Bicycle Mechanics - why the variation in speedometer charts on tire size --> circumference

so the user's guide to every speedometer has a little chart giving the circumference for common tire sizes.
But the exact circumference (almost always listed in millimeters) for a given tire size is not consistent across the various speedometer user's guides that I have.

What I'm wondering is, is the variation actually meaningful? For example, say I have two bikes with identical 700x25c tires. One speedometer says to program 2124mm circumference, another says 2146mm. Is this meaningful? that is, does the speedometer have something else going on in its programming that I should plug in the number specified?
Or are all speedometers just multiplying wheel circumference by the number of rotations to get distance?

I assume the latter, and the differences in listed circumference are just because of variations in the tires that the manufacturers measured, or because they were just sloppy in producing the chart. But I'm curious to hear if others share my assumption.


Just for the record, here's an example of the existing variation. Circumference measurements for 700x25c tires from the various speedometer user's manuals I have...

2100 (Cateye Mity 2, from about 1995)
2105 (Filzer wireless)
2110 (Cateye Mity 8 and Cateye Velo 2)
2120 (Planet Bike Protege)
2124 (Ascent wireless)
2146 (E3 and Axiom, which I think are both Performance in-house sub-brands)

The variation is due to rim width. the wider the rim, the less tall the tire will be and that will affect the overall diameter.

i would the the blame of laziness goes with the owner, who should be measuring their own rolling circumference for the most accurate reading.

For the most accurate measurement you should also sit on the bike as your weight will deflect the tire and change the circumference. Also you should ensure that your tire is always at the same identical pressure, otherwise the circumference will drop and your readings will be affected likewise.

Those are just rough estimates. Actual tire sizes vary from brand to brand even though the nominal sizes are the same.
There is an area near my home with section line roads that have survey pins in the center of the intersections. Pin to pin is exactly one mile. In the past when I wanted a lot of accuracy I would calibrate my speedometers with those survey pins. But I quit worrying about it and no longer bother to do that.

glad to hear that others share my assumption. I just wanted to make sure that this all points back to "measure your own wheel circumference."
And also "speedometer makers are lazy or at least not careful when putting together their guides/charts."

AFAIK they just multiply the wheel rotations. And every tire manufacturer is different, and every bike and rider weighs a different amount, and everyone inflates their tires differently. The problem is that the circumference measured is not the circumference that one would tape with the wheel off the bike. Rather it is pi*D, where D is twice the radius from the hub center to the road in the center of the contact patch, the bike being loaded. So at best, the manufacturer is offering an approximation.

To get it closer for the initial installation, stretch out a tape on the floor next to a wall. Get on the bike and push yourself along the wall, while measuring valve-to-valve on the wheel with the magnet. That should be close for that tire and inflation. Then the next time you're out on a long measured course, see how your mileage matches up with that of the course designer. You can use ratio and proportion to tune it closer. Every time you fit a new brand or size of tire, you'll have to do it over if you want it to be accurate.

The problem is that the circumference measured is not the circumference that one would tape with the wheel off the bike. Rather it is pi*D, where D is twice the radius from the hub center to the road in the center of the contact patch, the bike being loaded. So at best, the manufacturer is offering an approximation.

Correct, measuring the circumference with a tape will not give you the right answer.
If you do a roll out measurement it will be better to measure several rotations and then average those.
And do this for the wheel with the magnet and computer sensor. The rear wheel carries a much larger proportion of the weight.

Calibration numbers given in mm or cm are the estimated circumference of the tire so the bike's travel distance is just that number times the number of revolutions. Avocet used to express the circumference in inches and others may use an arbitrary number.

Cat Eye, among others, uses the circumference in mm or cm as their calibration values. From their charts it appears they assume the tire is circular in cross section. Therefore, the wheel diameter is the rim's bsd +2* (tire width) and the circumference is pi times that number. For a 700x23 tire the diameter would be 622 + 2*23 = 668 mm and the calibration number would be 668*3.1415..... = 2099 mm or 210 cm. And, in fact, that is the value their chart gives for a 700x23 tire.

Variations in calibration number from manufacturer to manufacturer are possibly due to many bike tires not running true to their marked size. I've seen tires marked 700x23 as small as 21 mm in width and I have some tires marked 700x32 that actually measure 26 mm both wide and high.

