Noot / Erwt Notes

Note #7 Sunday 25 September 2011

My very shaky knowledge of analog electronics means that transistors and I do not get along well. It was most concerning when my H-bridge didn’t actually work as the sim had predicted:

H-Bridge Simulation

It’s a sim. It’s 100% accurate! Right? At least the digital ones are (… sometimes). I pouted a bit then rebuilt the circuit and it worked fine. Damn transistors. I am getting about 300mV drop from battery to motor, which might be an issue with NiMHs. I also discovered that, of course, there is a way to let out the magic smoke with this bridge, or at least trigger the getting-really-hot state. Both inputs high.

AVR

I also started wiring up the AVR:

AVR wiring

It seems my AVR Dragon would only program the chip at 3.3V. That isn’t really a problem though, as I’ll only do initial programming bootloader programming via ISP. Got the motors running with some test code. No luck communicating via serial yet.

Serial Interface

My original plan was to use an FTDI Cable for serial interfacing. But thinking about it more, it might be a reasonable idea to build a tiny interface board around an ATMega8U2 as used in the Arduino. It would be about the same price (Ignoring assembly costs for now), but it would make a nice starting point for future FT2232 / JTAG replacement work. Then I could replace the whole FT2232 section on the Noot with an ATMega and crystal. The Versaloon Code has a simple JTAG implementation that’d be worth looking at.

Encoding

Spent a while puzzling over potential wheel encoding solutions. I really think it’ll be necessary if you want to Logo turtle stuff. Options to investigate:

Detecting motor flux with a hall-effect sensor on the side of the motor?
Mounting a magnet on some shaft and hall sensing that.
Back-EMF, which would need a better H-bridge
Some sort of breakbeam / reflectance thing on the wheels.

If all else fails I can just measure battery voltage and guestimate motor speed (awfully).