Accurate Odometer

Zog The Undeniable wrote:


They're all perfectly accurate if calibrated correctly. Assuming you're
in America, do the highways have mileposts? You could set the computer
for the wheel circumference specified in the manual (probably a bit
inaccurate) then ride a few measured miles and work out the % error,
finally entering the real rolling circumference into the computer.

I must do this myself at some point; many UK dual carriageways have
posts every 100 metres (why it's metric I have no idea; all our road
signs are in miles).

They should all be very repeatable if working properly, and very
accurate if calibrated properly. Some models let you enter the
circumference to the nearest mm, and some only to the nearest cm.
The Cateye Mity 3, for instance, in most respects a fine computer,
only uses the nearest cm. For me, a 700 x 23 typically needs about
2095 mm, so I can pick 209 cm and have it read a little low or 210
and have it read high. One to avoid is the Planet Bike protege 9.0,
which takes calibration numbers to the nearest mm, and then either
rounds off to the nearest cm or some internal distance unit. I've
tried to tweak one in to high accuracy, and found that 1mm changes
made no difference, until a point when it made a too-large
difference.

Dave Lehnen

Do a web search for "Sigma BC800" and you'll find some hits link to
car and motorcycle rally sites. There you will find that the cheap
$15 bicycle computer is used on cars and motorcycles when they need
accurate distances. Here's one:
http://mywebpages.comcast.net/billingsley/sigma.htm
A few sites even have tables comparing accuracy of the bicycle
computer to the car and motorcycle stock and aftermarket odometers and
the results are pretty surprising.

Carl Fogel writes:

As a cross-check, my $15 Nashbar odometer varies perhaps 0.03
miles on either side of 15.25 miles on my daily ride, which may
say more about how much I swerve and how tight I cut my corners
than it says about the accuracy of the speedometer.

Or variation in tire pressure

Possibly tire pressure is involved, but unlikely, since the tiny
odometer variation doesn't seem to go in one direction.

That is, I don't see my distance slowly increasing (or decreasing)
until I remember to pump my tires up again.

The variation could also be due to weight changes, since the
standard Fogel often varies by several chocolate doughnuts.

Don't rely on others reports if you don't believe them. Just measure
the rollout distance with different inflation pressures on your own
bicycle. I have done it and it isn't just a millimeter. I'm
satisfied that 2093mm is good enough for my Avocet 700-25 tires at
average inflation. I don't care if it changes a little with inflation
because the tires hold air to my satisfaction for more than a month.

Many posters here on rec.bicycles.tech warn of the dangerous
inaccuracy of measuring an unloaded wheel, but I've never seen a
post that mentioned any measured difference in tire circumference
between a loaded and an unloaded bike. (Indeed, the warnings often
fail to mention which way they think the loading affects the
outcome.)

What dangers have been reported? As inflation pressure decreases, so
does the rollout distance for one revolution. As I pointed out, when
driving a car over Botts Dots on roads you'll notice the slamming
effect get harsher with increasing speed. That is because the change
in rolling radius causes a momentary acceleration of the wheel that
through inertia cannot occur at higher speeds as easily. Therefore,
the slamming effect bends wheel suspension. The bumps do not get
larger nor does wheel bounce but the the rolling radius changes almost
instantaneously. Since most car tires are radials (with a constant
circumference belt), I find that a dramatic demonstration of rolling
radius.

I wouldn't be surprised if the difference was more theoretical than
measurable, but it would fun if someone posted the results of some
careful loaded and unload tire measurements.

Be surprised!

Jobst Brandt

On Sat, 12 Feb 2005 19:07:35 -0800, "Dan"
wrote:


wrote in message
.. .
On Sat, 12 Feb 2005 17:29:07 -0800, "Dan"
wrote:


wrote in message
.. .

As a cross-check, my $15 Nashbar odometer varies perhpas
0.03 miles on either side of 15.25 miles on my daily ride,
which may say more about how much I swerve and how tight I
cut my corners than it says about the accuracy of the
speedometer.


Carl Fogel

Or variation in tire pressure

Dear Dan,

Possibly tire pressure is involved, but unlikely, since the
tiny odometer variation doesn't seem to go in one direction.

That is, I don't see my distance slowly increasing (or
decreasing) until I remember to pump my tires up again.