Also, loading the tire and doing a rollout can give slightly different distance than the assumed circumference so that may be another source of difference. Actually, a carefully done rollout is the most accurate way to determine the true calibration for your particular tire and weight.

Your example values of 2124 and 2146 is a difference of 1% or 1 mile in a century. I personally want better accuracy than that but many riders don't care.

Your example values of 2124 and 2146 is a difference of 1% or 1 mile in a century. I personally want better accuracy than that but many riders don't care.

In the data I report above, it's worse than 1%. The difference between 2100 (low value) and 2146 (high value) is 2%. That's an even bigger deal, and as you say, below the level of accuracy I want in measurement.
Now, I could arguably throw out the 2146 from analysis (since it's such a far outlier, we could just say it's erroneous, and even though it showed up in two user's manuals, those are both Performance in-house brands so they presumably were working from the same chart). And we might also throw out the low value (2100) since it is from the mid-90's whereas everything else is mid-2000's and since. Then the low value is 2105 and high value is 2124, which is still a variation of 0.90%

In the past I've ridden a lap or two on a running track (400m or 800m) with the speedometer set in km, to try to nail exactly 0.40km or 0.80km. As long as I'm not weaving, that's about the best roll-out measurement I could hope for. (unless the tire behaves differently on a rubberized track than it does on pavement)

I measure a course with google maps, then adjust the wheel size setting by the percentage that the speedo is off. Setting it this way over 20-25 miles is as accurate as google maps at 3x the distance so that's close enough for me.

I always wondered about weighting the bike for rollout tests - I don't see how the circumference would change. (Tire squish ~= contact patch size != circumference). I started to test it once, but I got sidetracked thinking about the difference in effective torque from the shorter effective radius of a loaded/squished tire, then I saw a cool butterfly... and then I quit worrying about rollout measurements and went for a ride.

A single rollout with or without a bit of weight on the bars is sufficient to be within less than 1% accuracy. All the other time one would spend is better done on the road, riding.

What I'm wondering is, is the variation actually meaningful? For example, say I have two bikes with identical 700x25c tires. One speedometer says to program 2124mm circumference, another says 2146mm. Is this meaningful? that is, does the speedometer have something else going on in its programming that I should plug in the number specified?
This is a 0.1% difference. That's 0.1 miles per 100.

Since all 25mm tires are not all the same and things like pressure and load produce different effective circumferences, the best thing to do is measure your setup (with you and your load on the bike). People might prefer entering a number from a chart rather than doing the real measurement.


Or are all speedometers just multiplying wheel circumference by the number of rotations to get distance?
Yes. Speedometers (other than GPS) detect wheel rotations and compute the speed/distance from the circumference entered.

I always wondered about weighting the bike for rollout tests - I don't see how the circumference would change. (Tire squish ~= contact patch size != circumference).
The patch size is related to the effective circumference.

The more load you put on the tires, the wider the patch gets and the closer the hubs get to the ground. The effective radius (the distance between the center of the hub and the ground) decreases (which means the circumference decreases as well).

As already suggest...don't waste time on this. Do a couple rough rolls with inflation and some weight...input a "close enough" estimate and ride.

If you are going to agonize over every little 10th of a MPH or 100th of a mile - I question why you have a bike to begin with.

To be honest, most folks I have setup for are happy with the generic references in the manuals.

=8-)

This is a 0.1% difference. That's 0.1 miles per 100.

Wrong! The original calculation was correct at 1% or 1 mile per 100. Check your math.

In the data I report above, it's worse than 1%. The difference between 2100 (low value) and 2146 (high value) is 2%. That's an even bigger deal, and as you say, below the level of accuracy I want in measurement.
Now, I could arguably throw out the 2146 from analysis (since it's such a far outlier, we could just say it's erroneous, and even though it showed up in two user's manuals, those are both Performance in-house brands so they presumably were working from the same chart). And we might also throw out the low value (2100) since it is from the mid-90's whereas everything else is mid-2000's and since. Then the low value is 2105 and high value is 2124, which is still a variation of 0.90%)
I have no idea where the 2146 could have come from but, as you say, it's probably an outlier or a typo. All of the Cat-Eyes I've seen use 210 - 211 cm for 700x25 tires and that's calculated from the circular cross section assumption I mentioned above. If the tires are true 25 mm then the actual rolling circumference will be very close to that if the tires are inflated to say 100 psi or thereabouts as the deflection is pretty small.

