Well, I finally got around to testing my code for the BoaArm, and I almost immediately ran into trouble. Firstly, the motor sensors are powered separately from the rest of the circuitry. After locating the pin used to enable the sensor power, and connecting it up to the new cpu, I ran into another issue, jitter.

When seeking to a position, the joint moves past it intended position, faster than program reacts. So what happens is that the motors seek back and forth, rapidly, trying to reach a position, but always skipping past it. There are two ways to fix this, I can either lower sensor resolution, so that the position its trying to seek to is wider, or I can decrease the programs response time.

Lowering sensor resolution isn't really a viable option, without first trying to decrease response time. There are several things I can try to increase response time. Firstly I can increase the speed at which the program checks and reacts to the motor position. I can try moving the motor control functions from the timer based interrupt handler into the ADC interrupt handler. As a last resort, I could also limit the boa arm to moving one joint at a time, which would decrease response time to a 1/3rd or a 1/6th of the current response time. Though only being able to move one joint at a time would not be good.

Over on the RoboPanda front, milw has completed his own cartridge and programmer. The picture does not show milw's final configuration, its lacking a pull up resistor for the !CS line of the original ROM, to prevent bus clashes when using the new chip. milw reports that, after he added the pull up resistor, the new chip appears to work perfectly. Which is great news. I was a little worried that the chip's power up requirements were not going to be meet properly, but I'm happy to be wrong about that. Now I just need to finish the programmer I am building, so I can try it myself.

Milw's Cartridge Mod