Starting construction of H-bridge module.

I got the components from ELFA and Lawicel a few days ago. The first step was to set up the LM2574-based step-down converter, supposed to deliver a stable 12V supply to the HIP4081, on a breadboard. The schottky diode which is part of the circuit had pins so thick, I had to sandpaper them down to be able to push them into the breadboard holes... :)

I had no batteries at hand that could deliver the 24V I plan to use for the finished Segway clone construction, so I had to borrow a battery from my lawn trimmer, providing 21V. The circuit proved to work fine, at least on a quick minimal load test.

More components ordered.

Next step is to design a powerful H-bridge that can drive larger DC motors that can actually drive a full-scale Segway clone. Yesterday I ordered a bunch of components, most notably:

• 1 HIP4081AIP H-bridge driver IC.
• 4 IRFZ44N N-channel MOSFET transistor, capable of controlling 49 amps at 55 volts.
• 1 LM2574N-12 step-down switch regulator to provide the 12 volts required by the HIP4081AIP, from a 24-28 volt power supply (sealed lead-acid batteries).

My plan is to build a prototype motor driver on a breadboard and try it out for smaller currents, measuring timing and voltage levels on eg. the MOSFET gates. When it all seems to work, I will design a PCB capable of handling the full 25-40 amps required by larger motors.

One source of inspiration (an open-source, open-hardware motor driver):

http://www.robotpower.com/products/osmc_info.html

Balancing robot starts to work.

After a little Arduino sketch hacking the robot is now balancing. It works pretty well on carpets, not so well on wooden floors, probably due to small friction. Carpets seem to reduce oscillations in the control algorithm.