BTW, I use 209 cm (Cat-Eye says 210) for 700x23 tires and once rode a 10K TAC certified race course on my bike and it agreed with the course distance (6.214 miles) within a few feet. Good enough. Various model tires differ more than that.

The range from the smallest to the largest (including the 2146) is about 2%. When you inflate your tires are you consistent to within 2%?

Sometimes close enough is close enough.

If my speedometer say's I'm going 19.2 mph, I expect that I'm actually traveling at 19.2 mph and not 19.0!

I get really pissed when I find out it's wrong. (really, I do).

So now let me describe how I calibrate my computer. Tell me if this is accurate or not.

I set it according to the instructions, Sheldon Brown's numbers or even the default number it comes at. Just to get a baseline.

Then I ride the bike for a few miles over a course that I've measured on Google Earth. I generally use the Yolo Causeway because it's almost perfectly flat and straight and 3.11 miles from levee to levee. It's also on my way to work. As I ride, I try to stay as straight and centered as I can.

Then I take the distance my computer reads and divide that number by the distance I rode. Then I multiply that number by whatever number I told my computer was the size of my tire.

I did this today, actually. I had entered 2136 as my tire size. I rode the causeway. At the other end, it said I rode 3.15 miles. 3.11/3.15=0.99 .99x2136=2115 (my new wheel size).

The patch size is related to the effective circumference.

The more load you put on the tires, the wider the patch gets and the closer the hubs get to the ground. The effective radius (the distance between the center of the hub and the ground) decreases (which means the circumference decreases as well).

Nope, radius is tied to circumference only for a circle. Once you put the tire on the ground it is no longer a circle. I'm not saying there is no "stretching" of the outer part of the tire, because in practice there is a very slight amount - with a flat steel tape around a tire I can feel it but not measure it. But for the most part the deflection is in the cross section getting wider, and the outer circumference of the tire does not change.

If my speedometer say's I'm going 19.2 mph, I expect that I'm actually traveling at 19.2 mph and not 19.0!

I get really pissed when I find out it's wrong. (really, I do).

So now let me describe how I calibrate my computer. Tell me if this is accurate or not.

I set it according to the instructions, Sheldon Brown's numbers or even the default number it comes at. Just to get a baseline.

Then I ride the bike for a few miles over a course that I've measured on Google Earth. I generally use the Yolo Causeway because it's almost perfectly flat and straight and 3.11 miles from levee to levee. It's also on my way to work. As I ride, I try to stay as straight and centered as I can.

Then I take the distance my computer reads and divide that number by the distance I rode. Then I multiply that number by whatever number I told my computer was the size of my tire.

I did this today, actually. I had entered 2136 as my tire size. I rode the causeway. At the other end, it said I rode 3.15 miles. 3.11/3.15=0.99 .99x2136=2115 (my new wheel size).

Ow! My head hurts...........:eek:

I removed the computers off my bikes years ago.

Who really wants to know how slow they are?

I'm curious about how you folks that are using Google Earth know that it's accurately calibrated? Let alone the fact that a three figure working distance implies at best a limited accuracy. For example 3.11 miles plus or minus .01 miles implies an error of plus or minus 53 feet. And that's only good if you can actually trust that Google Earth is accurately calibrated to that level. It's also allowing for up to a possible 3% error over that distance even if you CAN trust it to be accurate to the two decimal places. If we want to be truly anal about our input setting for these things then I'm sorry but that is not good enough.

At that point I'd rather trust someone marking my valve position, roll for three revolutions while sitting in the saddle with my weight in the usual riding position and regularly used tire pressure and then use a steel tape to measure the distance to the nearest 1/16 inch. At that point I'd be working with roughly a 3360mm distance plus or minus the allowable 1.6 mm (1/16 inch) error. That's a possible plus or minus error of only .05% and I've got far more control over the results. Now THAT is being suitably OCD about the whole exercise.... :D

Nope, radius is tied to circumference only for a circle. Once you put the tire on the ground it is no longer a circle.

