Danny Griffin HTMAA 2023

Week 05 :: Electronics Production

How do we prototype a circuit design?

1.1 Group Assignment:

Characterizing the in-house rules for our in-house PCB Production Process.

Responsive image — Working with Simon Lesina-Debiasi, Alex Htet Kyaw, and guided by Jen O'Brien.

First steps for this week involved a hands-on demo of PCB milling using the Modela machine in RPL.

It's interesting to think that the digital side of the setup is basically "set," and the parameters we control are the tightness of the physical setup. The endmill is adjusted via the set screw on the collet.

Using double-sided tape, we fix the FR-4 board as firmly as possible to the spoilboard.

The Mods setup to generate gcode is quite similar to the vinyl cutter. The main distinction here is the layer of selecting the appropriate endmill, since we use a 1/64" bit for milling traces, and 1/32" bit for cutting the outline.

Milling the provided test file, we see that .010" is seemingly a reliable dimension for use with the 1/64"(.015") endmill. This relationship gives us the rule-of-thumb ratio: [.01 to .015]

Handling the board, we got a feel for the roughness of the copper. Scraping the surface with a metal ruler was suitable for removing many of the resultant burrs.

1.2 Group Assignment:

Send a PCB out to a board house.

Importing Gerber format files into Kicad's Gerber Viewer.

Exploring the layers, I also came acrosss silkscreen vectors

I decided that I would try to use PCBWay to order a breakout board for the ESP-32CAM, Specifically this one from Electronic Clinic.

I just downloaded the files which were provided on the breakout board page, and quickly realized that I couldn't open these filetypes. Although the Gerber files are ASCII text, they're unintelligible to edit from a text editor. It seems like Eagle would be the program of choice for

To see how I use this PCB when it eventually arrived, please see my Final Project Progress

2. Make and test the Development Board that you designed.

My first step for this week was to revisit last week's assignment, and design my own PCB more in line with my final project goal. This schematic shows the XIAO ESP32C3 connecting to two output devices, a DC vibrating motor, and an OLED display. I've tried the OLED display in week 02 with the XIAO RP2040, so I have some familiarity with the SDA/SCL pins used for serial communication. The aspect I'm less sure about, is driving the DC motor, which is requiring me to dive into transistors, resistors, and diodes.

I have not tested this circuit layout yet, so I'm making some assumptions about what might eventually be necessary. Specifically, I'm anticipating that eventually, I'd like to use the battery (+/-) pins on the underside of the board.

In designing the form factor of the PCB, I'm trying to think about the DC vibration motor one one side, with a display OLED on the other.

Milling process in action.

My first attempt is a catastrophe, which I can trace to two errors. Firstly, I clearly did not use enough tape to fix my stock onto the spoilerboard. You can see this by the uneven depth of milling across the surface, where one side of the boarding was coming loose and kicking up from the surface. The second error is a blunder on my part - I did not realize that the settings in Mods had been set to 40 offsets by the previous user, instead of the recommended 4. This oversight is the reason that so many unnecessary toolpaths were run.

A better outcome on the second try. This image shows the board drying after using Gojo soap for cleaning.

First attempt at stuffing a board. I made the mistake of using non-lead solder for the very first blob to try to stick the ESP32C3 onto the board. I found this really difficult to work with, especially for the initial positioning of elements.

One mistake I made was failing to account for the fact that the OLED screen that I had been using has some header pins already soldered onto it. I realize now that I can incorporate some wired connections from my board to this element.

First attempt at desoldering. I was too focused on making nice-looking blobs, and didn't think about how much excessive solder I was adding to the board. When I wanted to make changes, the process turned out to be quite unforgiving.

Unfortunately, the board does not perform any of its intended functions, so far. Now that I've seen the process through from schematic to physical board, I at least understand some of the parameters that I'm working with.

To get the board functioning the way I want, I think I need to break out the steps a bit more systematically. I tried too many things in one round of development, so I'm left with a debugging situation that's not in my favor.