Bike Computer Calibration Help Please

[eck]

I did a ride today with a new computer, which I'd calibrated using the "roll the bike along the floor with a tape to measure the circumference" method.
It gave 212cm, so I that's what I put in to the computer.
When I got back, I was well chuffed with my average speed of 30kph. Until I suspected the computer was reading a bit on the generous side. 

The computer gave the distance as 67.62k. I've measured the route on bikely (66.8k), toporoute (66.71k), and mapometer (66.64k), a mean of 66.72
Can I re-calibrate it using the good old cross-multiplication method?
ie If 212cm gives 67.62k, then 66.72 must be 209.2?

Or is it not as simple as that?  Would my not inconsiderable weight sitting on the bike make as much as 3cm difference to the circumference.

Funnily enough, the ready reckoner table for 700x23 says 209.6, a lot closer what I think it might be. 

[Greenbank]

Just go with 209.6, it's a better figure than 212cm.

The good news is that it only drops your average speed from 30kph to 29.66kph.

[PaulF]

Sounds sensible enough to me

[Biggsy]

The computer gave the distance as 67.62k. I've measured the route on bikely (66.8k), toporoute (66.71k), and mapometer (66.64k), a mean of 66.72

Those route calculators don't take into account all your little wobbles.  Your computer might be giving you a truer answer to how far you travelled.

You will get a more accurate calibration by sitting on the bike while measuring, but that'll only make a small difference unless the tyres are really soft.

[teethgrinder]

I seldom bother measuring my wheel. It's not very easy to get a very accurate reading anyway.
If you want to be extreme about it, you'll need to measure your tyre pressure and keep it the same at all times aswell as keeping your bike and yourself the same weight at all times and also account for tyre wear. But that's being a bit OTT.
209 is my default setting for 700c road wheels. My result in the Mersey 24 TT is almost always 1% less than I have on my computer. Given that the course is measured as accurately as possible, using the shortest possible route and also that I think that TT courses are always a tiny bit over distance on purpose so that any record breaking ride won't be false because of the course being under distance, then I reckon I must have mine set very well.
212 sounds too high to me, unless you have big and chunky tyres.

[Biggsy]

On second thoughts, maybe weight makes more difference than we think.  I've just measured an Open Corsa 700x23 on Open Pro as 212 cm without weight, too.  (Not convenient for me to measure with weight at the moment).

[MSeries]

think about it, the circumference is 2 x Pi x radius.

so the difference is 2 x Pi x (r₁ - r₂)

suppose the weight of the rider squashes the radius by 5mm we get

2 x Pi x 5 for the change in the circumference

31.4 mm

A 5mm average squash may not be unreasonable.

[GruB]

What is the ETRO value on your tyres?

[Biggsy]

What is the ETRO value on your tyres?

eck said "700x23", so that'll be 622mm inner diameter + 23mm depth = 645mm outer diameter.

But the actual diameter depends on manufacturer accuracy/honesty and the width of the rim.

[teethgrinder]

On second thoughts, maybe weight makes more difference than we think.  I've just measured an Open Corsa 700x23 on Open Pro as 212 cm without weight, too.  (Not convenient for me to measure with weight at the moment).

I always sit on my bike if I'm measuring. The measurement is down to less than a millimeter. I doubt that I ever get it exactly right.

Close enough is good enough for me.

think about it, the circumference is 2 x Pi x radius.

so the difference is 2 x Pi x (r₁ - r₂)

suppose the weight of the rider squashes the radius by 5mm we get

2 x Pi x 5 for the change in the circumference

31.4 mm
A 5mm average squash may not be unreasonable.

You mean that the tyre is squashed 5mm when the rider sits on the bike?

That sounds like a lot to me, or about right for a mountain biker with low air pressure to absorb shock, which is common.

I think that the rider weight would make a few millimeters difference to the circumference. 2mm error on the circumference measurement (or 1mm on the radius) will work out at roughly 1% error on the computer readings, so on one hand, it's worth getting it right, but on the other, I don't think it's worth worrying about getting it spot on, because it won't be spot on all of the time unless you are very determined to keep it there.

209 is 3 less than 212 and eck's reading are almost 1% out, according to bikey etc, so 209 or 210 wouldn't be far wrong.

[MSeries]

Yes, teethgrinder that is exactly what I proposed. I've not mesaured it though.


1 mm average squash makes more than 6mm difference in the circumference.


FWIW I don't use bike computers anymore

[Greenbank]

The computer gave the distance as 67.62k. I've measured the route on bikely (66.8k), toporoute (66.71k), and mapometer (66.64k), a mean of 66.72

Those route calculators don't take into account all your little wobbles.  Your computer might be giving you a truer answer to how far you travelled.

More importantly, all of those route calculators (Bikely, Toporoute, Mapometer) all use the same underlying data (via Google Maps). The difference is due to their own algorithms for placing intermediate points, calculating distances, removing spurious points, averaging, etc.

But three interpretations of the same data all agreeing to within 160 meters over 66km is no great surprise given they use the same underlying data.

If you used another mapping program you may get a figure that agrees with your computer, or disagrees with it even more.

[teethgrinder]

FWIW I don't use bike computers anymore

If I wasn't so vain, I doubt that I would.
I like numbers too much.

[Tewdric]

As TG says, it's really difficult to measure the circumference of the tyre by rolling it along a tape measure. It's counterintuitive but using the recommended setting has always proved more accurate for me!  I really took my time last time I did it and drew a line the whole width of the tyre and rim with a  fine felt pen, using a set square, and then I used the set square again to determine the start and end points of the roll.  I still got a more accurate distance result using the preset 23c setting! (with Pro Race 3 incidentally) 

[Deano]

I just use whatever it says in the table that comes with the computer instructions.  2096 for 700c x 23mm.

[Wobbly John]

Put a dab of emulsion paint on the tyre. Ride your bike a few yards. Measure the distance between paint splots on the road. Clean paint off tyre - job done.