Accuracy of cycle computers

[FranklynMint]

Greetings fellow cyclists
I have a mid-price cycle computer, which is set up according to instructions (double-checked and triple checked). On a long ride, recently (recommended by a friend, who said it was about 40+ miles, he hadn't measured it) i was quietly impressed by the distace i rode, which was longer than he'd said. But looking at a map later, I thought something was amiss. I reckon that the computer is about 1 mile out every 15. Is that par for the course with these gadgets? I'm not interested in measuring to the millimetre, nor do I have more than a passing interest in the average speed, etc, but it's nice to have an idea of how far you've done.
Arrivederci
FM

 

[Pete Biggs]

FranklynMint wrote:
> Greetings fellow cyclists I have a mid-price cycle
> computer, which is set up according to instructions (double-
> checked and triple checked). On a long ride, recently
> (recommended by a friend, who said it was about 40+ miles,
> he hadn't measured it) i was quietly impressed by the
> distace i rode, which was longer than he'd said. But
> looking at a map later, I thought something was amiss. I
> reckon that the computer is about 1 mile out every 15. Is
> that par for the course with these gadgets? I'm not
> interested in measuring to the millimetre, nor do I have
> more than a passing interest in the average speed, etc,
> but it's nice to have an idea of how far you've done.
> Arrivederci FM

Modern cycle computers tend to be as accurate as each other
so don't worry about the model. Just make sure it's working
and that the tyre circumference setting is correct. The
result will be true.

Bike ride distances will be greater than map measurements
because you can't measure all the bends properly on a normal
map, and all the little micro-turns and wobbles add up.

~PB

 

[Doctor J. Frink]

On Tue, 15 Jun 2004 10:01:00 GMT, FranklynMint
<[email protected]> wrote:
>Greetings fellow cyclists I have a mid-price cycle
>computer, which is set up according to instructions (double-
>checked and triple checked). On a long ride, recently
>(recommended by a friend, who said it was about 40+ miles,
>he hadn't measured it) i was quietly impressed by the
>distace i rode, which was longer than he'd said. But
>looking at a map later, I thought something was amiss. I
>reckon that the computer is about 1 mile out every 15. Is
>that par for the course with these gadgets? I'm not

That comes out at nearly 7% error, which I would think is
quite large. There are two main sources of error for these
things that I see:

1) The accuracy of the wheel circumference measurement
2) Sensitivity of the magnet + sensor (ie does one
revolution of the wheel register as one revolution in
the sensor)

The wheel circumference (C) is something that you should be
able to get quite accurately with a roll-test (where you sit
normally on the bike, keeping the front wheel straight, mark
the start and end of one (or more) wheel revolutions while a
friend pushes you and measure).

This should get it to within 0.5% at the worst (5mm error on
a 2000mm C wheel is 0.25% error). As C varies with tyre
pressure, ride weight etc this is probably more like 1-2% in
practice. Generally this bit is as accurate as you can make
it (within setting resolution of mostly 1mm on the computer,
ie about 0.001%).

What could be causing you trouble, assuming your C is set as
accurately as you can get it is the magnet sensor set up.
You want it as near to the hub as possible so that the
magnet sweeps past the sensor as slowly as possible. This
gives the sensor longer to react. If you've got it nearer
the rim and you're going at a decent pace the sensor may be
saturated and thus record a different number of revolutions
than there have actually been.

If your sensor works properly your distance should simply be

revolutions x C

so your only significant error lies in C (measurement error
and how it varies as you ride).

>interested in measuring to the millimetre, nor do I have
>more than a passing interest in the average speed, etc, but
>it's nice to have an idea of how far you've done.
>Arrivederci FM

Speed and distance are related, so getting one accurate
means the other should be as well. If you get your C to
within a few mm and get your sensor properly set up you
should easily achieve 1%, which is really quite accurate
when you think about it (1 mile in a hundred). Don't
forget that something that's 10miles in a straight line
might not be 10miles on a bike that weaves a bit (as a
front wheel tends to do a bit) so your distance reading on
the bike may well be accurate even if it doesn't quite
match what the map says.