I agree with njkayaker. As the wheel rolls the distance traveled is the effective circumference of circumscribed by the radius from the center of the wheel's axis and the weighted wheel's contact patch on the road. On a smooth road this is a perfect circle, 2 r pi, where r is the effective (weighted radius).

I'm curious about how you folks that are using Google Earth know that it's accurately calibrated? Let alone the fact that a three figure working distance implies at best a limited accuracy. For example 3.11 miles plus or minus .01 miles implies an error of plus or minus 53 feet. And that's only good if you can actually trust that Google Earth is accurately calibrated to that level. It's also allowing for up to a possible 3% error over that distance even if you CAN trust it to be accurate to the two decimal places. If we want to be truly anal about our input setting for these things then I'm sorry but that is not good enough.

At that point I'd rather trust someone marking my valve position, roll for three revolutions while sitting in the saddle with my weight in the usual riding position and regularly used tire pressure and then use a steel tape to measure the distance to the nearest 1/16 inch. At that point I'd be working with roughly a 3360mm distance plus or minus the allowable 1.6 mm (1/16 inch) error. That's a possible plus or minus error of only .05% and I've got far more control over the results. Now THAT is being suitably OCD about the whole exercise.... :D

I don't put a lot of faith into Google Earth being deadly accurate, either.

When I measured a tire's rollout recently, I made sure to only do one revolution, so as not to display OCD behavior.

If my speedometer say's I'm going 19.2 mph, I expect that I'm actually traveling at 19.2 mph and not 19.0!

I get really pissed when I find out it's wrong. (really, I do).

So now let me describe how I calibrate my computer. Tell me if this is accurate or not.

I set it according to the instructions, Sheldon Brown's numbers or even the default number it comes at. Just to get a baseline.

Then I ride the bike for a few miles over a course that I've measured on Google Earth. I generally use the Yolo Causeway because it's almost perfectly flat and straight and 3.11 miles from levee to levee. It's also on my way to work. As I ride, I try to stay as straight and centered as I can.

Then I take the distance my computer reads and divide that number by the distance I rode. Then I multiply that number by whatever number I told my computer was the size of my tire.

I did this today, actually. I had entered 2136 as my tire size. I rode the causeway. At the other end, it said I rode 3.15 miles. 3.11/3.15=0.99 .99x2136=2115 (my new wheel size).


Ow! My head hurts...........:eek:

Actually it's easier, I think, than having somebody hold you on your bike while you do a rollout test.

Basically, I just stop, reset my computer, ride the set distance, remember the number at the other end of the course and then sit down with a calculator and do the math when I get to work. No more complicated than what everyone does at the gas station after they fill up and want to know what mpg their car is getting.

I removed the computers off my bikes years ago.

Who really wants to know how slow they are?

That's a good point but I like to know how far I've been, and how many miles my chain stays within tolerance.

I don't have room at my house for much of a rollout test (gravel driveway, gravel street) so I tried Google Maps. I didn't know if it would work until I tried it. All I want out of my speedometer is for my Garmin log to match Google Maps or MapMyRide, and it works well enough for that. (My Garmin is a F50 = not a GPS)

The weighted/unweighted rollout could be it's own thread but the tire is really not a circle when it's on the ground - it has a flat spot. Tire deflection does not change the circumference except possibly at the "sharp" corners where the tires turns from round to flat. I'm anal about that because I got called out in a class presentation in kinematics about 15 years ago. I was measuring torque/power transfer through a bicycle drivetrain or something like that and I used the distance from the axle to the ground instead of the rollout for one of the calculations. I don't remember the project but I do remember when I saw what I did, while standing there at the projector. :)

In the data I report above, it's worse than 1%. The difference between 2100 (low value) and 2146 (high value) is 2%. That's an even bigger deal, and as you say, below the level of accuracy I want in measurement.
Now, I could arguably throw out the 2146 from analysis (since it's such a far outlier, we could just say it's erroneous, and even though it showed up in two user's manuals, those are both Performance in-house brands so they presumably were working from the same chart). And we might also throw out the low value (2100) since it is from the mid-90's whereas everything else is mid-2000's and since. Then the low value is 2105 and high value is 2124, which is still a variation of 0.90%

