> >
>

# Calibrating distance on computers - long measured courses - surveyors wheels?

Notices
Electronics, Lighting, & Gadgets HRM, GPS, MP3, HID. Whether it's got an acronym or not, here's where you'll find discussions on all sorts of tools, toys and gadgets.

# Calibrating distance on computers - long measured courses - surveyors wheels?

05-11-22, 08:56 PM
#1
KC8QVO
Senior Member

Join Date: Apr 2013
Posts: 1,081

Bikes: Surly Disk Trucker, 2014 w/Brooks Flyer Special saddle, Tubus racks - Duo front/Logo Evo rear, 2019 Dahon Mariner D8, Both bikes share Ortlieb Packer Plus series panniers, Garmin Edge 1000

Mentioned: 2 Post(s)
Quoted: 369 Post(s)
Liked 103 Times in 89 Posts
Calibrating distance on computers - long measured courses - surveyors wheels?

All,

I am curious how some of you are calibrating bike computers over longer measured course lengths - at least a mile if not 2-3+.

I have done the wheel count method over a measured distance. I want to say I used 200 feet, maybe 300 feet. On the 2 bikes I ride there is a slight variation in the distance over a routine ride I do. I find that a bit odd - if the wheel count method should be accurate then the numbers should line up closer than what they do.

That brings me to my question on calibrating over longer courses. Surveyors measuring wheels can keep track of distance while rolling, as opposed to stretching out a tape (even a long 100-300 foot reel tape) and resetting it. Plus - it would allow you to measure continuously however far you want to go. I think a lot of them max out at 9,999 feet. So you would be able to fit a mile in there at 5,280. Or, if you went further than 9,999 feet the counter would just start over. If you factor that in to the counter turning over - say the reading is 3,200. At 9,999 before that would be 13199 feet, or about 2.5 miles.

Just curious what others have done. My theory with the measured course is it would average out other inaccuracies that could be at play over a shorter distance - and that is what I believe the root of my mismatched distance readings is. They are close, but I notice they are different. I'd like to get them closer - and I don't know which one of the two is the most accurate.
05-12-22, 05:17 AM
#2
Tourist in MSN
Senior Member

Join Date: Aug 2010
Posts: 9,545

Bikes: 1961 Ideor, 1966 Perfekt 3 Speed AB Hub, 1994 Bridgestone MB-6, 2006 Airnimal Joey, 2009 Thorn Sherpa, 2013 Thorn Nomad MkII, 2015 VO Pass Hunter, 2017 Lynskey Backroad, 2017 Raleigh Gran Prix, 1980s Bianchi Mixte on a trainer. Others are now gone.

Mentioned: 41 Post(s)
Quoted: 2774 Post(s)
Liked 934 Times in 761 Posts
Why? If you have less than a half percent error in your computer, does it matter?

If you were setting up a track for competition, use a surveyor tape measure with all the bells and whistles (temperature compensation, correct tension on tape, etc.).

When I measure the distance of one wheel revolution with a tape measure and come up with something like 85.25 inches, if I was off by a quarter of an inch (it was really 85.00 inches), my error is a hair less than 0.3 percent. Get a 25 or 50 foot tape measure and do three or six revolutions on the wheel to increase precision.

I think you are getting precision and accuracy a bit mixed.
Likes For Tourist in MSN:
05-12-22, 05:21 AM
#3
Kai Winters
Senior Member

Join Date: Oct 2008
Location: Northern NY...Brownville
Posts: 2,066

Bikes: Specialized Aethos

Mentioned: 1 Post(s)
Quoted: 62 Post(s)
Liked 93 Times in 55 Posts
meh, don't care that much
05-12-22, 06:18 AM
#4
njkayaker
Senior Member

Join Date: Sep 2007
Posts: 13,338
Mentioned: 27 Post(s)
Quoted: 3538 Post(s)
Liked 941 Times in 656 Posts
Originally Posted by KC8QVO
All,

I am curious how some of you are calibrating bike computers over longer measured course lengths - at least a mile if not 2-3+.

