Speedo Error

M

Mangozac

Senior member
 QLD, Australia
Ever since getting my M Roadster I've had a feeling that the speedo is indicating higher than the actual speed. Recently I had a chance to confirm this when a friend was in the car and used a GPS speed app on his phone. At 100km/h I was doing 10% slower than the speedo indication!

At first we suspected it could be the 19" CSL replicas on the car so I punched the stock and new sizes into https://tiresize.com/calculator/ and compared them. The thing is, even though the rims are a larger diameter, the tyre profile is smaller and hence the overall difference in circumference does not change significantly (it actually reduces despite the larger rim!):

Stock: 255/40r18 Circumference = 81.7"
New: 265/30r19 Circumference = 79.3"

Handily, that site also calculates the speedo error of 2.7% faster. That's still a lot less than the measured 10% though. So I did some more digging.

Firstly, we need to convert the circumferences to impulses per km. I understand that the wheel speed sensors output 10 pulses per revolution.

Stock: Circumference = 2.07518m = 481.9 revs/km = 4819 impulses/km
New: Circumference = 2.01422m = 496.47 revs/km = 4965 impulses/km

I then grabbed the AKMB coding in NCSExpert, which reports the following:

Z_KAHL_TACHO - IMPULSES PER KILOMETER FOR SPEEDO
4961_imp/km
4974_imp/km
4539_imp/km
4846_imp/km
4812_imp/km
4764_imp/km <---- this one is selected

This is interesting - note that there is one option that is extremely close to our calculated impulses/km for the stock size (4812 imp/km), yet the one in use is 1% lower.

I wondered if perhaps the previous owner had monkeyed around with any of the settings so I recoded the car back to its vehicle order and everything stayed the same.

I have read that BMW calibrates speedometers to read 5% faster, but the question is where is that done? If its done before the Z_KAHL_TACHO calculation is performed then that could explain what's going wrong here: 5% (BMW rules) + 1% (Z_KAHL_TACHO) + 2.7% (tyre circumference) = 8.7%, which is not too far off from the (approximate) 10% we measured using the GPS.

So, to bring me back to the BMW 5% error I would need to set Z_KAHL_TACHO to what I calculated from the new circumference, the closest match being 4961_imp/km. Note, I am aware that there are techniques to set the Z_KAHL_TACHO value to other values not listed, but it's a little more complex and risky.

Has anybody else found similar issues and made such adjustments?
 
my head hurts after readin this.
sorry, i can't help.
maybe do a small calculation in the subconcious part of the driving section of ur brain to offset the real speed...?
 
It's normal. Speedometer always shows more and never the same when compared to gps readings. It's simple rule that all manufacturers has to follow, it never can be lower than actual speed but it can be higher. Faster you go, bigger difference you get, nothing to worry about.
 
On stock size mine has less than 2% error.
Speedo must not under read but may over read by 10% and pass MOT in UK.
 
I know it's normal for the speedo to read over and it was fine on my other Zeds, but this one seems to be significantly out and it's annoying! I'll change the settings as I discussed and report back ;)
 
Scooba_Steve said:
Speedo must not under read but may over read by 10% and pass MOT in UK.
Click to expand...
Is the accuracy of the speedometer tested as part of the MoT?

If so, how would they do this and over what range of speeds?
 
ZedFourM said:
Scooba_Steve said:
Speedo must not under read but may over read by 10% and pass MOT in UK.
Click to expand...
Is the accuracy of the speedometer tested as part of the MoT?

If so, how would they do this and over what range of speeds?
Click to expand...

No the Speedo isn't checked during a MOT. How could they with out either every MOT station being equiped with an Rolling Road or immediate access to a 70Mph motorway.

OP before changing anything I would double check your "over read" with something a bit more accurate than a phone app. I'm not saying its wrong, but it may be and if you adjust for a 10% error based solely on that you may be over doing it.
 
MACK said:
OP before changing anything I would double check your "over read" with something a bit more accurate than a phone app. I'm not saying its wrong, but it may be and if you adjust for a 10% error based solely on that you may be over doing it.
Click to expand...
Good advice, but I was only going to adjust it 5% anyway, to correct for the larger wheels, etc. Changing any more requires a value not listed and hence advanced coding, which I'd prefer to avoid.
 
mac27040 said:
is it measured of the back wheels or the fronts as they are different sizes?
Click to expand...
Edit: it's measured from the rear wheels, as I based my above calculations on.
 
Mangozac said:
mac27040 said:
is it measured of the back wheels or the fronts as they are different sizes?
Click to expand...
Edit: it's measured from the rear wheels, as I based my above calculations on.
Click to expand...

Interesting. You would think manufacturers would use the ABS sensors on non driven wheels to eliminate inaccuracies for wheel spin etc.
 
MACK said:
Interesting. You would think manufacturers would use the ABS sensors on non driven wheels to eliminate inaccuracies for wheel spin etc.
Click to expand...
On manual BMWs one of the rear ABS sensors is used (auto cars have a speed sensor on the gearbox). It's standard to sense speed from the driven wheels, one reason being for use on a rolling road.
 
I know the max tolerances is 10% and assumed it was tested similarly to the braking effort. Kind of defeats the point if it isn't tested... Every day is a classroom day, will read up on it.
Had a look at mine again, its no more than 2 mph or 2% out, would annoy me so it's well worth coding to see if it helps.
 
Scooba_Steve said:
Had a look at mine again, its no more than 2 mph or 2% out, would annoy me so it's well worth coding to see if it helps.
Click to expand...
Ducklakeview said:
Scooba_Steve said:
On stock size mine has less than 2% error.
Click to expand...
This... The problem isn't your speedo....
Click to expand...
So this is quite interesting and seems that, even accounting for different wheel size, my car is quite for out for some reason. The plan of attack is therefore to modify the Z_KAHL_TACHO setting to suit my current wheel circumference and then grab a quality GPS (I'm sure my Dad has a portable marine one) to use as a speed reference and see exactly how far out it is. A couple of percent as above will be just fine.
 
Have you considered measuring distance? If the speedo is out then so will mileage be, and distance should be easier to measure.
 
I don't go past the closes distance measurement speedo check very often so the GPS is easier. Will keep it in mind if I do end up on that road though!
 
You must log in or register to reply here.
Back
Top Bottom