In the past I've ridden a lap or two on a running track (400m or 800m) with the speedometer set in km, to try to nail exactly 0.40km or 0.80km. As long as I'm not weaving, that's about the best roll-out measurement I could hope for. (unless the tire behaves differently on a rubberized track than it does on pavement)

If you calculate the diameter of the wheel from the numbers above, subtract 622mm from it (the diameter of 700C wheels), and divide by 2, you'll get a tire height. The values range from 23mm for the Cateye to 30mm for the Axiom. Notice I said 'tire height' and not tire diameter. While most tires are roundish, there are some that have a higher profile in the middle so tire height can vary some.

It would also depend on how they measured the tire height for use in the computer. I can envision the computer designer sending Egor out to the warehouse to measure tires. Egor grabs a tire...while not paying attention to the pile of tires he grabs it from, holds the tire between thumb and forefinger (squeezing slightly) and measures the diameter with his calipers. 'Oh', he says, 'It's 30 mm." He never realizes that he is pinching the tire up.

Or maybe the evil computer designer was bad at math and decided that 30mm tire diameter was a nice round number. Or perhaps he rode his commuting bike to work the day he needed the tire diameters and just took that one as the general measurement for 700C tires.

For calibrating the computer yourself, you don't need to ride around a track. Pump the tires up, mark the tire with something...a dab of paint on the tire does nicely...then, with you on the bike, roll the bike in a straight line for one or two revolutions. (If you are really picky, do it for 3 or 4 but then you are just being anal about it:rolleyes:) Measure the distance between the spots and there you have your circumference to about as accurate as anyone could possibly need.:thumb:

I have no idea where the 2146 could have come from but, as you say, it's probably an outlier or a typo. All of the Cat-Eyes I've seen use 210 - 211 cm for 700x25 tires and that's calculated from the circular cross section assumption I mentioned above.
Actually the entire chart for those two speedometers (E3 and Axiom) were high outliers - it's a consistent bias across different tire sizes on the chart. Somebody was probably just new to the office.



I set it according to the instructions, Sheldon Brown's numbers or even the default number it comes at. Just to get a baseline.

I didn't know Sheldon had numbers. Just looked it up. Here:
http://www.sheldonbrown.com/cyclecomputer-calibration.html

Wow, I never would have thought we'd have this long a debate on such a minor issue. You don't even need a helper to do a simple weighted rollout that will be accurate w/in about .5%. Find a straight line on pavement - an expansion joint in concrete will do, or lay a metal tape measure down. Mark the pavement and put your valve stem directly over the mark (OK, mark the tire and pavement and line them up if you're REALLY obsessive. Lean on the bars as you roll the bike parallel to the straight line until your valve stem/mark is at the bottom again. Measure the distance and convert to metric if necessary. Done.

Wrong! The original calculation was correct at 1% or 1 mile per 100. Check your math.
I misread one of the numbers. Do you normally criticize the least important part of a post?


The point is that he has to measure his actual tire.

The chart doesn't know what tire he has.


Since all 25mm tires are not all the same and things like pressure and load produce different effective circumferences, the best thing to do is measure your setup (with you and your load on the bike). People might prefer entering a number from a chart rather than doing the real measurement.

Nope, radius is tied to circumference only for a circle. Once you put the tire on the ground it is no longer a circle. I'm not saying there is no "stretching" of the outer part of the tire, because in practice there is a very slight amount - with a flat steel tape around a tire I can feel it but not measure it.

The only part of the tire that matters is the point where the perpendicular line from the center of the hub intersects the ground. This is the radius of the virtual circle whose circumference the computer needs to compute speed/distance.

The fact that the actual tire isn't truly a circle doesn't matter at all.


But for the most part the deflection is in the cross section getting wider, and the outer circumference of the tire does not change.
Since the tire is deformed at the bottom, the tire isn't a circle. Therefore, circumference is the wrong word. It's the distance from the center of the hub to the ground that changes and the circle that this is the radius-of changes too.