As we're generally not looking for scientific precision it
doesn't matter overly much. My gf's computer was giving a
few % different to my computer so we tweaked hers to give
the same results. This isn't because mine was any more
accurate but just to make things fair (and hers was more
tweakable).

Frink

--
Doctor J. Frink : 'Rampant Ribald Ringtail' See his mind
here : http://www.cmp.liv.ac.uk/frink/ Annoy his mind here :
pjf at cmp dot liv dot ack dot ook "No sir, I didn't like
it!" - Mr Horse

 

[Mark Tranchant]

FranklynMint wrote:

> But looking at a map later, I thought something was amiss.
> I reckon that the computer is about 1 mile out every 15.
> Is that par for the course with these gadgets?

You need to set it to measure the rolling circumference of
your wheel. I don't know what your instructions say, but the
easiest way is to ride over a line of something like washing-
up liquid and keep going, then measure the distance between
as many lines as you can.

Depending on your computer, divide by the appropriate
combination of 2 and pi if you need the diameter or radius.

1 part in 15 is just under 7%: you should be able to measure
your wheel better than that. There are three sources of
inaccuracy:

1) Inaccurate wheel measurement
2) Missed or duplicate wheel counts
3) Precision (rounding) errors

These are in order of importance for typical computers: 2)
and 3) are practically negligible if set up properly.

Don't forget that the computer measures the distance
travelled by your front wheel - if you weave, it'll measure
longer distances than the road actually travelled.

--
Mark.

 

[Pete Biggs]

Mark Tranchant wrote:

> Don't forget that the computer measures the distance
> travelled by your front wheel - if you weave, it'll
> measure longer distances than the road actually travelled.

Not just front wheel weaving, but many or most roads
include slight kinks that won't show up on anything other
than very large-scale maps. All that will add up
considerably on a ride.

~PB

 

[FranklynMint]

Thanks for the practical advice, folks. I'll measure the circumference accurately and move the sensor nearer to the hub.
Cheers.

 

[Ian G Batten]

In article <[email protected]>,
FranklynMint <[email protected]> wrote:
> Greetings fellow cyclists I have a mid-price cycle
> computer, which is set up according to instructions (double-
> checked and triple checked). On a long ride, recently
> (recommended by a friend, who said it was about

Over ten miles this morning mine agreed to within about 0.05
miles (~100 yards) with the GPS box I had clipped to the
handlebars. Given they're measuring different things in
different ways, given I was running the GPS at a slow update
rate and given the fact that the computer had recorded 100
yards _more_ I'd say that's about bang on. The more is
reassuring, as the GPS box will be measuring a succession of
short chords to the route.

> But looking at a map later, I thought something was amiss.
> I reckon that the computer is about 1 mile out every 15.

Unless you're talking about riding in a dead straight line
along 15 miles of salt flats, the front wheel will go
further than you think. Once you factor in bends, it goes a
_lot_ further. I don't know id ~6% is typical, but the
difference between GPS tracks and GPS point-to-point
bearings is usually at least that.

ian

 

[Doctor J. Frink]

On Tue, 15 Jun 2004 11:43:00 GMT, Trevor Barton <[email protected]> wrote:
>
>If you have a car drive a fixed distance and then ride the
>same, and then adjust your cycle computer by some
>proportion to make the two read the same. You would need
>top do it over a fair distance, some miles I guess, but
>it's likely that in a modern car with correctly inflated
>tyres the odometer will be pretty accurate.

Here you're introducing an unknown error (ie just how
accurate is your car's odometer, both in the absolute
measurement, its resolution and under what car conditions is
it accurate) along with the fact that a car covers ground in
a different way to a bike (the normal wobble of the bike,
and differing suspension putting different loads on the
wheels which affects wheel circumference). All you're
getting then is the distance you would have covered in your
car, which you haven't done, cos you're on your bike.

