Hi,

This is just in the idea fase, but at some point I want to hack my hometrainer. It has a little computer that shows you the speed you cycle, the time, and other "yadayada" non-interesting information. Anyway...

So, what I like to do, is listen to the board-computer's incoming sensor signal, with a Atmega8 chip, and connect make the signal digital, so I can send it over a UART-to-USB cable to my computer, and do more interesting things with it... Like a little 3d program that shows you a 3d world where you cycle, and the faster you cycle in real live, the faster you go in the world.... Or something simpler to start with.

I am all-ready-to-go on the software side. I probably hook it up first to a game I allready wrote a long time ago. On the AVR side, I am also good to go.... I just need to get the signal somehow, in whatever format it is, into the AVR.

So, for the hardware side of things. On this hometrainer, type "Nordic"... There is no external wheel exposed, so I can't mount any simple sensors to detect speed myself. (Just anticipating a question)
What does exist is two wires that come from "some sensor", and go to this little on board computer, as I described in the beginning.

So, what could those two wires be leading to, is my question to the forum. I am not really in the mood to take the whole bike apart just now, but maybe I do that, if I can't get the info otherwise.

I was mostly expecting 3 wires. Voltage, Ground and signal back.... So what can two wires provide me? One seems to go to the CPU directly. Not alot of analogues stuff is on the circuit, so I expect it could be digital allready. The second seem to be connected directly to voltage. Maybe there is some on/off circuitry in the sensor. Or maybe the resistance changes? (but then I'd expect 3 wires again.)

Does any one has any good ideas?

I post some pics later (Need to take them first)...

Thanks for any comments
DaC

 

Pics 1/2

 

Pics 2/2

 

Ok, it was easier then I thought. The only two wires to the sensor instead of 3 should have been a giveaway, that it indeed is a simple circuit breaker / maker "sensor".

Just for anyone interested:

How I found out is, like this: I linked up a 3k resistor, with a led, and coupled it to the contacts of the sensor that goes into the computer. Then I cycled and watched the behaviour of the led. On, Off, or Dim? Result: The led is normally in the state off. However when I cycle, it blinks once for each rotation of my pedals.

So this means that I can directly connect one side of the sensor output to + voltage (not more then 5 volt), and the other side directly to my AtMega8, any digital input will do.

 

Gee, that was a lot of reading to arrive at, "I don't need anything".

 

Sorry, but sometimes I need to talk to someone to get my brain-cells all lined up... This was perhaps one of those occasions, maybe someone has use of this information, if not, feel free to delete the thread ;o)

 

Not within my power

Actually, any "pro" would expect, just from seeing 2 wires connected to a brain chip. that the result is going to be pulses. "One per pedal revolution" is important information.
Bring that with you when you come back for design help.

 

I am still a bit confused how the "switch" part in the sensor works.. I can imagine a mechanical switch would be unreliable after a while and lots of stutter in the signal... Then again, how else could it have been implemented, with only a +V going to it, and a Signal coming from it. Well, maybe I still will take it apart some day to find out how this works in detail.

 

Magnets...or light pulses can do this.

 

Sure, but my question is more. If there is only a + voltage, and no ground going to the sensor. How would the light be activated. 3 wires would do the trick. + and - to the light.
+ to a photodiode in series with a resistor which connect to -
signal back from the midpoint between photodiode and resistor... something like that. I am just not sure how to do that with only two wires.

 

Then I guess it's a magnet and a Hall sensor.

 

Magnet and a reed switch would do it

 

*After reading up* A reed switch would definitly work. A Hall sensor seems to need 3 contacts once more according to Wiki.

A reed switch is a mechanical switch though... So I guess it would wear out after a while, and my software (or hardware) needs to account for stutter.

 

http://www.hamlin.com/technical-detail-hall-sensor.cfm

I'm thinking about changing my signature line to:"You can always find incomplete or frankly wrong information on the internet".

 

The little wheel under the ski of my old NordicTrack had a magnet and a reed switch plugged into the "computer". When the computer died, I installed a cheap Schwinn bicycle computer and set the distance per revolution all the way down to 188mm to match the little wheel. Surprisingly, the computer had no lower range limit; good to know if you want to make a measuring wheel or similar hack.

 

http://www.hamlin.com/technical-detail-hall-sensor.cfm

I'm thinking about changing my signature line to:"You can always find incomplete or frankly wrong information on the internet".

But they cannot put anything on the internet that is wrong. I read that on the internet somewhere...