Bike Forums

Bike Forums (http://www.bikeforums.net/forum.php)
-   Electronics, Lighting, & Gadgets (http://www.bikeforums.net/electronics-lighting-gadgets/)
-   -   Bike computer accuracy - magnet placement on spoke (http://www.bikeforums.net/electronics-lighting-gadgets/736151-bike-computer-accuracy-magnet-placement-spoke.html)

dave_mac 05-17-11 04:31 PM

Bike computer accuracy - magnet placement on spoke
 
Hi everyone, I just bought a new bike computer and something that has really been bothering me is that I can't understand how it can be accurate when it doesn't really tell you where to place the magnet on the spokes.

Tell me if I'm wrong here but if the magnet is closer to the hub the magnet will make more RPM's vs if it's closer to the outer edge so how is it possible for the computer to accurately track distance and speed? Or is the inaccuracy negligible?

Allen 05-17-11 05:02 PM

Negative.
It does not matter how far out from the hub the magnet is placed because the magnet will only pass by the sensor once a rotation.

S=D/T
You program the computer with the circumference of the wheel and the sensor counts each time the magnet passes it by.
The inter part of the wheel and the outer part take the same amount of time to make one rotation. The edge of a wheel moves at a faster rate than the inner part because it is moving a greater distance but they take the same amount of time to complete one rotation.
The bike moves forward the same distance as the circumference as the wheel each rotation.
So what your computer is calculating is: speed equals the circumference of the wheel (the distance the bike moves each rotation of the wheel) divided by the time it takes the wheel to complete one rotation.

colleen c 05-17-11 05:06 PM

RPM of the rim remain the same regardless of where you place the magnet on the spoke. One turn of the wheel at the tire is equal to one turn at the spindle.

Magnet sensor read each individual pulse when the magnet passes the sensor. Since your bike will move X amount of feet with one revolution of the wheel, the sensor will see only one pulse for that one revolution regardless of the magnet is closer to the outside or inside of the wheel.

Edit: forgot to include a formula:

(rev/min) * (ft/rev) = ft/min

then:
(ft/min) * (mile/5280ft) * (60min/hour) = miles/hour

colleen c 05-17-11 05:46 PM

Quote:

Originally Posted by colleen c (Post 12656438)
RPM of the rim remain the same regardless of where you place the magnet on the spoke. One turn of the wheel at the tire is equal to one turn at the spindle.

Magnet sensor read each individual pulse when the magnet passes the sensor. Since your bike will move X amount of feet with one revolution of the wheel, the sensor will see only one pulse for that one revolution regardless of the magnet is closer to the outside or inside of the wheel.

Edit: forgot to include a formula:

(rev/min) * (ft/rev) = ft/min

then:
(ft/min) * (mile/5280ft) * (60min/hour) = miles/hour

Just to clarify, in the real world, computer use either cm/rev or mm/rev and more like rev/milisec or sec. It still the same principle but just a matter of more conversion. You get the idea.

ItsJustMe 05-17-11 05:51 PM

Um, no, the inside of the wheel turns the same number of times as the outside. How could it possibly turn more or less times? Swing your arm around three times. Did your bicep swing around more times than your hand?

For maximum accuracy, you should put the magnet and sensor close to the hub of the wheel. Theoretically it doesn't matter, but the problem is that near the rim, the magnet will be moving past the sensor much faster (but the same number of times per minute) and at higher speeds, the sensor may not pick up every time the magnet goes past it. It will start to "skip" and you will suddenly start getting drastically lower speed readings. This may or may not happen with any particular computer, but it doesn't hurt to move the sensor towards the center. As a side benefit, with the magnet nearer the hub, it will be less likely to unbalance the wheel.

no1mad 05-17-11 06:46 PM

The installation instructions weren't all that clear about whether to mount closer to the hub or rim. What it did make clear was that the spoke mounted magnet had to pass within like 2 mm of the sensor- up near the rim is too far away for them to "see" each other.

Another benefit about mounting closer to the hub (asides from the wheel balancing which i hadn't thought of) is that it's a bit more protected than if it was mounted near the rim.

dave_mac 05-17-11 07:11 PM

Makes sense. The logic I was using was if you're going around a track (or your hub in this instance) the inside lane is always fastest but the S=D/T formula totally makes sense. Thanks.

rm -rf 05-17-11 08:59 PM

Quote:

Originally Posted by ItsJustMe (Post 12656636)
Um, no, the inside of the wheel turns the same number of times as the outside. How could it possibly turn more or less times?
..snip..

Ask this guy. :innocent:

http://a6.sphotos.ak.fbcdn.net/hphot..._1637284_n.jpg

formicaman 05-18-11 12:00 AM

Glad I am not the only idiot out there whose brain can't seem to wrap itself around the basic fact that it doesn't matter where the damn magnet is. Did you major in communications too?

Richard Cranium 05-20-11 06:37 AM

Quote:

Tell me if I'm wrong here but if the magnet is closer to the hub the magnet will make more RPM's vs if it's closer to the outer edge so how is it possible for the computer to accurately track distance and speed? Or is the inaccuracy negligible?
Yes, of course RPM is constant. But like some one mentioned - it is important that you give the magnet enough time to "pull" the sensing switch in the pickup unit.