Measuring the wheel circumference directly using a rolling
measurement (either directly or on a known length track)
should give you much better accuracy, well under 1% if done
properly which surely puts most car odometers to shame. It's
also quicker and doesn't require an unnecessary long drive
for both car and bike.

I would be more inclined to set the car odometer to the
*bike's* but then that might be illegal ;0).

Frink

--
Doctor J. Frink : 'Rampant Ribald Ringtail' See his mind
here : http://www.cmp.liv.ac.uk/frink/ Annoy his mind here :
pjf at cmp dot liv dot ack dot ook "No sir, I didn't like
it!" - Mr Horse

 

[Doctor J. Frink]

On Tue, 15 Jun 2004 11:50:45 GMT, FranklynMint
<[email protected]> wrote:
>Thanks for the practical advice, folks. I'll measure the
>circumference accurately and move the sensor nearer to the
>hub. Cheers.

Oh, one final thing about moving closer to the hub, if it's
a wireless sensor this may reduce battery life (further away
it is the more power it has ot use to reach the computer),
although it's hard to say by how much (probably not much if
it's only a few cms). I expect if it's well within the
limits of the sensor it shouldn't affect the accuracy.

Frink

--
Doctor J. Frink : 'Rampant Ribald Ringtail' See his mind
here : http://www.cmp.liv.ac.uk/frink/ Annoy his mind here :
pjf at cmp dot liv dot ack dot ook "No sir, I didn't like
it!" - Mr Horse

 

[Ian G Batten]

In article <[email protected]>,
Trevor Barton <[email protected]> wrote:
> some miles I guess, but it's likely that in a modern car
> with correctly inflated tyres the odometer will be pretty
> accurate.

Well...

Take a car running on 175/60R15s (a tyre size I've made up,
but is the sort of thing that a family car might run on).
The circumference of the tyre is 3.14 *
((15*25.4)+(175*0.6)*2) = 1.855m (oh, I love mixed metric
and imperial sizing). But tyres ship with 10mm of tread, and
are safe down to 3mm, so that's 3.14 * ((15*25.4)+(175*0.6-
7)*2) = 1.812m at the other extreme. So there's potentially
+/-1% of error just in the wear in your tyres.

ian

 

[David Off]

Doctor J. Frink wrote:

> On Tue, 15 Jun 2004 11:43:00 GMT, Trevor Barton
> <[email protected]> wrote:
>
>>If you have a car drive a fixed distance and then ride the
>>same, and then adjust your cycle computer by some
>>proportion to make the two read the same. You would need
>>top do it over a fair distance, some miles I guess, but
>>it's likely that in a modern car with correctly inflated
>>tyres the odometer will be pretty accurate.
>
>
> Here you're introducing an unknown error

agreed, my car odo is about 5% accurate. Bike computers can
suffer from contact bounce of the reed switches at higher
speeds so will tend to over read all things being equal.

The council should have an accurate 'measured mile'
somewhere which can be used to calibrate odometers, try the
highways department for information.

 

[Just Zis Guy]

Ian G Batten wrote:

> The circumference of the tyre is 3.14 *
> ((15*25.4)+(175*0.6)*2) = 1.855m (oh, I love mixed
> metric and imperial sizing). But tyres ship with 10mm
> of tread, and are safe down to 3mm, so that's 3.14 *
> ((15*25.4)+(175*0.6-7)*2) =
> 1.812m at the other extreme. So there's potentially +/-1%
> of error just in the wear in your tyres.

And that's before you've driven them for a year without
checking the pressures ;-)

Guy
--
May contain traces of irony. Contents liable to settle after posting.
http://www.chapmancentral.co.uk

 

[Dr Curious]

"Doctor J. Frink" <[email protected]> wrote in message
news:[email protected]...