The weighted/unweighted rollout could be it's own thread but the tire is really not a circle when it's on the ground - it has a flat spot. Tire deflection does not change the circumference except possibly at the "sharp" corners where the tires turns from round to flat. I'm anal about that because I got called out in a class presentation in kinematics about 15 years ago.
If you tied a string around the midline of the tire, the length of the string would not change as the pressure or the load on the tire was changed. I suppose it's this measurement you keep calling (incorrectly) "circumference". No one cares about that measurement! That is not the circumference value that the computer needs.


I was measuring torque/power transfer through a bicycle drivetrain or something like that and I used the distance from the axle to the ground instead of the rollout for one of the calculations.
It's this distance that is really the required piece of information (it's the radius of a virtual circle). The computers could certainly take this measurement but they happen to take circumference instead!

=====================


I agree with njkayaker. As the wheel rolls the distance traveled is the effective circumference of circumscribed by the radius from the center of the wheel's axis and the weighted wheel's contact patch on the road. On a smooth road this is a perfect circle, 2 r pi, where r is the effective (weighted radius).
Yup.

Taxi drivers in Las Vegas were busted for letting the pressure down to increase the recorded meter readings.
A bylaw was written to state a legal minimum.

As long as we are obsessing over minutia, has anyone measured the actual change in radius between a loaded and unloaded 700x23 tire at say 110 psi?

Equivalently, what has anyone found as the difference between the published and/or calculated "circumference" of a 700x23 tire (210 cm) and the actual value measured from a properly conducted (i.e. weighted bike) rollout?

As long as we are obsessing over minutia, has anyone measured the actual change in radius between a loaded and unloaded 700x23 tire at say 110 psi?

Equivalently, what has anyone found as the difference between the published and/or calculated "circumference" of a 700x23 tire (210 cm) and the actual value measured from a properly conducted (i.e. weighted bike) rollout?

Has anybody found a difference in the roll-out measurements between the two?

For a 25mm tire, the OP talked about two circumference numbers: 2124 and 2146. The corresponding radii are 338.05 and 341.55. The difference is 3.5 mm (or about 1%).

It seems that the difference between an unloaded and loaded tire might be about 1-2 mm.

It's this distance that is really the required piece of information (it's the radius of a virtual circle). The computers could certainly take this measurement but they happen to take circumference instead!

No, you have it backwards. Virtual circles have nothing to do with it. It's how far the tire moves in a revolution, which is the distance of the outer edge of the tire. This distance does not change with a change in the height of the axle. I can explain it to you but I can't understand it for you. But that's okay because we both have our computers set the way we want them.

No, you have it backwards. Virtual circles have nothing to do with it. It's how far the tire moves in a revolution, which is the distance of the outer edge of the tire. This distance does not change with a change in the height of the axle. I can explain it to you but I can't understand it for you. But that's okay because we both have our computers set the way we want them.


The overall outer circumference of the tire may not change because when you load down and deform the contact patch the rest of the tire shifts slightly to accomadate it. The cords in the tire and the air pressure in the tube will see to that. But the deformation at the contact patch slightly reduces the radius right at the point that truly matters. There will now be a virtual circle's radius that corresponds to the LOADED rollout distance of the tire and THAT circumferance is what would be inputted into the computer.

The iron clad test would be to do rollouts for both loaded and unloaded tires and see if there truly is a difference and how much difference there is.

No, you have it backwards. Virtual circles have nothing to do with it. It's how far the tire moves in a revolution, which is the distance of the outer edge of the tire. This distance does not change with a change in the height of the axle. I can explain it to you but I can't understand it for you. But that's okay because we both have our computers set the way we want them.

It's the distance from the point in the center of the contact patch and the center of the hub. With less pressure or more load, this one point is closer to the rim (and the hub). (We don't care what is happening to the "outer edge" of the tire anywhere else except at this one, single point in the middle of the contact patch.)

The actual distance is between the rim radius and the radius of the unloaded inflated tire. If you remove the tire from the rim and measure the roll-out (circumference) or the radius, you have modeled the case with "excessive load" or "zero pressure".

http://sheldonbrown.com/cyclecomputer-calibration.html#rollout


For cyclecomputers that require a radius value, divide the result by 6.2832 (2 x π) to get the radius.

