Speedometer Accuracy

[Crossrodes]

I was just wondering if anyone knows the answer to this question:  How accurate is the speedometer on a '79 R65?  I know that Jap bike were inaccurate.

[Barry]

BMW issued a service bulletin on speedo accuracy which said something like + 10% to -0%  + 2 MPH so they should never read low but may be 10% high or even a bit more. Except for the 85MPH speedo's they seem to mostly read 10% high.

[Bob_Roller]

You can check your speedometer against the distance markers on limited access highways, chose a distance and then measure the time it takes to travel that distance .

From what I've seen in print, the 85 mph speedometers were the most accurate, I've got one on my '81 r65, it reads 2 mph high across the entire scale .

We have information only radar sites on a lot of the main arterial streets here in the suburb of hell I call home .

I get a consistent 2mph high from all of the radar sites that I go through .

[Darwin_R65]

I mounted a GPS on my handlebar and when the

Speedo read 60kmh the GPS said 59.
Speedo read 80kmh the GPS said 75
Speedo read 110kmh the GPS said 102

Under 60kmh the 2 agreed but as i got over the speedo continued to under read.

Tyre size, even brand will affect speedo reading as well.

John

[Crossrodes]

Thanks guys.  My speedo is kilometers only (bike made for the German market).  It goes up to 200kph but I also know that this is a replacement speedo.  Whether it was replaced in Germany or here in Canada I don't know.  When I see some distance markers I'll check it.  The warning speed radars that they use here don't seem very accurate.  Being an old Radar Tech I'm betting that they rarely get calibrated.  I'm going riding with some friends today.  I'll ask if anyone knows if their speedo is accurate and check it against that if it is.

[Barry]

Now I'm at home I can post the service bulletin about permissible speedo error.  I was 0.4 mph out  it's 10% + 2.4 mph.

[tvrla]

Barry, or anyone - any idea what the +2.4 is all about???

As if it's bad enough having a 10% discrepency - adding another 2.4mph on top of that??!!!

I find it odd that Germans who are so exacting would allow such a slopppy tolerance.

[nhmaf]

There was (or is) a law regarding motor vehicle speedometer accuracy which mandated that speedometers from manufacturers MUST NOT read low, e.g. say that you are going 55 MPH for example when in fact you are going 60 MPH.   I used to have a printout of it but cannot find it at the moment.  In many cases, it is therefore easier to simply build in a fixed offset/error to help ensure that, across the statistical sampling of units, that you almost never create any units which read low.   By skewing the "average" in this respect, you move the centerpoint of your bell-curve (assuming a normal distribution) so that only a tiny percentage of truly faulty units would read low, and you cull those out during production testing as much as possible.   Then, you have the additional tolerance/variances which could add another 0 - 10% on top of that.

[Barry]

Barry, or anyone - any idea what the +2.4 is all about???

As if it's bad enough having a 10% discrepency - adding another 2.4mph on top of that??!!!

I find it odd that Germans who are so exacting would allow such a slopppy tolerance

Well I'm no expert on speedo's but in calibrating gauges there are two types of error and two adjustments to make. The 10 % figure is a span error that is an error that gets proportionally bigger as the gauge reads higher and the 2.4 mph is a zero offset error which remains the same over the whole scale of the gauge. Both types of error would occur in manufacture and both could be mostly adjusted out.

I don't doubt they could have done a lot better than 10% plus 2.4 mph if they tried so you have to accept the size of the error is deliberate.  If they were trying to market an analogue gauge on it's accuracy I'd say they could get within 2 %.  

The German mentality is to think of a safety margin then double it.  A good example is spring rates. leaving aside suspension springs which have to be right other springs on German bikes and cars for example throttle return springs tend to be twice as strong as really needed.

[tvrla]

I can sort of follow you, Barry. The extra 2.4 is still rather odd, considering there's already a 10% margin. Why add another 2.4 mph?

Interesting about springs! Are you saying we could go with throttle springs half as strong?

[Crossrodes]

It's interesting that Honda cars and motorcycles were also out (speedo showed a higher speed).  The  problem was that the error also applied to odometers so people who leased cars were charged more money for the mileage they put on their cars.  There was/is a lawsuit when people discovered this. So today Honda cars and motorcycles have fairly accurate speedos and odometers.  I wonder if BMW did or will face the same problem?

[Crossrodes]

I rode with some friends today.  One had an 84 (I believe) R65.  I asked him about his speedo.  He said it was out by 10%.  So I checked mine against his.  They were both the same at 100kph.

[Barry]

Re: Speedometer Accuracy
I can sort of follow you, Barry. The extra 2.4 is still rather odd, considering there's already a 10% margin. Why add another 2.4 mph?
 

It is odd but span and static error was always the standard way of expressing error when I was calibrating gauges. It's the magnitude of the errors thats odd although 2.4 would only be a 2% static error on a 120mph speedo and could be corrected by repositioning the needle on the spindle. The 10 % span error was deliberate on their part and  presumably could be adjusted out by increasing the hair spring tension or alternatively by increasing the drive cup magnetic air gap.

Interesting about springs! Are you saying we could go with throttle springs half as strong?

Well maybe not quite half as strong but it's a recognised issue. Didn't someone set up in business to manufacture and sell lighter weight springs for airhead carbs.  

On my other German vehicle a Merc 190E, the throttle return spring was so strong it would make your ankle ache. I swapped it for a lighter one 18 years ago and it's been fine ever since.

German manufacturers at least in that era tended to over engineer things. A motoring journalist once wrote that the spring on a Merc ashtray was strong enough to kick start Concorde. A bit of an exaggeration perhaps but the ashtray spring along with several others was unnecessarily strong.

One other spring that was too strong on an my airhead was the clutch operating arm return spring. I measured the clutch lever pull at the handlebars with and without the spring using a spring balance. I forget the figures but the spring was adding something like 25 - 30% to the clutch lever weight. The spring is only there to return the cable and take the weight off the thrust bearing so I reduced the spring tension and was rewarded with a lighter clutch. Later models have a different arrangement and are probably lighter anyway.

Theproblem was that the error also applied to odometers so people who leased cars were charged more money for the mileage they put on their cars.

That's interesting as oddmeters are usually more accurate. My Merc is as close to spot one as I can measure certainly within 1%. The bike under reads by 2.5% with the current rear tire fitted but tha's a bit of a variable anyway with airheads as there's at least a 3% diameter variation in the range of  tires typically fitted to airheads.

[tvrla]

Interesting data, Barry. Thanks for that!

[Crossrodes]

Here's some info on the lawsuit in Canada: http://www.thestar.com/business/article/580143--honda-offers-to-settle-odometer-suit

There was a suit in the US also.