I have done the wheel count method over a measured distance. I want to say I used 200 feet, maybe 300 feet. On the 2 bikes I ride there is a slight variation in the distance over a routine ride I do. I find that a bit odd - if the wheel count method should be accurate then the numbers should line up closer than what they do.

That brings me to my question on calibrating over longer courses. Surveyors measuring wheels can keep track of distance while rolling, as opposed to stretching out a tape (even a long 100-300 foot reel tape) and resetting it. Plus - it would allow you to measure continuously however far you want to go. I think a lot of them max out at 9,999 feet. So you would be able to fit a mile in there at 5,280. Or, if you went further than 9,999 feet the counter would just start over. If you factor that in to the counter turning over - say the reading is 3,200. At 9,999 before that would be 13199 feet, or about 2.5 miles.

Just curious what others have done. My theory with the measured course is it would average out other inaccuracies that could be at play over a shorter distance - and that is what I believe the root of my mismatched distance readings is. They are close, but I notice they are different. I'd like to get them closer - and I don't know which one of the two is the most accurate.
No one cares that much.

​​​​​​The very-involved approach might not be better than taking the average of multiple shorter runs.

Do you think you can ride a bike straight for 2.5 miles?

You see a "slight variation" but don't reveal the numbers. It sounds like one is always lower than the other. That would be a "difference", not "variation".

What are the two bikes? Is the difference always lower for one bike? If so, make you ride one bike straighter (faster)?

Note that this depends on load and tire pressure.

Use circumference numbers for each based on the average distances measured. That should get the numbers closer (if one is always less than the other).

Last edited by njkayaker; 05-12-22 at 06:30 AM.
05-12-22, 06:46 AM
#5
Chuck M
Butted Hi-Tensile

Join Date: Sep 2020
Location: Oklahoma
Posts: 1,127

Bikes: Hi-Ten bike boomers, a Trek Domane and some projects

Mentioned: 1 Post(s)
Quoted: 369 Post(s)
Liked 967 Times in 478 Posts
It seems people can get overly concerned with accuracy on these devices. Some rides I do with a gps watch and some rides with a computer with gps and wheel sensor. On the rides with the computer, I don't know how much of the calculated distance is done by gps data and how much by wheel revolutions. But I don't think I've noticed enough of a difference I would want to be worried about.

Off topic as far as wheel calibration but related to accuracy concerns is I also run with a group. For several seasons I was the one that went out on the path and marked our mileage and turn around points. Many would complain to me that the markers should have been this many or that many feet before or after where I marked them. Or they would tell me they only got 5.94 miles or went 6.03 for a six mile run according to their watch or phone app. Those without watches or smart phones just simply thanked me for my time I spent doing this for the group.
__________________
"It is the unknown around the corner that turns my wheels." -- Heinz Stücke

05-12-22, 07:46 AM
#6
VegasTriker
Senior Member

Join Date: Feb 2006
Posts: 2,670

Bikes: Catrike 700, Greenspeed GTO trike, , Linear LWB recumbent, Haluzak Horizon SWB recumbent, Balance 450 MTB, Cannondale SM800 Beast of the East

Mentioned: 18 Post(s)
Quoted: 474 Post(s)
Likes: 0
Liked 173 Times in 134 Posts
If you live in an area laid out in a grid as many places outside of the eastern US you can check the accuracy of your computer by simply riding. In my area most streets were built so that the main avenues are a mile apart with a secondary road at the half mile point. I can choose to read the computer at the beginning of one road and go several miles to the same point at a road a couple miles away and then again read the computer. I'm pretty amazed at how accurate an inexpensive computer is if you set the wheel diameter correctly. The instruction sheet that comes with the computer is an estimate. If you want accurate readings do a runout on the wheel. Tie a narrow piece of string around the wheel, douse it with a colored liquid like India ink or food coloring and ride the bike a few feet on cement pavement. The distance between two lines on the pavement will be an accurate value to plug in the computer.
05-12-22, 08:17 AM
#7
Iride01