> As C varies with tyre pressure, ride weight etc this is
> probably more like 1-2% in practice.

This is interesting.

So the outside circumference of an inflated tyre is larger
than that of a flat tyre is it?

Curious

>
> What could be causing you trouble, assuming your C is set
> as accurately as you can get it is the magnet sensor set
> up. You want it as near to the hub as possible so that the
> magnet sweeps past the sensor as slowly as possible. This
> gives the sensor longer to react. If you've got it nearer
> the rim and you're going at a decent pace the sensor may
> be saturated and thus record a different number of
> revolutions than there have actually been.
>
> If your sensor works properly your distance should
> simply be
>
> revolutions x C
>
> so your only significant error lies in C (measurement
> error and how it varies as you ride).
>
> >interested in measuring to the millimetre, nor do I have
> >more than a passing interest in the average speed, etc,
> >but it's nice to have an idea of how far you've done.
> >Arrivederci FM
>
> Speed and distance are related, so getting one accurate
> means the other should be as well. If you get your C to
> within a few mm and get your sensor properly set up you
> should easily achieve 1%, which is really quite accurate
> when you think about it (1 mile in a hundred). Don't
> forget that something that's 10miles in a straight line
> might not be 10miles on a bike that weaves a bit (as a
> front wheel tends to do a bit) so your distance reading on
> the bike may well be accurate even if it doesn't quite
> match what the map says.
>
> As we're generally not looking for scientific precision it
> doesn't matter overly much. My gf's computer was giving a
> few % different to my computer so we tweaked hers to give
> the same results. This isn't because mine was any more
> accurate but just to make things fair (and hers was more
> tweakable).
>
> Frink
>
> --
> Doctor J. Frink : 'Rampant Ribald Ringtail' See his mind
> here : http://www.cmp.liv.ac.uk/frink/ Annoy his mind here
> : pjf at cmp dot liv dot ack dot ook "No sir, I didn't
> like it!" - Mr Horse

 

[Doctor J. Frink]

On Tue, 15 Jun 2004 13:45:16 +0100, Dr Curious <[email protected]>
wrote:
>
>"Doctor J. Frink" <[email protected]> wrote in
>message news:[email protected]...
>
>
>> As C varies with tyre pressure, ride weight etc this is
>> probably more like 1-2% in practice.
>
>
>This is interesting.
>
>So the outside circumference of an inflated tyre is larger
>than that of a flat tyre is it?
>
>Curious

Take bike.

Deflate front tyre.

Wheel descends to floor to rest on rim.

Note distance r_flat from hub to floor (ie the radius of the
circle that we would be rolling on).

Pump up tyre.

Wheel rises from floor as tyre inflates.

Note distance r_inflated from hub to floor (ie the radius of
the circle that we would be rolling on).

You'll find that r_inflated > r_flat.

The actual circumference of the tyre itself may not change
but the effective wheel radius does, and that's what we're
really measuring (ok, labelling it Circumference wasn't
totally accurate but it's close enough).

This is also affected by how you ride the bike, how much you
weigh etc, as under normal pressures where the tyre meets
the ground the tyre deforms perceptibly, meaning r_loaded <
r_unloaded. This is why you must be on the bike when
measuring C.

To summarise: measure how far one wheel revolution takes you
(effectively its circumference) under the conditions you
normally ride.

Frink

--
Doctor J. Frink : 'Rampant Ribald Ringtail' See his mind
here : http://www.cmp.liv.ac.uk/frink/ Annoy his mind here :
pjf at cmp dot liv dot ack dot ook "No sir, I didn't like
it!" - Mr Horse

 

[Mike Dodds]

> >> As C varies with tyre pressure, ride weight etc this is
> >> probably more like 1-2% in practice.

