rollagain

NOTE: Mods, feel free to move this if you think it belongs somewhere else.
--------------------------
To the best of my knowledge, all cyclometers measure bicycle speed by counting pulses from a magnet attached to a spoke. If someone does it differently, feel free to correct my ignorance.

I got to thinking about using a different method that would give more accurate readings, especially at low speeds, by using a light beam. Light source on one side of the wheel, photocell on the other side; the detector circuit would count the interruptions of the beam from each individual spoke. Compared to the magnetic-pulse method, this scheme would increase the precision of the reading by a factor equal to the number of spokes in the wheel.

Small details, of course; the light source would have to be modulated and/or color filtered to prevent interference and false signals from environmental sources. The assembly would need to read the spokes far enough inward from the rim to avoid being triggered by the valve stem. Probably other little problems; the pulse train would be a little uneven due to cross-pattern spoke spacing, but I think that one could be fixed with a corrective algorithm in the detector circuit.

BIG problems: mud, snow, ice, maybe even just rainwater or puddle-splashes. Not sure if a shield of some sort could solve those.

It got crazier from there--maybe. I was thinking of high-frequency-modulated light sources, coupled with little mirrors, and realized I've got one right in front of me--an optical computer mouse. What would happen if you pointed one of those at, say, the sidewall of your tire? I wouldn't be surprised if it would give a reading, but you'd have to have it awfully close or else rework the design of it. Same problems with fouling from mud, etc.

Thoughts?

fietsbob

Furlongs per Fortnight?

rollagain

Sure, or leagues per lunar cycle if you like. ;-)

prathmann

It could give slightly faster response when your speed is changing but wouldn't really be different in accuracy from the standard method. The old Avocet models used a Hall effect sensor together with a special circular band at the hub that produced multiple pulses per revolution (20?) for the same effect

europa

The simple magnet triggered cyclometers are cheap and reliable. For those wanting a bit more, GPS units are now solid and reliable, though not cheap. I'm not saying "don't do it", just that you may have trouble finding a niche for the new system.

On the other hand, if all you're after is a fascinating project, go for it.

I'm reminded of car and motorcycle ABS systems which, I believe, use an optical system to read if the disc the sensor reads (don't know its name) is still spinning.

HillRider

There is nothing so simple we can't complicate it.

TimothyH

Garmin wireless speed sensors using a magnetometer which references against the earth's magnetic field.

Others use accelerometers.


-Tim-

dieselrover

If precision is your goal, then todays microelectromechanical sensors that you find in the wheel-hub devices of Wahoo and Garmin et al. will typically measure on the order of 12 bits of resolution (~4000 values) with maybe a conservative 8 bits of usable data (~250 values). These are based on accelerometers, measuring the force of gravity as the wheel spins around and around or magnetometers as stated before.

You'd need a lot of spokes to reach that level of precision.

https://support.garmin.com/faqSearch...Jd3Tnyif9jRSy6


s

bike_galpal

Moose

Your method would still rely on calculating speed using a formula relating to the tire's roll out distance. The system would be just as compromised by the the variations in circumference of tires as the system you are trying to improve. Doubtful that any accuracy achieved by counting spokes rather than the magnet's complete revolution would overcome the margin of error in the tire's actual roll-out distance.

pvillemasher

Solution looking for a problem

rollagain

That's mostly what I expected. Fun to play with ideas, though, and thanks everyone for the contributions.

sweeks

I still have one of these! Simple by today's standards, but it worked well.
Steve

Dean51 

+1

Dean

HillRider

I had Avocet cyclometers starting in 1986. The early ones were very good, the later ones not so much.

sweeks

Mine was on my first "new" bike, a Motobecane Mirage I bought in 1975. After all this time, I'm restoring this bike because I like it. I could use the old Avocet computer... it would match the Avocet saddle.
Steve

gregf83 

Your solution would have higher resolution but be no more accurate than a system counting wheel revolutions.

The challenge in either system is to determine how far you travel per wheel revolution.

KD5NRH

Others do neither: https://www.walmart.com/ip/Bicycles-...eter/401844238

TimothyH

That thing doesn't make sense.

The inner scale goes from zero to 400+ and is either RPM or km/hr.

79pmooney

I never thought of this before but I wonder if the circuitry in ABS is similar to that of the SawStop, a system that detects a change in electrical resistance from flesh contacting a power saw blade developed for chop saws used by lumber mills. (I spent a little time working for an engineer designing chopsaws for one of the early outfits to buy a license for the stop. In those days you would draw blood from a small nick. Apparently they stop the blade even faster now.) One significant difference - the blade and stop get destroyed but we get to use out ABS repeatedly.

Ben

europa

You can't ride at 400 km/hr? All you need is a decent hill

KD5NRH

Wheel RPM makes sense. OTOH, that would mean the gauge can only read an accurate speed with one wheel size somewhere around 24".

alias5000

You can have that much easier with existing technology. Mount a magnet on each (/every other) spoke and divide the wheel circumference in your bike computer by the number of magnets installed.

Retro Grouch 

The way that you improve precision is by reducing measuring errors. Success can only be achieved by minimizing the biggest errors.

I'm thinking the greatest measuring errors in a wheel driven odometer come from inaccurate wheel circumference input measurements and riding an inconsistent path.

It doesn't matter how many wheel pulses you count per revolution if your circumference measurement is off.
It doesn't matter how accurately you count each revolution of the wheel if repeated rides don't exactly follow the same path.

wphamilton

The optical switch speedometer has been tried, usually with a spinning disk with holes mounted on the wheel, and has the great advantage of more frequent data, and more precise data. Which is a big "don't care" for the usual speedometer usage but can be useful otherwise. I think it may be possible but tricky to use the spokes, but if you're going to have a shield anyway for environmental factors why not just use the perforated disk in the first place?

Using an optical/laser mouse seems like a shortcut until you have to place it right up against a surface and try to make that work. If you're going to have to buy a few high quality lenses and experiment to make a mouse work it defeats the purpose of using the mouse. Expensive, extra trouble. Your first plan with a laser with detector is more practical.