===================


But the deformation at the contact patch slightly reduces the radius right at the point that truly matters. There will now be a virtual circle's radius that corresponds to the LOADED rollout distance of the tire and THAT circumferance is what would be inputted into the computer.
Yup.

Measure that radius. Create a perfectly rigid tire with that radius. Roll it out and measure it. Roll-out distance: C = pi()*2*r!

Has anybody found a difference in the roll-out measurements between the two?

It seems that the difference between an unloaded and loaded tire might be about 1-2 mm.

The difference between weighted and unweighted is more important for those of us with rear wheel sensors because the rear tire deforms significantly more than the front. Rear wheel computers need a smaller circumference calibration than front wheel computers.

The difference between weighted and unweighted is more important for those of us with rear wheel sensors because the rear tire deforms significantly more than the front. Rear wheel computers need a smaller circumference calibration than front wheel computers.

Yup, that makes sense. You can compensate for the larger load by using higher pressure in the rear.

Sheldon Brown says there's another problem with using the rear wheel:


Unless you want to count "miles" ridden on a stationary trainer, it is best if you measure the roll-out of the front wheel and mount the computer sensor there. The rear wheel "creeps" along the road surface as you pedal, and can skid during braking, so it gives a less-accurate readout.

http://sheldonbrown.com/cyclecomputer-calibration.html#rollout

Okay guys, I'm going to quit smoking crack now. The rollout test proved me wrong. I had to think of it in terms of the edge of the rim, and no matter how much tire is there, for a given tire, pressure, and load the rim is mostly a constant distance from the ground.

My only consolation is that I was right about it back in college, but I've lived a life of lies ever since.

Yup, that makes sense. You can compensate for the larger load by using higher pressure in the rear.

Sheldon Brown says there's another problem with using the rear wheel:



http://sheldonbrown.com/cyclecomputer-calibration.html#rollout

I carry about 10 psi more in the rear tire to compensate for the difference in weight.
I wish Sheldon was still with us. I would argue with him that the rear wheel more accurately measures distance traveled because it runs a straighter path. The trick is to input the best circumference calibration for the wheel. As said in my earlier post, I calibrate to a one mile surveyed distance.

Well i changed both my tires in June on different days and at different LBSs. So i ended up with different tires. haha
I was using the bookchart indicating about 2230 mm and it seemed high.
So today i put masking tape on both tires and rolled out beside my metric tape.

The front Conti TP 37mm came to 2155, and seemed to be exactly right on my ride to the DQ, where i go all the time. 1.07 miles.
The back Schwalbe Marathon 35mm came to 2170.
That is backwards and way off accoring to the charts.

I think what you really want to measure is speed (getting from point A to point B), and yet you're all measuring velocity.

Fretting over 1-2% speedo's (the mis-named instrument) calibration really isn't going to solve the issue.

I think what you really want to measure is speed (getting from point A to point B), and yet you're all measuring velocity.

Fretting over 1-2% speedo's (the mis-named instrument) calibration really isn't going to solve the issue.

"Speedometer" is correct.

People are not generally interested in knowing the velocity (speed in a particular direction).

http://www.physicsclassroom.com/class/1dkin/u1l1d.cfm


Velocity is a vector quantity. As such, velocity is direction aware. When evaluating the velocity of an object, one must keep track of direction. It would not be enough to say that an object has a velocity of 55 mi/hr. One must include direction information in order to fully describe the velocity of the object. For instance, you must describe an object's velocity as being 55 mi/hr, east. This is one of the essential differences between speed and velocity. Speed is a scalar quantity and does not keep track of direction; velocity is a vector quantity and is direction aware.

....I wish Sheldon was still with us. I would argue with him that the rear wheel more accurately measures distance traveled because it runs a straighter path. The trick is to input the best circumference calibration for the wheel. As said in my earlier post, I calibrate to a one mile surveyed distance.