All of which goes to explain something for me. I took the
Cateye Astrale 8 off my tourer and put it onto a new (second
hand road bike) and put an old Halfords 12 function on my
tourer. Both were set with the wheel setting (circumferance)
to 2180 for a 700 x 38. I then did my usual morning commute
and found that instead of 6.35 miles the distance shown was
6.27 miles or abotu 100 yards less.

That must entirely down to my weight bearing down on the
rear tyre (assuming both computers can count and calculate
correctly. The Astale 8 magnet is on the rear wheel, but the
Halfords magnet is on the front wheel.

 

[Dr Curious]

"Doctor J. Frink" <[email protected]> wrote in message
news:[email protected]...
> On Tue, 15 Jun 2004 13:45:16 +0100, Dr Curious
> <[email protected]> wrote:
> >
> >"Doctor J. Frink" <[email protected]> wrote in
> >message news:[email protected]...
> >
> >
> >> As C varies with tyre pressure, ride weight etc this is
> >> probably more like 1-2% in practice.
> >
> >
> >This is interesting.
> >
> >So the outside circumference of an inflated tyre is
> >larger than that of a flat tyre is it?
> >
> >Curious
>
> Take bike.
>
> Deflate front tyre.
>
> Wheel descends to floor to rest on rim.
>
> Note distance r_flat from hub to floor (ie the radius of
> the circle that we would be rolling on).
>
> Pump up tyre.
>
> Wheel rises from floor as tyre inflates.
>
> Note distance r_inflated from hub to floor (ie the radius
> of the circle that we would be rolling on).
>
> You'll find that r_inflated > r_flat.
>
> The actual circumference of the tyre itself may not change
> but the effective wheel radius does, and that's what we're
> really measuring (ok, labelling it Circumference wasn't
> totally accurate but it's close enough).
>
> This is also affected by how you ride the bike, how much
> you weigh etc, as under normal pressures where the tyre
> meets the ground the tyre deforms perceptibly, meaning
> r_loaded < r_unloaded. This is why you must be on the bike
> when measuring C.
>
> To summarise: measure how far one wheel revolution takes
> you (effectively its circumference) under the conditions
> you normally ride.

...

I'm sorry but this all totally irrelevant.

The radius of the wheel is irrelevant.

When a tire is flat, all this means is that the wheel/tyre
is no longer circular. It's now a part circle with a flat at
the bottom. And so there's more surface area in contact with
the ground. But the actual outside circumference remains
exactly the same. It wouldn't even matter if you were riding
on oval wheels. All you ever need measure are the number of
revolutions multiplied by the circumference of the tyre.

Curious

...


>
> Frink
>
> --
> Doctor J. Frink : 'Rampant Ribald Ringtail' See his mind
> here : http://www.cmp.liv.ac.uk/frink/ Annoy his mind here
> : pjf at cmp dot liv dot ack dot ook "No sir, I didn't
> like it!" - Mr Horse

 

[Dave Larrington]

Mike Dodds wrote:

> All of which goes to explain something for me. I took the
> Cateye Astrale 8 off my tourer and put it onto a new
> (second hand road bike) and put an old Halfords 12
> function on my tourer. Both were set with the wheel
> setting (circumferance) to 2180 for a 700 x 38. I then
> did my usual morning commute and found that instead of
> 6.35 miles the distance shown was 6.27 miles or abotu 100
> yards less.
>
> That must entirely down to my weight bearing down on the
> rear tyre (assuming both computers can count and calculate
> correctly. The Astale 8 magnet is on the rear wheel, but
> the Halfords magnet is on the front wheel.

I, however, have a Mity3 and an Astrale, both set to
circumference = 1440 mm and both looking at the same magnet.
Distance and average speed are sometimes considerably
different...

