So far, I’ve been quite happy with my SpokePOVs. They’ve worked fairly well together without any real signaling or coordination between them. But I have had a few problems, and I thought I’d write them up in case someone has similar problems. I’ve been having a lot of random resets, sync issues, and … uh … IC socket stability. Yeah, that’s what I’ll call it.
I’m suspecting a couple of reasons for the random resets. Since I like to really rip along on my bike about as fast as I can manage, centripetal force begins to cause some problems. In particular, I think that it’s causing the AA batteries to pull hard enough out that one or more comes out of contact with their battery clips and causes the microcontroller to do a hard reset. I’ve made a temporary fix by bending the clips in a little, but it’s still not perfect as they still band back after a day or two. I’ll probably either drill some holes at the tops of the tabs and tie the two tabs together, or place all of the batteries towards the hub of the tire and just use the external power connector on the boards. The other reason they might be resetting is the trace that gets cut on the board for three batteries (I’m using the higher voltage blue LEDs). I’ve noticed that flexing the boards can make two of them reset, and I suspect that I didn’t cut deep enough, and the foil of the trace just touches and comes apart from itself. But the more obvious indicator of this condition is that if I flex the board just right, the LEDs come on really dim, and my multimeter reads 3v on the power input.
Sometimes a random reset will cause my pattern of images to get confused from the others and causes a ghost of one image to come over the one displayed by the other two. I’ll fix this eventually by tying together one hall sensor to all three SpokePOVs and fiddling with the rotation offset in the firmware. Even if I do this, one could be out of sync from a reset (which I plan to fix anyway). To combat this, I’m actually thinking of tying a line to either another microcontroller, or between an adjacent POV that would have an interrupt that sends what rotation and frame it’s on so that other POVs could sync up to it. I haven’t bothered to write software or wire anything up yet though, it’s just an idea.
The last problem that I have is having ICs come out of their sockets, particularly my SPI EEPROM. If I get going fast enough, it actually begins to rock out of its socket a little. Luckily it hasn’t come all the way out yet. I’ll probably just epoxy it in one of these days.