Some Proper Engineering

Today involved quite a bit of planning of how I'm going to continue developing the robot.

Motor Modelling

Firstly, I drafted a simple electrical model of a motor, which will hopefully be reasonably representative.  I've treated it approximately as a voltage source, where the voltage is proportional to the angular velocity (RPM), with a series resistance and inductance.  The torque that the motor generates is proportional to the current flowing in the circuit.  Note that the motor will have internal friction - using a simple model, this can be approximated by a leakage current bypassing the torque-producing current.  The magnitude of this is limited by the amount of friction in the system.

Optical Sensor Adjustment

I found a nut which is the right size for calibrating the gap between the optical sensors and the reflecting surfaces.  Actually adjusting the sensors was a challenge, as the wheel assemblies needed removing to access the screws that mount the sensors.  Ideally the sensor brackets would be reversed, but the mounting of the wheel assemblies makes this impossible.  Fortunately, the adjustment shouldn't need repeating.

When the wheel sensors are wired up, it should also be possible to adjust their position around the wheel to get a good quadrature output.  That should be quite simple using an oscilloscope with the appropriate motor running.  The wheel units won't need to be removed for this adjustment.

Motor Control Board

I decided that, before committing to redesigning the motor control board, it would be worth looking into modifying the programming of the PEEL device to allow PWM to be applied on the board.  The existing connection between the parallel interface board and the motor control board carries the three floor sensor indications, the four wheel sensor indications, as well as the four motor control flags.  Apart from the problems with extracting the PEEL from the board, there are spare connected lines that can be used to carry the PWM signals.

Robot Layout

Since I intend to scrap the existing boards, and I'd already removed most of them, I stripped all of the boards from the chassis and investigated the possibility of mounting the Mini-ITX board directly to the chassis.  To keep the mounting simple, any new PCB should be mountable on the same mounting holes as the Mini-ITX board.  I drew out the positions of the holes that would be needed on the chassis, in pencil, then sketched the layout I wanted on the robot.  In doing so, I recognised that the heatsinks and RAM modules on the PC board should be mounted front-to-back, so that the motion of the robot naturally provides some airflow to the unit.

Next week, I need to decide what hardware I'm going to use in terms of microcontrollers and the motor control circuit.  It'd also be wise to continue thinking about possible objectives.