The variation could also be due to weight changes, since the
standard Fogel often varies by several chocolate doughnuts.

Many posters here on rec.bicycles.tech warn of the dangerous
inaccuracy of measuring an unloaded wheel, but I've never
seen a post that mentioned any measured difference in tire
circumference between a loaded and an unloaded bike.
(Indeed, the warnings often fail to mention which way they
think the loading affects the outcome.)

I wouldn't be surprised if the difference was more
theoretical than measurable, but it would fun if someone
posted the results of some careful loaded and unload tire
measurements.

Carl Fogel

Carl-

Since the computer is really just an event counter, I can (offhand) think of
three possible errors: path wobble, the distance between events and the
event count. Thinking about this, I wouldn't be surprised to discover that
all 3 errors occur during a 15 mile ride. I seem to have about 1% confidence
in my odometer. 1% of a 26-inch wheel is about 1/4 inch which correlates to
a 1/8-inch in ride height variation due to tire pressure variation (or the
donut load varies - I don't like the doughy ones). A 1% distance error also
amounts to about a 0.5 degree error in bearing - easy enough error for a
doughnut man. I am sure that the count could be off +/- due to any number of
influences.

In my engineering world I have come to the conclusion that for most
applications, 2 significant figure work is good enough. 2 significant figure
work assumes about 5% error. In some work I aspire to a higher standard and
try to achieve 3 significant figure accuracy, about 1%. Anything beyond
3-digit work requires a level of attention to detail that simply doesn't pay
off for practical work. Depending on how much money I have, my checkbook
significant figure accuracy varies wildly.

My Dad was a field surveys topographer for USGS. He did the fieldwork for
the topographic maps we all know. They aspired to do highly precise
measurements but it was a difficult task and required knowledge of many
natural and human influences along with a sound foundation in statistics and
probability. Not worth it in bike ride distance measurements.

We could go beyond the Newtonian world and discuss 4-dimensional space-time
and the influence of mass and velocity but I don't think we should.

Dan

Dear Dan,

To cut to the chase . . .

My cheap cyclocomputers come up with the same distance for a
15.25 mile daily ride +/ 0.03 miles, or +/- 0.2% accuracy.

Carl Fogel

I wrote:

Don't rely on others reports if you don't believe them. Just
measure the rollout distance with different inflation pressures on
your own bicycle. I have done it and it isn't just a millimeter.
I'm satisfied that 2093mm is good enough for my Avocet 700-25 tires
at average inflation. I don't care if it changes a little with
inflation because the tires hold air to my satisfaction for more
than a month.

Make that: 2096 for an Avocet Road 700-25c tire.

Jobst Brandt

On Sun, 13 Feb 2005 05:32:58 GMT,
wrote:

Carl Fogel writes:

As a cross-check, my $15 Nashbar odometer varies perhaps 0.03
miles on either side of 15.25 miles on my daily ride, which may
say more about how much I swerve and how tight I cut my corners
than it says about the accuracy of the speedometer.

Or variation in tire pressure

Possibly tire pressure is involved, but unlikely, since the tiny
odometer variation doesn't seem to go in one direction.

That is, I don't see my distance slowly increasing (or decreasing)
until I remember to pump my tires up again.

The variation could also be due to weight changes, since the
standard Fogel often varies by several chocolate doughnuts.

Don't rely on others reports if you don't believe them. Just measure
the rollout distance with different inflation pressures on your own
bicycle. I have done it and it isn't just a millimeter. I'm
satisfied that 2093mm is good enough for my Avocet 700-25 tires at
average inflation. I don't care if it changes a little with inflation
because the tires hold air to my satisfaction for more than a month.

Many posters here on rec.bicycles.tech warn of the dangerous
inaccuracy of measuring an unloaded wheel, but I've never seen a
post that mentioned any measured difference in tire circumference
between a loaded and an unloaded bike. (Indeed, the warnings often
fail to mention which way they think the loading affects the
outcome.)

What dangers have been reported? As inflation pressure decreases, so
does the rollout distance for one revolution. As I pointed out, when
driving a car over Botts Dots on roads you'll notice the slamming
effect get harsher with increasing speed. That is because the change
in rolling radius causes a momentary acceleration of the wheel that
through inertia cannot occur at higher speeds as easily. Therefore,
the slamming effect bends wheel suspension. The bumps do not get
larger nor does wheel bounce but the the rolling radius changes almost
instantaneously. Since most car tires are radials (with a constant
circumference belt), I find that a dramatic demonstration of rolling
radius.

