Anyone good at Maths. Cateye cycle computer calibration.

Kona-Ian

Senior Retro Guru
Just testing new cycle computer again today.
Circumference measured 2080mm.
Looking at the chart that's for > 700x19C.
Popped out for quick test and computer says 4.3miles and GPS said 4.4miles.
On chart doesn't have 26x1.9 closest was 26x1.95 and that was 2050mm, when tested that it was 0.3miles out.
That was yesterday but went further, so more miles you do the more it goes out by, obviously.
So by using circumference measurement it's 0.1 miles out....wondering how much to change the 2080mm by.

Sorry if any of that is totally confusing :)
 

CassidyAce

Senior Retro Guru
Re:

Yep, that's confusing. Some of the discrepancy could be due to rounding, though. For example, if the computer registers 4.34 miles and the GPS registers 4.35 miles and both round to one decimal place, the computer will say 4.3 miles and the GPS will say 4.4 miles, giving the impression of a 0.1 mile discrepancy when, actually, it's one tenth of that, 0.01 miles. More than that, GPS will not be completely accurate and tyres sizes will not be exact either (not to mention deformation). However, 26x2.30 might give you a more accurate measurement given that 26x2.35 comes in 2083mm.
 

FluffyChicken

Archivist
Retrobike Rider
Feedback
View
Re:

circumference, measured on the bike by rolling or off the bike just around the tyre?

anyway.
you need the circumference to increase.. so
(i.e
for each revolution it needs to say you have traveled a bit further)

4.4/4.3 * 2080

of course tweak as you get more measurements over longer distances.
 

Kona-Ian

Senior Retro Guru
Yes chicken, measured by marking ground and doing a complete turn of the wheel.
Have adjusted it to 2090 but not been out yet.
 

FluffyChicken

Archivist
Retrobike Rider
Feedback
View
Re:

works out as 2128 from my maths, by that assumes 4.3 and 4.4 are perfectly correct.

I'd go half and half and try 2100 as a start and work from there.

add a bit if you are short, take some off if long.
GPS could of course be wrong.
 

Kona-Ian

Senior Retro Guru
Thanks for the help Chicken and Cassidy, I will try that tomorrow, couldn't go out today ended up doing other stuff that needed doing.

The GPS was just a phone app. I prefer to use a cycle computer but was using it to compare accuracy.
 

mattr

Old School Grand Master
Kona-Ian":3lrjlxa5 said:
The GPS was just a phone app. I prefer to use a cycle computer but was using it to compare accuracy.
You'll not get any accuracy from the GPS on a phone (or GPS generally!) As soon as you have any corners or gradients, or tree cover or, basically anything other than long, flat, straight road. You'll be adding in errors.

Not to mention sampling and positioning errors!
(you even get errors between two handlebar mounted GPS, doing the same route, at the same time)
 

hamster

Retro Wizard
Feedback
View
The GPS will generally measure straight line distances between each of the points it logs. So you get a 'threepenny bit' effect round corners.
Selecting the low power mode on phone GPS will reduce the number of samples and on some turns off some positioning channels, reducing accuracy.
Furthermore, if the satellites are in poor alignment as they fly overhead, you get degradation of accuracy. The GPS on my boat warns this, phones generally don't.

Some GPS apps will plot onto mapping data and then recompute the distance.
 

FluffyChicken

Archivist
Retrobike Rider
Feedback
View
Newer GNSS style ones are better, the don't just rely on the GPS but also the other sat's we can use up there.
They can sample faster than 1Hz but many don't bother or allow the settings.
There is another new generation with increased accuracy, started around late last year when the regulations changed, but no idea who uses the newer chipsets.

Phones etc will always be worse of as they only have small antennas for it.

It's all down to the app once the GNSS setup has been chosen.

Newer and more money does tend to mean better.

Think though at 20mph you're going ~9 meters every second (1Hz). It's good enough and I'd trust the wheel if setup correctly much better, especially if you've measured how far it travels fully kitted out over 10 or 20 meters and then worked out the circumference for one rotation. It's hard for that calculation to got wrong.
GNSS/GPS is just convenient.
 
Top