Battery and data suggestions

[creid]

I love this project and am going to attempt something Similar but probably with an arduino or atmega. I have two suggestions regarding the problems ive seen a lot of you have.

1 Batteries - There is no reason your batteries have to be mounted to the board. Take your batteries off the board and mount the battery holders around the rim of the wheel. If you do that centrifical force becomes irrelevant and you can run as many colors and diodes as you want. You should also consider surplus cellphone batteries theyre all over ebay cheap.

2. Power overall - You already have a spinning wheel going through a magneticfield.... thats a generator. A wire loop buffered through capacitors could recharge your batteries as you ride perhaps? (It might be too spikey etc)

3. Data transfer/ storeage - if you interface a bluetooth (or ethernet chip) you make it wireless. Use volatile ram and you can load the image from a cell phone on the fly. Use bitplanes (as they used to in the first pc's at least) . I know nothing about interfacing dynamic ram but its certainly megafast. The only downside of this is you may have to upgrade your MCP or mount a second one just to handle memory etc.

Just ideas. Im hoping to design a car wheel pov using an arduino with bluetooth and dynamic ram (I hope!) to have a 256 or more color dynamic display pov. The work here is a huge inspiration.

[karlgg]

You'd need a lot more magnet (and coils to match) to generate the juice all those LEDs need. And might make your wheel more of a convoluted contraption than you'd think. But it should certainly be doable.

As for adding bluetooth (or any radio) to a spinning doodad surrounded (I assume) by a wire cage... I'm not sure what that would do to the signal. My radio knowledge is old and mostly forgotten, but seems to me Doppler issues and antenna directionality (even in "omindirectional" antennas) would make a horrible connection while moving in circles.

Anybody know more about that sort of thing?

[karlgg]

Also, dynamic memory is generally just teeny tiny capacitors. You charge them to write (if you want a "1" in there), and then read the voltage back. But they're small enough that they lose the charge pretty quick, and have to be constantly 'refreshed' by reading and re-writing.

It's not 'hard' per se to interface with them, just more of a pain than most people want to go through. (Especially since it only takes one late refresh to completely corrupt whatever you have in memory) There are chips especially designed to do nothing but babysit the DRAM so the processor doesn't have to... But I have no experience with using them directly. Maybe there are some suitable for the AVR.

[creid]

Thanks a lot kardgg. Im certain youre right on the charging system. Actually might be a good idea for an onboard power system ala a car regulator. I doubt everyone is going to keep their pov's on 24/7. Still as you said.. why not just plop batteries back inthe charger.

The only reason i was thinking dynamic memory is its is massively cheap, perfect (fast) for display systems etc. I doubt ill ever do anything so complex (or am capable) but the ultimate way to go with these systems is to use a 3 plane display ram block (the way pc systems at least used too). Then you dont use the onboard system to do anything but repaint the display over and over based on memory. You can then completely offboard everything else so you sit in the car/on the bike / whatever and transfer what you want displayed via bt/wifi/xbee/ whatever.

Ill wait til wireless tech prices drop before i try any of that. Im not putting an $80 bt module on each tire of my car (or two wheels of a bike) to be vibrated, pressured, soaked, sandblasted etcc etc heh.