I wouldn't be surprised if the difference was more theoretical than
measurable, but it would fun if someone posted the results of some
careful loaded and unload tire measurements.

Be surprised!

Jobst Brandt


Dear Jobst,

Er, you need to read more carefully. Your unreported
"dangers" are not the same as my rather obviously
tongue-in-cheek "dangerous inaccuracy."

I'd be more surprised if you were to include the figures
that you say are dramatic. I can't even tell from your post
whether you are saying that the loaded tire produces a
larger or smaller circumference. I'd be willing to believe
your figures if you'd simply post them.

You find 2093mm a good average figure for your tires,
presumably measured loaded. What was the unloaded
measurement? It "isn't just a millimeter."

If you reply with a 4-character post, it will be the first
time that I know of an actual difference appearing in this
newsgroup.

I hope it's not a secret.

Cheers,

Carl Fogel

On Sun, 13 Feb 2005 06:10:52 GMT,
wrote:

I wrote:

Don't rely on others reports if you don't believe them. Just
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
measure the rollout distance with different inflation pressures on
your own bicycle. I have done it and it isn't just a millimeter.
I'm satisfied that 2093mm is good enough for my Avocet 700-25 tires
at average inflation. I don't care if it changes a little with
inflation because the tires hold air to my satisfaction for more
than a month.

Make that: 2096 for an Avocet Road 700-25c tire.

Jobst Brandt


Dear Jobst,

Hmmm . . . so would I have been wrong to rely on your 2096mm
figure?

I scarcely know what to believe now.

Is the difference between a loaded and an unloaded tire's
circumference greater than the difference between 2093 mm
and 2096 mm?

Cheers,

Carl Fogel

To cut to the chase . . .

My cheap cyclocomputers come up with the same distance for a
15.25 mile daily ride +/ 0.03 miles, or +/- 0.2% accuracy.

Carl Fogel

Carl: Over a distance of 15.25 miles, a variation of 158 feet doesn't seem
like more than might be accounted for due to normal minor deviations during
the ride. For example, do you consistently approach and pass through each
intersection in the exact same manner, regardless of the presence of traffic
or the condition (color) of the light? Might there be a stop sign you come
up to where, with cross-traffic present, you might make a right turn for a
bit and then cut across when safe to do so?

Have you made comparisons between average speed & distance traveled?

--Mike-- Chain Reaction Bicycles
www.ChainReactionBicycles.com


wrote in message
...
On Sat, 12 Feb 2005 19:07:35 -0800, "Dan"
wrote:


wrote in message
. ..
On Sat, 12 Feb 2005 17:29:07 -0800, "Dan"
wrote:


wrote in message
.. .

As a cross-check, my $15 Nashbar odometer varies perhpas
0.03 miles on either side of 15.25 miles on my daily ride,
which may say more about how much I swerve and how tight I
cut my corners than it says about the accuracy of the
speedometer.


Carl Fogel

Or variation in tire pressure

Dear Dan,

Possibly tire pressure is involved, but unlikely, since the
tiny odometer variation doesn't seem to go in one direction.

That is, I don't see my distance slowly increasing (or
decreasing) until I remember to pump my tires up again.

The variation could also be due to weight changes, since the
standard Fogel often varies by several chocolate doughnuts.

Many posters here on rec.bicycles.tech warn of the dangerous
inaccuracy of measuring an unloaded wheel, but I've never
seen a post that mentioned any measured difference in tire
circumference between a loaded and an unloaded bike.
(Indeed, the warnings often fail to mention which way they
think the loading affects the outcome.)

I wouldn't be surprised if the difference was more
theoretical than measurable, but it would fun if someone
posted the results of some careful loaded and unload tire
measurements.