Join Date: Jul 2017
Location: Mississippi
Posts: 11,038

Bikes: Tarmac Disc Comp Di2 - 2020

Mentioned: 42 Post(s)
Quoted: 4458 Post(s)
Liked 2,987 Times in 2,081 Posts
How many miles you rode really doesn't mean anything. The only thing that counts is who crosses the line first.
05-12-22, 08:24 AM
#8
prj71
Senior Member

Join Date: Jun 2015
Location: North Central Wisconsin
Posts: 3,769
Mentioned: 13 Post(s)
Quoted: 2257 Post(s)
Liked 734 Times in 490 Posts
Never. The mileage at the end of the ride is close enough to not care about that much accuracy.
05-12-22, 08:32 AM
#9
Barry2
LR÷P=HR

Join Date: Sep 2019
Location: SF Bay Area
Posts: 1,382

Bikes: Holdsworth 1979 Special, C-dale 1993 MT3000 Tandem & 1996 F700CAD3, Cervelo 2022 R5 & 2018 R3, JustGo Runt, Ridley Oval, Kickr Bike 8-)

Mentioned: 0 Post(s)
Quoted: 525 Post(s)
Liked 581 Times in 360 Posts
Most really modern bike computers use GPS tracking to calculate distance, there is no calibration.
If you add a wheel sensor, some computers use the GPS to automatically set the wheel circumference to match the GPS for distance.
Then in areas of low/no GPS signal, the computer uses the wheel sensor for speed and distance.
As a bonus, the wheel sensor can also allow the computer to register when you start and stop more quickly than solely using GPS data.

Hope that helps

Barry
Likes For Barry2:
05-12-22, 08:36 AM
#10
pdlamb
Senior Member

Join Date: Dec 2010
Location: northern Deep South
Posts: 7,627

Bikes: Fuji Touring, Novara Randonee

Mentioned: 32 Post(s)
Quoted: 1935 Post(s)
Liked 1,194 Times in 758 Posts
Depending on where you live, the people who build and maintain the roads may have already done the surveying for you and marked them with mile markers. I like to do a roll-down test: I check my odometer reading at the mile marker near the top of the ridge and roll down two miles. If correction is needed, I can get within 0.5% this way.

Note this depends on where you live and ride. In some parts of the midwest, they seem to put up mile markers like some people put up bouquets of balloons to celebrate a child's birthday. (Maybe that explains some of the numbering there, too!)
05-12-22, 01:53 PM
#11
Tourist in MSN
Senior Member

Join Date: Aug 2010
Posts: 9,545

Bikes: 1961 Ideor, 1966 Perfekt 3 Speed AB Hub, 1994 Bridgestone MB-6, 2006 Airnimal Joey, 2009 Thorn Sherpa, 2013 Thorn Nomad MkII, 2015 VO Pass Hunter, 2017 Lynskey Backroad, 2017 Raleigh Gran Prix, 1980s Bianchi Mixte on a trainer. Others are now gone.

Mentioned: 41 Post(s)
Quoted: 2774 Post(s)
Liked 934 Times in 761 Posts
Originally Posted by VegasTriker
... If you want accurate readings do a runout on the wheel. Tie a narrow piece of string around the wheel, douse it with a colored liquid like India ink or food coloring and ride the bike a few feet on cement pavement. The distance between two lines on the pavement will be an accurate value to plug in the computer.
That is a very good idea to use something to mark the pavement. Thanks for posting.
05-13-22, 04:16 AM
#12
KC8QVO
Senior Member

Join Date: Apr 2013
Posts: 1,081

Bikes: Surly Disk Trucker, 2014 w/Brooks Flyer Special saddle, Tubus racks - Duo front/Logo Evo rear, 2019 Dahon Mariner D8, Both bikes share Ortlieb Packer Plus series panniers, Garmin Edge 1000

Mentioned: 2 Post(s)
Quoted: 369 Post(s)
Liked 103 Times in 89 Posts
The 2 computers are off about 7% between the two. One is higher than the other. The 7% difference is calculated from an average of many of the same rides. Each bike's ride individually is very close and repetitive (if I take the upper and lower distances of each bike I get about a 2% difference ride to ride span).