Believe it or not, when i used to get fierce while riding rollers, i could get pickup sending units to "fail" or skip when riding at speeds near 50 mph. This was mostly because the unit was simply nearing being worn out. But it is proof that there was actual "mechanical" activity in older cat-eye and avocet units.

I assume they might still be using "reed switches." I don't know for sure. It seems I do still hear 'ticking."

By the way the "numbers" in the setup routine represent centimeters - even if you put in the wrong number - its hardly 50 feet per mile, in other words your only going to be off about 1 mile per hundred.

But I digress.

ItsJustMe 05-20-11 09:02 AM

Quote:

Originally Posted by Richard Cranium (Post 12668477)
By the way the "numbers" in the setup routine represent centimeters

Millimeters, surely. My comp is set to something like 2055. If that was centimeters my wheel would be 20 meters in circumference, which works out to 21 feet in diameter.

Iowegian 05-20-11 10:18 AM

Whoops. Never mind.

dougmc 05-20-11 11:33 AM

Quote:

Originally Posted by rm -rf (Post 12657435)

Clever answer!

In case somebody doesn't know what's going on here ... http://en.wikipedia.org/wiki/Rolling_shutter

ItsJustMe 05-20-11 11:34 AM

Actually it is a reed switch. You can put a volt meter across it and see the resistance go from infinite to zero when a magnet is nearby.

High end ones might use a hall effect transistor, that would be the smart way to do it. Much more reliable than an inductive loop since a hall effect switch is extremely fast and will give you essentially an on/off switch no matter how slow the magnet is moving (or even if it's stopped altogether). Hall effect transistors are pretty cheap.

It's possible that newer comps have already switched to hall effects, I wouldn't know, my computer is a number of years old at this point.

Iowegian 05-20-11 03:40 PM

Egads! you are correct. I guess that eliminates the whole signal conditioning problem although I wonder what would happen if you park your bike with the magnet next to the sensor. Might have to try that out tonight.

Well, I edited my mis-post and I've been thinking it's time for a new avartar...

rscamp 05-20-11 04:03 PM

Yes, FWIW the speedo sensors I have are reed switches. I haven't had them skip but it is certainly possible for them to miss a very brief passing of the magnet even though the moving part has very little mass.

If the magnet is stopped at the switch it will just hold it closed (they are NO (normally open) if memory serves).

These switches are also used for other applications such as defeating throttle with brake on ebikes...

ItsJustMe 05-20-11 05:41 PM

I think I know another reason they use reed switches. that makes it trivial to build computers that are auto-on without drawing much current. With a clean reed switch, the switch can just be hooked to an interrupt line on the microprocessor, and it can go into deep sleep mode and wake when the switch state changes. If it were a hall effect switch it would have to be powered constantly to know when the magnet moved so it could wake up, and the battery would only last a few weeks at best. It might be a little better with a coil but still, it would need more sensitive circuitry that would have to be powered.

Iowegian 05-23-11 04:33 PM

I am (obviously) not an analog guy so bear with me on this....It seems to me that whether the switch is NO or NC it has to be connected to some source of power to generate a transition to whatever is counting the rev's. I'd imagine it's a normally open switch that momentarily closes to complete some circuit and this generates a clock pulse to a counter or something along those lines. But wouldn't that quickly drain the battery if the switch was left in a closed state?

ItsJustMe 05-23-11 07:52 PM

Quote:

Originally Posted by Iowegian (Post 12683627)
I am (obviously) not an analog guy so bear with me on this....It seems to me that whether the switch is NO or NC it has to be connected to some source of power to generate a transition to whatever is counting the rev's. I'd imagine it's a normally open switch that momentarily closes to complete some circuit and this generates a clock pulse to a counter or something along those lines. But wouldn't that quickly drain the battery if the switch was left in a closed state?

That's all true, but a microcontroller awaiting a state change in deep sleep mode can pull as little as a few nanoamps. I've done a little work with AVR controllers, and in deep sleep mode waiting for a state change, battery drain is such that a standard CR2032 button cell will last on the order of 3 or 4 years. It's true that if the switch happens to be closed, the current draw will go up a bit, but the pullup resistors can be quite high resistance and the drain would probably only kill the battery in a year or so. Since the odds of the wheel sitting where the switch is closed is probably about 1 in 100, in general you're still looking at multiple years, unless you happen to park it with the switched closed and leave it alone for 6 months.

Juha 05-24-11 01:05 AM

Quote:

Originally Posted by ItsJustMe (Post 12684430)
Since the odds of the wheel sitting where the switch is closed is probably about 1 in 100, in general you're still looking at multiple years, unless you happen to park it with the switched closed and leave it alone for 6 months.

And this would only be a problem for wireless computers, or if you used a wired computer and left it on the bike.

ItsJustMe 05-24-11 05:31 AM

Quote:

Originally Posted by Juha (Post 12685332)
And this would only be a problem for wireless computers, or if you used a wired computer and left it on the bike.

I guess I've never considered that. I've never taken a computer off the bike except once a year or so to change the battery, or twice a year when I need to get to the "set" button to change the time.


All times are GMT -6. The time now is 06:46 PM.