--

Dave Larrington - http://www.legslarry.beerdrinkers.co.uk/
===========================================================
Editor - British Human Power Club Newsletter
http://www.bhpc.org.uk/
===========================================================

 

[James Annan]

Mike Dodds wrote:

>
> All of which goes to explain something for me. I took the
> Cateye Astrale 8 off my tourer and put it onto a new
> (second hand road bike) and put an old Halfords 12
> function on my tourer. Both were set with the wheel
> setting (circumferance) to 2180 for a 700 x 38. I then
> did my usual morning commute and found that instead of
> 6.35 miles the distance shown was 6.27 miles or abotu 100
> yards less.
>
> That must entirely down to my weight bearing down on the
> rear tyre (assuming both computers can count and calculate
> correctly. The Astale 8 magnet is on the rear wheel, but
> the Halfords magnet is on the front wheel.

As has already been mentioned (I think), the front wheel
actually does go slightly further than the rear. Ride
through a puddle and it will be very obvious. Weight
distribution may also make a difference, I have not thought
about their relative magnitudes.

James

 

[Michael Macclan]

On Tue, 15 Jun 2004 14:38:03 +0100, Dr Curious wrote:

>
> I'm sorry but this all totally irrelevant.
>
> The radius of the wheel is irrelevant.
>
> When a tire is flat, all this means is that the wheel/tyre
> is no longer circular. It's now a part circle with a flat
> at the bottom. And so there's more surface area in contact
> with the ground. But the actual outside circumference
> remains exactly the same. It wouldn't even matter if you
> were riding on oval wheels. All you ever need measure are
> the number of revolutions multiplied by the circumference
> of the tyre.
>
>

I may be wrong but I think there's a fault in your
reasoning. Consider the following questions:

Do you accept that the circumference of the wheel rim
(without tyre) is less than the circumference of the wheel
(measured on the outside of the tyre) when the tyre is
fully inflated?

Do you accept that when you are riding a bike with a flat
tyre then the flat is always at the bottom of the wheel?

Do you accept that that the effect of having a (fully) flat
tyre is to reduce the effective circumference of the tyre to
that of the wheel rim?

If you accept each of the above (and I can't see how you
could do otherwise) then you must also accept that a bike
with a flat tyre travels a smaller distance with each
revolution of that wheel than it would if the tyre were
fully inflated.
--
Michael MacClancy Random putdown - "He had delusions of
adequacy." - Walter Kerr www.macclancy.demon.co.uk
www.macclancy.co.uk

 

[Doctor J. Frink]

On Tue, 15 Jun 2004 14:38:03 +0100, Dr Curious <[email protected]>
wrote:
>
>The radius of the wheel is irrelevant.
>
>When a tire is flat, all this means is that the wheel/tyre
>is no longer circular. It's now a part circle with a flat
>at the bottom. And so there's more surface area in contact
>with the ground. But the actual outside circumference
>remains exactly the same. It wouldn't even matter if you
>were riding on oval wheels. All you ever need measure are
>the number of revolutions multiplied by the circumference
>of the tyre.

Sorry, but that's wrong for these measurements. Try it!
Take a bike, do a rolling test (unloaded) a few times,
measure the distance covered. Deflate the tyre until it
deforms significantly. Do the rolling test again. You get
different results.

A severe test I just did (with a pumped and ~75% deflated
tyre so it's deformed but not squishing about) made a
difference of a few cm on just one revolution. It's likely
to be less in real world conditions but it's still going to
be a few mm out.

I'm probably backtracking, but it's not really the absolute
circumference we're interested in, it's how far the bike
goes for one revolution of the wheel the sensor is attached
to (ie the effective circumference) and that distance
definitely (and I've physically proved it to myself) depends
on tyre pressure/load. We're dealing with real things that
deform here, not pure mathematical models.

Frink

--
Doctor J. Frink : 'Rampant Ribald Ringtail' See his mind
here : http://www.cmp.liv.ac.uk/frink/ Annoy his mind here :
pjf at cmp dot liv dot ack dot ook "No sir, I didn't like
it!" - Mr Horse