natural Philosopher]

No.

[The Natural Philosopher]

I didnt say what you said I said. I am correct but you are wrong.

the radius of the part that touches the road

No, it is meaningless. all that matters is the circumference. The squashed bit of tyre and its tread moves with repectr to the rim. It has to.

Given that the wheel itself is circular

Completely wrong. The tyre tread HAS to move with respect to the rim Because the circumference of the rim is smaller than that of the tread.

You are faced with two basic facts. One revolution of the tyre no matter what shape it is must move the car one circumference forward, and one revolution of the tyre must happen in the same time as one revolution of the wheel.

That this conflicts with your simple picture if a RIM rolling along the road, is where it gets interesting.

And it is in the fact that the tread has to move faster than the rim where the tyre is compressed. And the sidewalls distort to allow this to happen.

The picture you have in your head is probably like this:

However anyone whop has actually played with a setup like that knows that the outer ring ("tyre") rotates slower than the internal pinion ("wheel rim"). By a huge amount

Unless the tyre is actually so under inflated that its *slipping on the rim* it is patently obvious that the tyre has to be rotating at the same rate as the wheel, therefore all that matters is the circumference. Whether a standard steel belted radial changes its *circumference* when it's inflated, is the only point that matters.

Who cares?

"effective radius" is meaningless in the context of a wheel and tyre that is not round. Circumference still has meaning, radius does not, except as circumference divided by two pi.

[dennis]

Maybe but maybe not. Don't forget they fit the same speedo to several models and they change the wheel and tyre sizes depending on the trim level so the inaccuracies are going to be quite high.

With digital ones they can change the parameters to match which is why they can be more accurate unless the dealer has done a special and swapped the wheels/tyres.

The driver fitting winter tyres may well cause errors too as they are unlikely to be the same size.

[DICEGEORGE]

HGVs are limited to 56mph 90kph by EU law, but their speed limit is 60mph as speed limits must at 10mph's

d-wales/

" the national speed limits for HGVs over 7.5 tonnes, travelling on a singl e carriageway will increase from 40mph to 50mph. The speed limit for HGVs o ver 7.5 tonnes, travelling on dual carriageways will increase from 50mph to 60mph.

The limits in Scotland are staying the same. European speed limiter require ments also remain unchanged and must be set at 56mph or lower."

"?All UK speed limits must be in multiples of 10 mph, as there is n o requirement for vehicle speedometers to show increments of less than 10 m ph?."

[pamela]

I think Jack Johnson was saying the speedo uses the "squashed type" radius when calculating a speed to display and therefore that radius is the one to measure.

Irrespective of "squashed but equal length" circumferences.

Seems sensible.