Carl Fogel

Carl-

Since the computer is really just an event counter, I can (offhand) think
of
three possible errors: path wobble, the distance between events and the
event count. Thinking about this, I wouldn't be surprised to discover that
all 3 errors occur during a 15 mile ride. I seem to have about 1%
confidence
in my odometer. 1% of a 26-inch wheel is about 1/4 inch which correlates
to
a 1/8-inch in ride height variation due to tire pressure variation (or the
donut load varies - I don't like the doughy ones). A 1% distance error
also
amounts to about a 0.5 degree error in bearing - easy enough error for a
doughnut man. I am sure that the count could be off +/- due to any number
of
influences.

In my engineering world I have come to the conclusion that for most
applications, 2 significant figure work is good enough. 2 significant
figure
work assumes about 5% error. In some work I aspire to a higher standard
and
try to achieve 3 significant figure accuracy, about 1%. Anything beyond
3-digit work requires a level of attention to detail that simply doesn't
pay
off for practical work. Depending on how much money I have, my checkbook
significant figure accuracy varies wildly.

My Dad was a field surveys topographer for USGS. He did the fieldwork for
the topographic maps we all know. They aspired to do highly precise
measurements but it was a difficult task and required knowledge of many
natural and human influences along with a sound foundation in statistics
and
probability. Not worth it in bike ride distance measurements.

We could go beyond the Newtonian world and discuss 4-dimensional
space-time
and the influence of mass and velocity but I don't think we should.

Dan

Dear Dan,

To cut to the chase . . .

My cheap cyclocomputers come up with the same distance for a
15.25 mile daily ride +/ 0.03 miles, or +/- 0.2% accuracy.

Carl Fogel

wrote:

My cheap cyclocomputers come up with the same distance for a
15.25 mile daily ride +/ 0.03 miles, or +/- 0.2% accuracy.

I get the same distance to within 1/100 mile when riding to work. It's
always 5.57 miles there and 5.60 miles back. The reason for the
difference is that I have to make a right turn at a roundabout on the
way back, which is the long way round if you're riding on the LH side of
the road.

On Sun, 13 Feb 2005 07:49:10 GMT, "Mike Jacoubowsky"
wrote:

To cut to the chase . . .

My cheap cyclocomputers come up with the same distance for a
15.25 mile daily ride +/ 0.03 miles, or +/- 0.2% accuracy.

Carl Fogel

Carl: Over a distance of 15.25 miles, a variation of 158 feet doesn't seem
like more than might be accounted for due to normal minor deviations during
the ride. For example, do you consistently approach and pass through each
intersection in the exact same manner, regardless of the presence of traffic
or the condition (color) of the light? Might there be a stop sign you come
up to where, with cross-traffic present, you might make a right turn for a
bit and then cut across when safe to do so?

Have you made comparisons between average speed & distance traveled?

--Mike-- Chain Reaction Bicycles
www.ChainReactionBicycles.com

Dear Mike,

The rut that I enjoy riding might bore a normal rider
witless in terms of pure bicycling.

I understand that you fellows expect to see things like stop
signs and intersections with traffic.

On a good day, I do no braking at all in fifteen miles for
anything except turns.

On an average day, I have to stop at the sole traffic light
on my daily ride instead of rolling right through.

On a bad day, I have to slow down a little for some damned
car, but there's no swerving involved.

I can't think of any place where my line varies as much as
three feet to either side, other than when I pass cars in a
20 mph zone on a one-way street through the city park.

It's so ridiculously regular that I've thought that slower
days have led to slightly longer odometer readings and
theorized that at lower speeds, there's more swerving.

Unlike the distance, the scenery varies quite a bit. So does
the wildlife--squirrels, rabbits, foxes, beaver, muskrat,
prairie dogs, coyotes, deer, ducks, geese, swans, hawks,
horned owls, screech owls, burrowing owls, bald eagles,
great blue herons, pelicans, cormorants, kingfishers,
flickers, robins, hummingbirds, crows, ravens, roadrunners,
grackles, doves, quail, turkeys, box turtles, snapping
turtles, softshell turtles, garter snakes, bullsnakes,
rattlesnakes, red racers, lizards, toads, bullfrogs,
salamanders, tarantulas, and tarantula hawks.

The badgers, porcupines, bobcats, elk, bear, and cougar are
mostly on the far side of the ridge, so I see them on
weekends when I take my dogs to the countryside.

It's 15.25 miles, give or take a hundred and fifty feet.
Lapsed trials riders tend to follow the same line.

Carl Fogel