What I don't like about the "measuring the circumference of the wheel" method is this does not take in to account the compression of the tire when under load.

That is why I am trying to get an accurate course distance measurement. Then I can tune the calibration factor on the computer per bike so that the reading is repetitive and correlates to the measured course.
05-13-22, 04:59 AM
#13
njkayaker
Senior Member

Join Date: Sep 2007
Posts: 13,338
Mentioned: 27 Post(s)
Quoted: 3538 Post(s)
Liked 941 Times in 656 Posts
Originally Posted by KC8QVO
The 2 computers are off about 7% between the two. One is higher than the other. The 7% difference is calculated from an average of many of the same rides. Each bike's ride individually is very close and repetitive (if I take the upper and lower distances of each bike I get about a 2% difference ride to ride span).
What are the two bikes? What size tires? How did you determine the effective circumstances? When did you last measure both? Did anything change since then?

I suspect a simple and proper rollout would be closer. Maybe, one if them is wasn't done well.

Originally Posted by KC8QVO
What I don't like about the "measuring the circumference of the wheel" method is this does not take in to account the compression of the tire when under load.
That approach is the wrong way to do it.

(But it might not matter too much.)

Originally Posted by KC8QVO
That is why I am trying to get an accurate course distance measurement. Then I can tune the calibration factor on the computer per bike so that the reading is repetitive and correlates to the measured course.
There's no reason this would be better than a properly done roll-out.

I suspect one of the bike's measurements isn't as good as it could be.

Last edited by njkayaker; 05-13-22 at 05:23 AM.
05-13-22, 05:00 AM
#14
Chuck M
Butted Hi-Tensile

Join Date: Sep 2020
Location: Oklahoma
Posts: 1,127

Bikes: Hi-Ten bike boomers, a Trek Domane and some projects

Mentioned: 1 Post(s)
Quoted: 369 Post(s)
Liked 967 Times in 478 Posts
Just curious... Are these two identical computers on two different bikes or are you moving the unit from bike to bike? And if they are two separate computers, are they both using the same satellite system, GPS, GLONASS or Galileo?
__________________
"It is the unknown around the corner that turns my wheels." -- Heinz Stücke

05-13-22, 05:24 AM
#15
njkayaker
Senior Member

Join Date: Sep 2007
Posts: 13,338
Mentioned: 27 Post(s)
Quoted: 3538 Post(s)
Liked 941 Times in 656 Posts
Originally Posted by Chuck M
Just curious... Are these two identical computers on two different bikes or are you moving the unit from bike to bike? And if they are two separate computers, are they both using the same satellite system, GPS, GLONASS or Galileo?
I doubt they have GPS. The OP is not very forthcoming with details.
05-13-22, 05:45 AM
#16
Tourist in MSN
Senior Member

Join Date: Aug 2010
Posts: 9,545

Bikes: 1961 Ideor, 1966 Perfekt 3 Speed AB Hub, 1994 Bridgestone MB-6, 2006 Airnimal Joey, 2009 Thorn Sherpa, 2013 Thorn Nomad MkII, 2015 VO Pass Hunter, 2017 Lynskey Backroad, 2017 Raleigh Gran Prix, 1980s Bianchi Mixte on a trainer. Others are now gone.

Mentioned: 41 Post(s)
Quoted: 2774 Post(s)
Liked 934 Times in 761 Posts
Some computers occasionally have a hiccup, I have a wireless VDO computer that at very slow speed, such as when I roll my bike outside my front door on my condo, I will get a measurable distance measured in several hundredths of a mile from moving my bike 20 feet. It has happened many times on the VDO wireless computer, but has never happened on my wired computer.

