On Thursday, March 14, 2019 at 9:01:22 PM UTC-7, James wrote:
On 15/3/19 1:47 pm, Mike A Schwab wrote:


Well, actually the reed switch is effectively a relay. Takes 1/100 of
a second to close in the gap between the sides of the magnet.
Computer could be sampling much more often. And a faster speed
doesn't give the reed switch enough time to close.


Why does it need to be "sampling"?

The switch may be connected to a digital input that generates an
interrupt upon a rising (or falling) edge, for example. The ISR could
then read and reset a timer/counter since the last rising edge was
detected. A counter can keep track of distance travelled.


At 60 m/s (in excess of 200km/h), with a nominal 2m circumference, the
wheel spins 30 revs per second. A 100 uS timer period would capture the
speed of 1 rev with better than 1% accuracy at this speed.

So long as it could count to 100,000 (not impossible for a 4 bit
computer to handle roll over), it could also measure speed at 1km/h.

The screen could be updated with the latest distance or filtered speed
every second or when it has changed.

--
JS

James - If, as Frank says, they are using a 4 bit uP, that counter is a whole lot easier said than done. As I pointed out the counter has to be at least 20 bits and more likely 24. Small, cheap uP are that way for a reason - they have very little support circuitry. Dealing with large numbers is not their strong point.

Also Zen's idea that you can make enough of ANY bicycle component to qualify it for custom fabrication is WAY over the top. I wonder if he knows what it cost to manufacture custom chips that will be outdated next year because someone gives their's another feature. Speed averaging for display isn't the world's easiest problem and they are doing that in a continuous update.

Since the altitude problem requires two inputs minimum to an ADC there is no possibility of making a custom chip. And he doesn't seem to know what a gate array is or what it does.