I'm learning a lot of interesting things as I take apart my retired Canon Pixma MX340 multi-function inkjet, like how they implemented freewheel capability with just a single coil of metal between two gears. Another ongoing lesson: my design priorities for a hobbyist project differs from Canon for their mass produced commercial project.

Here's the latest example: once I've removed all gears and cams from the paper tray gearbox, I'm left looking at this trio of levers. (Highlighted in pink) Each actuate some aspect of the paper tray during its page-feeding process. If I were to implement such a mechanism in a 3D-printed hobbyist project, I would install a micro servo for each of these functions and write microcontroller code to fine-tune coordinating servo motion timing. Come to think of it, Canon engineers probably have exactly that rig on their workbench. They would need something dynamically adjustable so they can fine tune timing in their sequence, test performance of one design against another. Once a sequence is declared to be the best tradeoff across all design criteria, though, what happens?

For a hobbyist project, I would stop there. I push my microcontroller source code (Arduino sketch or similar) to GitHub, along with 3D printing CAD/STL and maybe a KiCad schematic, then I move on to my next project.

For Canon engineers, there's a lot of work still ahead. They will need to convert their software-defined paper feed sequence to mechanical hardware. Translating servo-actuated motion into lever-actuated motion, design cams to move those levers, and a gearbox to turn those cams to the right position at the right time. This takes work but it is an one-time up-front investment. Once the design is converted into injection-molded plastic, it can be cranked out at a tiny fraction of the cost of manufacturing something with multiple servo motors.

For a MX340, this trio of levers marked the end point in my effort to trace through all activity powered by the paper feed (Y-axis) motor. Backtracking from here, there was the big paper tray rubber-coated roller, ink disposal peristaltic pump, print carriage parking pawl, front door panel opening mechanism, and probably others I didn't even notice. If I were to do it my easy way, that's at least eight different motors. Canon engineers put in the time to optimize down to a single motor.

This results in a lower per-unit parts cost and higher long-term reliability (cams > servos) which more than makes up for the higher up-front design effort when amortized across thousands (millions?) of production units. But it wouldn't make sense for my own projects with a production volume of one.

This teardown ran far longer than I originally thought it would. Click here to rewind back to where this adventure started.