And the distance on my VDO computer is rarely repeatable within tenths of a percent where the wired computer is always spot on. The wired computer sensor sends a signal to the computer with every revolution, but the VDO wireless computer sensor sends a packet of information to the computer every several seconds, there can be several revolutions in that packet of information. And for some reason that I do not know, the precision on that wireless computer is not very good. I have had both computers for over a decade, the wired computer is consistently very precise, the wireless VDO is not. But I never care if my measured distance is off a few percent so I am not going to worry about this. If I cared that much about distance measurement, I have a separate GPS that I use sometimes, I don't call that a computer, I call it a GPS, it has no wheel sensor, I mostly use it for the map screen.

Originally Posted by KC8QVO
...
What I don't like about the "measuring the circumference of the wheel" method is this does not take in to account the compression of the tire when under load.
...
This might be a valid concern if you were riding a fat bike with only a few psi of pressure. Otherwise, this is not significant enough to matter, any error from that would be below 1 percent on most bike tires.

If you want to check this, ride a known repeatable distance, such as around a block with your tire at the max pressure on the tire label, then drop the pressure to the minimum pressure and ride the exact same route at the same speed to compare the difference. You will find minimal difference in the two distance measurements.

That said, I described above that my wireless VDO computer has poor precision, if I did that test at two different pressures, I would not trust my VDO computer to be precise enough.
05-13-22, 09:19 AM
#17
prj71
Senior Member

Join Date: Jun 2015
Location: North Central Wisconsin
Posts: 3,769
Mentioned: 13 Post(s)
Quoted: 2257 Post(s)
Liked 734 Times in 490 Posts
Originally Posted by KC8QVO
The 2 computers are off about 7% between the two. One is higher than the other. The 7% difference is calculated from an average of many of the same rides. Each bike's ride individually is very close and repetitive (if I take the upper and lower distances of each bike I get about a 2% difference ride to ride span).

What I don't like about the "measuring the circumference of the wheel" method is this does not take in to account the compression of the tire when under load.

That is why I am trying to get an accurate course distance measurement. Then I can tune the calibration factor on the computer per bike so that the reading is repetitive and correlates to the measured course.
Is this OCD? And why the need for such accuracy?
05-13-22, 09:20 AM
#18
Iride01

Join Date: Jul 2017
Location: Mississippi
Posts: 11,038

Bikes: Tarmac Disc Comp Di2 - 2020

Mentioned: 42 Post(s)
Quoted: 4458 Post(s)
Liked 2,987 Times in 2,081 Posts
7% error is still meaningless. Did you know that that when your auto manufacturer puts a odometer/speedometer in the cars they sell that that they are only required to have that much accuracy.

What will being more accurate do for you?
05-13-22, 10:00 AM
#19
pdlamb
Senior Member

Join Date: Dec 2010
Location: northern Deep South
Posts: 7,627

Bikes: Fuji Touring, Novara Randonee

Mentioned: 32 Post(s)
Quoted: 1935 Post(s)
Liked 1,194 Times in 758 Posts
Originally Posted by KC8QVO
What I don't like about the "measuring the circumference of the wheel" method is this does not take in to account the compression of the tire when under load.

That is why I am trying to get an accurate course distance measurement. Then I can tune the calibration factor on the computer per bike so that the reading is repetitive and correlates to the measured course.
You may be a bit OCD, but not unreasonably so. My rollout tests correlate pretty well with the computers' manufacturing instruction sheets. When I get on the bike and go for a ride, though, I find that going for the next smaller size tire on the sheet gives me a better match to the ride distance. E.g., a 700Cx32 tire matches mile marker and GPS ride distance better with a x28 calibration factor, or use the x25 calibration for a x28 tire.

Of course that's going to vary with rider weight and tire pressure. I suspect it will take some trial and error to find what works best for you.
05-13-22, 10:12 AM
#20
billridesbikes
Senior Member

Join Date: Nov 2019
Posts: 701
Mentioned: 1 Post(s)
Quoted: 347 Post(s)
Liked 413 Times in 249 Posts
Subtract 3% from the circumference of the computer on the bike with the longer distance, and add 4% to the circumference on the computer with the bike of the shorter distance. Now your devices are matched! Go live your live.

