Tweet
I guess I was thinking about an analog signal from a magnetic sensor versus a pulsed output from an optical encoder. If I think about it, an analog signal is going to be quantized into the digital realm, anyway. If you want a better level of precision, use better A/D conversion.
Absolute position can be done by analog methods like a rotary position sensor. Inside, it has three coils of wire; a drive coil, a sense coil and a position coil. The position coil is connected directly to the rotating object. As it turns, the coil rotates, relative to the drive coil, and changes the field that is picked up by the sense coil. The resulting voltage is used to determine the position of the two coils, relative to each other.
But, why bother? We don't really need absolute position detection. Even if we wanted it, we don't need it down to the millionth of a degree. We only need to be accurate to within the distance between sprocket holes... 1/8-inch? That's easily obtainable. Right?
I guess the two methods are about a horse apiece if you don't count preamp, filtering circuits and A/D. On those grounds, I'd probably choose a rotary, pulse encoder.
I thought, maybe, a reluctor wheel could be slipped in, sandwiched under the knob or something. Why bother? It's just as easy to super glue a metal disk to the knob.
The other thing, I was just supposing. If a splice ran through the sprocket, there might be a momentary change in rotation speed which might be detectable. I don't know. Maybe it's too much of a long shot?
(i.e. Yes, it might be detectable but, to what lengths are you willing to go for what kind of payoff and how small? In theory, possible. In practical terms, too much work for to little payoff, even if it does work.)
Absolute position can be done by analog methods like a rotary position sensor. Inside, it has three coils of wire; a drive coil, a sense coil and a position coil. The position coil is connected directly to the rotating object. As it turns, the coil rotates, relative to the drive coil, and changes the field that is picked up by the sense coil. The resulting voltage is used to determine the position of the two coils, relative to each other.
But, why bother? We don't really need absolute position detection. Even if we wanted it, we don't need it down to the millionth of a degree. We only need to be accurate to within the distance between sprocket holes... 1/8-inch? That's easily obtainable. Right?
I guess the two methods are about a horse apiece if you don't count preamp, filtering circuits and A/D. On those grounds, I'd probably choose a rotary, pulse encoder.
I thought, maybe, a reluctor wheel could be slipped in, sandwiched under the knob or something. Why bother? It's just as easy to super glue a metal disk to the knob.
The other thing, I was just supposing. If a splice ran through the sprocket, there might be a momentary change in rotation speed which might be detectable. I don't know. Maybe it's too much of a long shot?
(i.e. Yes, it might be detectable but, to what lengths are you willing to go for what kind of payoff and how small? In theory, possible. In practical terms, too much work for to little payoff, even if it does work.)
Comment