Nope, Sheldon, bless his soul, was right. Any pneumatic driving wheel will have some creep of the contact patch when it's providing an accerative force between the bike and the pavement. And since the back wheel spends far more time under acceleration and driving load at speed than the front does in braking the more accurate numbers will come from the front. The slight difference in the longer path covered by the front in slow speed turns is an occasional and minor thing compared to the contact patch slippage that would occur on even one hillclimb. And once up to any sort of decent speed where something more than a 10 degree lean is encountered in the turns the two wheels are so close to following the same track that again any error is insignificant.

Just had a thought that this would be a stupidly easy experiment to try. Mark off the effective loaded circumference for the two wheels at a given pressure. Mount odometers on each wheel calibrated to the results from the rollout test and go ride for a while. See if they match or one records more mileage than the other. Sort of maintain a rough log of the routes with notes on hills and the like. It's a VERY easy question to answer with real results.

Nope, Sheldon, bless his soul, was right. Any pneumatic driving wheel will have some creep of the contact patch when it's providing an accerative force between the bike and the pavement. And since the back wheel spends far more time under acceleration and driving load at speed than the front does in braking the more accurate numbers will come from the front. The slight difference in the longer path covered by the front in slow speed turns is an occasional and minor thing compared to the contact patch slippage that would occur on even one hillclimb. And once up to any sort of decent speed where something more than a 10 degree lean is encountered in the turns the two wheels are so close to following the same track that again any error is insignificant.

What is an accerative force?
I'm not about to believe that my rear wheel accelerates more than my front. I don't ride off-road. And I don't skid my rear wheel. I use both brakes.
On any normal ride the front wheel takes a longer path than the rear. How do you know how minor that is?
Speed and distance accuracy is going to depend on getting the right number in the computer.

Just had a thought that this would be a stupidly easy experiment to try. Mark off the effective loaded circumference for the two wheels at a given pressure. Mount odometers on each wheel calibrated to the results from the rollout test and go ride for a while. See if they match or one records more mileage than the other. Sort of maintain a rough log of the routes with notes on hills and the like. It's a VERY easy question to answer with real results.

Great idea! Please report the results you get!! ;)

Ya, right.... I'll get right on that...

I just moved and all my shop stuff, including ALL my bicycle parts, is packed in boxes in the garage pending the basement being renovated into my ideal retirement workshop. It'll be the end of October before the stuff starts coming out of the garage and gets organized. At this point the inside of a can of King Oscar's finest sardines looks downright spacious and parklike compared to my garage.

The good news is that the woodworking and model airplane area in the basement will have laminate flooring (it's 3/4 done anyway so I'm going with the flow), T-bar ceiling with T-bar light fixtures and some spot pots and most of the walls will be kitchen cabinet like "counters" for benches with storage cupboards above and full height cabinets for other storage. In the garage will be the metal working area with room for the bicycle working, motorcycle working and welding. Again kitchen cabinet like counters with drawer lowers and cupboard uppers will line the walls where there isn't a machine in the way.

So the bottom line is that if you folks want an answer to that one it'll have to come from a different corner of the country. Besides, for all that I'm posting here I don't remember the last time I had a speedometer on any of my bikes. I've still got a couple but I'm sure the batteries died years ago. Now I just ride until I bonk or arrive home... :D

Ya, right.... I'll get right on that...

I just moved and all my shop stuff, including ALL my bicycle parts, is packed in boxes in the garage pending the basement being renovated into my ideal retirement workshop. It'll be the end of October before the stuff starts coming out of the garage and gets organized. At this point the inside of a can of King Oscar's finest sardines looks downright spacious and parklike compared to my garage.

The good news is that the woodworking and model airplane area in the basement will have laminate flooring (it's 3/4 done anyway so I'm going with the flow), T-bar ceiling with T-bar light fixtures and some spot pots and most of the walls will be kitchen cabinet like "counters" for benches with storage cupboards above and full height cabinets for other storage. In the garage will be the metal working area with room for the bicycle working, motorcycle working and welding. Again kitchen cabinet like counters with drawer lowers and cupboard uppers will line the walls where there isn't a machine in the way.

So the bottom line is that if you folks want an answer to that one it'll have to come from a different corner of the country. Besides, for all that I'm posting here I don't remember the last time I had a speedometer on any of my bikes. I've still got a couple but I'm sure the batteries died years ago. Now I just ride until I bonk or arrive home... :D

Sounds like all we are getting is excuses!