The distance you measure on a map isn’t as accurate as you think it is anyway, especially for longer distances, as it depends on the method chosen by the cartographer, slight elevation changes, and curvature of the Earth at your location (the Earth is not a perfect sphere, although often modelled as one for maps.) GPS is probably the most accurate you can do and I’ve seen some articles showing only small differences between phone apps, watches, bike computers, and handhelds.
05-13-22, 10:18 AM
#21
Seattle Forrest
Senior Member

Join Date: Mar 2010
Location: Seattle, WA
Posts: 22,928
Mentioned: 80 Post(s)
Quoted: 17590 Post(s)
Liked 9,509 Times in 5,362 Posts
Originally Posted by Kai Winters
meh, don't care that much
05-13-22, 12:18 PM
#22
Tourist in MSN
Senior Member

Join Date: Aug 2010
Posts: 9,545

Bikes: 1961 Ideor, 1966 Perfekt 3 Speed AB Hub, 1994 Bridgestone MB-6, 2006 Airnimal Joey, 2009 Thorn Sherpa, 2013 Thorn Nomad MkII, 2015 VO Pass Hunter, 2017 Lynskey Backroad, 2017 Raleigh Gran Prix, 1980s Bianchi Mixte on a trainer. Others are now gone.

Mentioned: 41 Post(s)
Quoted: 2774 Post(s)
Liked 934 Times in 761 Posts
Originally Posted by Iride01
7% error is still meaningless. Did you know that that when your auto manufacturer puts a odometer/speedometer in the cars they sell that that they are only required to have that much accuracy.

What will being more accurate do for you?
That would surprise me, as the odometer is used to determine if your warranty has expired or not, so there is a reason for regulation to protect the consumer from manufacturer having an odometer read with that much inaccuracy. When I got an new engine block from Jeep under warranty, yeah, that was not cheap and it was near the expiration of their warranty. My odometer was very close to matching my GPS for distance traveled on everything I have owned since I bought my first GPS in 2001.

Speedometer, it is common knowledge that manufacturers usually show a faster speed than actual, I am not questioning that instrument.
05-13-22, 12:37 PM
#23
John Valuk
Full Member

Join Date: Oct 2020
Posts: 200
Mentioned: 0 Post(s)
Quoted: 53 Post(s)
Liked 119 Times in 77 Posts
Originally Posted by KC8QVO
What I don't like about the "measuring the circumference of the wheel" method is this does not take in to account the compression of the tire when under load.
I put a piece of masking tape on the tire, and another piece on the floor of my garage. Line up the marks, get on the bike, and walk it forward - with most of my weight on the saddle - to get two complete revolutions of the wheel. Do this with a tape measure laid out on the floor as a reference for moving pretty much in a straight line.

Mark the position on the floor where the tape on the tire ends up after the two full revolutions. Measure the distance between the two marks on the floor, and get the effective tire circumference from that.

Repeat this a few times to see how much scatter there is in the measurement (in my experience, not much).

This is just how I do it. I have not compared weighted vs. unweighted to see if that would make a measurable difference.

Last edited by John Valuk; 05-13-22 at 12:49 PM.
05-13-22, 03:25 PM
#24
KC8QVO
Senior Member

Join Date: Apr 2013
Posts: 1,081

Bikes: Surly Disk Trucker, 2014 w/Brooks Flyer Special saddle, Tubus racks - Duo front/Logo Evo rear, 2019 Dahon Mariner D8, Both bikes share Ortlieb Packer Plus series panniers, Garmin Edge 1000

Mentioned: 2 Post(s)
Quoted: 369 Post(s)
Liked 103 Times in 89 Posts
Lets cut through all the opinions and the "this way" and "that way" and simplify my question:

What is an accurate method to measure 1-3 miles for a course that I can repeat then tweak the computers reading?
05-13-22, 04:14 PM
#25
Iride01

Join Date: Jul 2017
Location: Mississippi
Posts: 11,038

Bikes: Tarmac Disc Comp Di2 - 2020

Mentioned: 42 Post(s)