Not a lot of people know it, as Michael Caine once remarked, but the BMW tolerance for speedometers is 10% plus 2.4 mph.
This means that at 30 mph it could read 35.4 mph, at 50 a reading of 57.4, and at 70 mph a reading of 79.4 mph. Maybe!
An industrial meter has to meet a 1.5 % tolerance, which makes one wonder why speedometers don't.
Speedo Error
Mike, I suspect the gauges in BMW's are actually pretty accurate, it is more likely to be the way that the speed is measured that isn't.
First up, from what I can work out, the signals for the various dials is measured digitally and then converted to analogue using an ADC. So the best measure of the speed is the self-test mode that gives a digital readout. There may be some errors in the alignment of the scale on the dial and so on, so this may make for the poorer tolerance - it is after all only a piece of paper stuck behind a needle.
One thing this approach does seem to do is to dampen the signals to give a more consistent reading. But as the speed is measured by the rotation of the wheels, then the exact tyre sizes and pressures all come in to play affecting the accuracy. Even a cold day will make your speedo read higher.
Another thing that screws up the speedo is bumps in the road and vibration within the wheels and sensor. If you go back to old fashioned analogue measurement systems they varied immensely between cars. In the sprite I hired, as 60ish the needle oscillated between 50 and 70 - so you only ever had an approximate idea of your speed. The rev counter on the other hand was much more stable. In the one I have just bought the needle only oscillates by a few mph once you get above about 40.
So how accurate is your speedo? Well I'm not sure that my satnav is totally accurate, but using that as a guide, I would say the mine under-reads by about 5%. So with the needle just under 75, I will actually be doing dead on 70 mph by the satnav.
The only way to measure speed more accurately is not to rely on the wheels at all, but to relate it to the speed at which the ground passes underneath the car, or to time the car between two fixed points a known distance apart.
Cheers R.
First up, from what I can work out, the signals for the various dials is measured digitally and then converted to analogue using an ADC. So the best measure of the speed is the self-test mode that gives a digital readout. There may be some errors in the alignment of the scale on the dial and so on, so this may make for the poorer tolerance - it is after all only a piece of paper stuck behind a needle.
One thing this approach does seem to do is to dampen the signals to give a more consistent reading. But as the speed is measured by the rotation of the wheels, then the exact tyre sizes and pressures all come in to play affecting the accuracy. Even a cold day will make your speedo read higher.
Another thing that screws up the speedo is bumps in the road and vibration within the wheels and sensor. If you go back to old fashioned analogue measurement systems they varied immensely between cars. In the sprite I hired, as 60ish the needle oscillated between 50 and 70 - so you only ever had an approximate idea of your speed. The rev counter on the other hand was much more stable. In the one I have just bought the needle only oscillates by a few mph once you get above about 40.
So how accurate is your speedo? Well I'm not sure that my satnav is totally accurate, but using that as a guide, I would say the mine under-reads by about 5%. So with the needle just under 75, I will actually be doing dead on 70 mph by the satnav.
The only way to measure speed more accurately is not to rely on the wheels at all, but to relate it to the speed at which the ground passes underneath the car, or to time the car between two fixed points a known distance apart.
Cheers R.
Arctic Silver '99 Z3 1.9 & Black '59 Frogeye 1275cc
You mean over-reads?Robert T wrote:I would say the mine under-reads by about 5%. So with the needle just under 75, I will actually be doing dead on 70 mph by the satnav.
In Autocar road M Coupe roadtest the speed was only out by 2mph at 100mph compared with timing gear. I can't see the rest of the raneg being any different, so I would say on the whole they're quite accurate. If you can base it on the gearing (ie 23.3mph/1000rpm in 5th) then it seems to be accurate.
Re: Speedo Error
At 78mph on the speedo my SatNav reads 70mphMike Fishwick wrote:
This means that at 30 mph it could read 35.4 mph, at 50 a reading of 57.4, and at 70 mph a reading of 79.4 mph. Maybe!
