-
Notifications
You must be signed in to change notification settings - Fork 11
Design Rationale
Since the particles are recognized by the amount of light they scatter, the sensing volume needs to be entirely dark, except for the laser. Ambient light on the photodiode can saturate the amplified photodiode signal and produce incorrect particle counts. To prevent ambient light from directly entering the sensing volume, the air follows a circuitous route to enter and exit the air channel. The ‘analog cap’ on top of the circuit boards further prevents light from hitting the top of the photodiode.
At the beginning of the project, I thought that laminar airflow would be ideal in order to have consistent flow velocities, and avoid turbulent eddies that could result in double counting particles. At this point, I am not sure if that is necessary. However, I still tried to design the flow channel to encourage ‘nice’ flow. All bends in the channel are designed to be as smooth of transitions as possible, and the channels were CNC milled rather than 3D printed to have a much smoother surface finish.
We wanted the device to be as small as possible. After numerous iterations, this layout was the one that resulted in the highest packing density.
The photodiode and laser are directly referenced to the PCB to minimize other sources of alignment error. The laser mount is held in via dowel pins that mate to holes on the PCB. To prevent over-constraining the assembly, one of the holes for a dowel pin is connected to the laser mount via a small flexure. Though a pin and a slot configuration can be used instead on the PCB, this is not available from all board houses. The photodiode is a ‘reverse gullwing package’, meaning that it looks down through a hole on the pcb. This allows the board to be soldered as a single sided PCB.