4 ms·
What about an array of photodiodes like this: o o o o X o o o o This is in an enclosed box with a hole over the center. Here, "o" represents a photodiod
by biot 11y ago
What about an array of photodiodes like this:
o o o
o X o
o o o
This is in an enclosed box with a hole over the center. Here, "o" represents a photodiode and "X" represents the center of the hole. When the solar panel is pointed directly at the sun, the sun is at position X and no photodiode is lit up. As the sun changes position, eventually it will light up one of the photodiodes. That completes a circuit which uses power from the panel to drive a motor, which is attached to a worm gear and rotates the panel towards the sun. Eventually the sun will be pointed at the X position again which breaks the circuit and stops the rotation. Each diode completes a circuit, driving one or both motors (in the case of the corner diodes) rotating them appropriately.
It'd be interesting as well to have some sort of mechanism which, when the panel receives power, keeps it locked in place. When the panel has no power for a period of time (ie: after sunset) the mechanism disengages and the panel falls back to its default position pointing towards sunrise.
- gkop 11y agoGood ideas! I had in mind a purely mechanical mechanism that takes the linear motion of the hydraulic piston or the water clock and turns it into a parabolic movement. I'm a software guy and could not find such a mechanism easily on Google, though.
- mapt 11y agoFinding vertical & finding north is not a problem, you can get that trivially in a tiny MEMS IMU if you don't want to do it manually, and from there determine the optimal path for solar tracking without added variables, supposing you have the timezone set by the distributor (or use a GPS for that at a few dollars a chip). The cost is in the rotating frame and the motors.