Gazza
"Understeer is when you hit the wall with the front of the car, oversteer is when you hit the wall with the rear of the car. Horsepower is how fast you hit the wall and torque is how far you take the wall with you"
Z3 S54 M roadster
, BMW Z1, BMW M3 CSL, Z4M Coupe
"Understeer is when you hit the wall with the front of the car, oversteer is when you hit the wall with the rear of the car. Horsepower is how fast you hit the wall and torque is how far you take the wall with you"
Z3 S54 M roadster
, BMW Z1, BMW M3 CSL, Z4M Coupe
Re: Speedo Error
Speedometers are required (by EU rules) not to under-read. 10% plus 2.4 mph sounds a lot like the 10% plus 4kmh maxium allowed by the relevant rules, so I suspect that BMW are merely restating the constraints of the rules that they're complying with.Mike Fishwick wrote: An industrial meter has to meet a 1.5 % tolerance, which makes one wonder why speedometers don't.
As has been pointed out, various factors such as tyre wear and pressure, and potentially different diameter tyres on different OEM wheels could affect the reading, so it shouldn't be surprising that manufacturers give themselves a few mph leeway on the lower limit.
My experience is the same as c_w - a consistent over-read of a couple of mph rather than a greater error at higher speeds: this suggests that the speedos are built to a low tolerance, they're just deliberately miscalibrated for legal reasons.
Paul
Having used my GPS for a number of years in various (mostly BMW) cars, I find the following - All over-read speed, usually by a constant value, rather than a % of the speed.
On our 5 series and 3 series it's about 3-4 mph and on the Z it's 4-5 mph. In the Audi it's 4 mph and in a recent VW 6-7 mph! With all the SPECS camera's everywhere now, cruise control is a godsend along with a GPS system you trust to set the speed correctly.
I think the technology is such that it's not difficult, the real world issue is probably more to do with how the manufacturers wish to deal with the issue of 'speed'.
My bike over-reads massively, but that's 'cos it's geared down for stronger acceleration (and it works off a gearbox output sensor)
On our 5 series and 3 series it's about 3-4 mph and on the Z it's 4-5 mph. In the Audi it's 4 mph and in a recent VW 6-7 mph! With all the SPECS camera's everywhere now, cruise control is a godsend along with a GPS system you trust to set the speed correctly.
I think the technology is such that it's not difficult, the real world issue is probably more to do with how the manufacturers wish to deal with the issue of 'speed'.
My bike over-reads massively, but that's 'cos it's geared down for stronger acceleration (and it works off a gearbox output sensor)
someone in a minority once wrote:I know I'm in a minority
You can have a difference of a couple of mph between a brand new tyre and one that is 'just legal' at the same temperature. If you also factor in ambient temperature, tyre temperature, wear and tear in mechanical components, minor errors in the digital to analogue conversion for the gauge, then it is easy to get a 10% variance.
That is why a zero tolerance law towards speeding would not work. Too many error factors that the driver has no control over. Even GPS has errors that vary according to how many satellites are in range at the time a reading is taken as well as the built-in errors in consumer GPS that are there for military reasons. (Don't get me started on the differences between consumer GPS and military GPS)
Overall the best you can do is get your car checked regularly and every couple of years get your speedo accuracy checked and if it has wandered too far from correct then get it re-calibrated (maybe once every ten years or 100K)
That is why a zero tolerance law towards speeding would not work. Too many error factors that the driver has no control over. Even GPS has errors that vary according to how many satellites are in range at the time a reading is taken as well as the built-in errors in consumer GPS that are there for military reasons. (Don't get me started on the differences between consumer GPS and military GPS)
Overall the best you can do is get your car checked regularly and every couple of years get your speedo accuracy checked and if it has wandered too far from correct then get it re-calibrated (maybe once every ten years or 100K)
Now with both the 411 bhp Supercharged Z3MR and a WRX STi Wagon getting a 3.4 litre transplant and an XKR with mods planned.
Fast is good, faster is better. Wheeeeee !!!
Fast is good, faster is better. Wheeeeee !!!
