CircuitScout

The current control board is a board that holds multiple separate modules, and each module can control a stepper motor and a servo. All modules are connected through I2C. I thought this Is a cool design since it is a distributed system and each module can be custom programmed. However, as there are only 6 steppers, and the CircuitScout function is not that complex, this kind of structure introduces some problems, the host ESP32 module needs to have frequent communication with each module to check the machine's status and decide what to do next. This is not a very efficient way. If the whole machine is controlled by a single MCU board, things will get easier. So I designed a new version of the main control board.

It's an RPi Pico W + 8 A4988 module. The RP2040 is dual-core, so I can use one core for communication between the machine and the GUI, and another core used for machine control. It's basically a PicoW + an A4988 Arduino shield, but I really enjoy designing my own board :)

Currently, the probes are driven by tiny linear servos, but these actuators are too weak and a little bit noisy. A better linear actuator needs to be designed. Currently, one idea is using a stepper motor with a built-in gear head combined with a  long screw.

The schematic has all the information about the connections between components, and normally the PCB debugging starts from locating desired signals from the schematic. However, for some simple PCBs, it's not that difficult to recognize which signal a pad holds, and the pads picked by the system may not be proper or the ones the users want, so we add another pad-selecting approach: users can pick the pads to be probed directly from the PCB layout in the GUI.

we have designed two kinds of probes, one can probe vertically (the one we used here). Another one can probe from a tilted angle, which can probe a pad from the side. We decided to use the vertical probe since it's currently easier to realize. The probing position is not the center of the pads, but add a tiny bias from the center, so the probe won't be blocked by the components. We will improve the design allowing probe tilting for use in cases where a pad can’t easily be accessed only by vertical movement and/or component shape inhibits easy access to a pad.