Week 6: Electronics Production

Assignment

Group Assignment

Please see this link (Session 3) for my group's assignment.

Group Pic

Website Update

As you may have noticed, this sub-page has a different look to it! During last week's presentations, I was inspired by Sara Ferandez's website. Its clean design and straightforward user interface was exactly the kind of aesthetic I've been looking for. I asked to meet up with Sara and she was kind enough to show me her method for documentation. This week, I am trying out her method. Thanks again Sarah for the tips and tricks!

Individual Assignment

Last week, we designed and simulated a PCB. I decided to design a board which demonstrated three different types of low pass filters. Although it was passing all of the design rules by the end of last week, there were a number of practical considerations that I needed to take into account. See below for what my layout looked like at the end of last week.

HW5 Circuit Board

Thanks to recitation, a really helpful design review session with Dimitar, insights from Matti and Miranda, chattng with Anthony after class, and many more hours of refinement than anticipated, I ended up with a board that looked something like this.

HW6 Circuit Board

Things I changed to improve manufacturability:

Next, I started working through the process of making my design come to life, using the Carvera Mill in the CBA electronics shop. I followed the instructions on Quentin's Carvera PCB Tutorial. Huge shout out to Quentin for this amazing document!

Step 1: Convert from PCB to Gerber to img

Use Quentin's gerber2img tool in order to convert the fabrication files exported from KiCAD (gerber files) to two color images. You need to make one for the edge cuts and one for the traces.

Convert Gerber2img

Step 2: Generate G Code

Next, I used the "Carvera Mill 2D PCB Program" which is hosted on https://modsproject.org/ to convert these images to GCode. Quentin provides directions for how to open and use the Carvera program in his tutorial. When I opened up the output, I could tell something was wrong because there were far fewer tracks than I expected (below left). A classmate helpfully pointed out that you need to make sure to check the "Black and white" rendering option on the gerber2img program. When I tried to export again, there were many more milling paths, but they weren't cutting my smallest traces (below middle). Those small traces are associated with the Op Amps in my circuit (below right). Turns out, that I had the board constraints in KiCAD improperly configured. My minimum clearance was 0.1 mils but needed to be 0.4 mils given the Carvera's tolrance. So, although my board was passing the ERC, it was improperly configured, which is why I didn't catch this issue earlier in the process.

Failed GCode

Next, I spent quite a bit of time re-working my traces in order to accommodate for the 0.4 mils clearance (see below). Fortunately, I had the pleasure of being around many of my fellow cba section classmates down in the electronics lab, which made this re-work much more enjoyable.

Op Amp Redesign

While I worked on this re-design, I couldn't help but notice that all the folks who had milled fine traces were no struggling to populate and debug their boards. I had a feeling that even though my board was now passing the updated design rules, that I too would struggle with populating more board if I went through with milling it.

It was a hard choice to make, because I was really eager to see this board through, but I'm trying to do a better job of "spiral development" and "supply side time management".

So. The next day, I opened up a brand new KiCAD project and got to work making a much simpler design: a board with a single passive RC filter using a potentiometer for the resistor so that I could dynamically alter the cutoff frequency. Functionally, this would also be cool and enable me to demonstrate a variety of low pass filters, but I would be able to use far fewer component and stick to bigger traces.

Step 3: (redesign) and use Carvera mill

Picking up where we left off, Step 3 was to use the Carvera mill to fabricate the board. Fortunately, Sara was setting up her board, so I got to watch her go through the whole process first, which was super helpful. Again, please refer to Quentin's Carvera PCB Tutorial for very specific step by step instructions.

In the end, it was really satisfying to go back through the whole process, starting from a clean slate and ending up with a fabricated board in a single day!

Op Amp Redesign

Step 4: Component Selection

Next, it was time to select and solder my components. Below, you'll find my notes from the process of component value selection.

Step 5: Code

First, I needed to follow this Tutorial from Seeed in order to load the SAMD21 board into the Arduino IDE. Then I figured it would be easy enough to find some code to generate sine waves, taking advantage of th DAC's full from ~0 Hz to 175 kHz!....what a silly thing to think!

Step 6: Testing

My plan of attack was to set the cutoff frequency of my filter at 318 Hz (potentiometer at midpoint = 5k) and then show that frequencies below 318 Hz should "pass" and frequencies higher than that should get attenuated. Alas, try as I might, I couldn't get this to work. All frequencies I was able to produce were the same magnitude at the output of my filter. Changing the resistance seemed to have an effect on the magnitude of the output, but this was independent of frequency.

I tried swapping out the capacitor (changing from 1 nF to 1 uF) but this was rather undirected. Essentially, I was out of time by this point. My next step would have been to cut the trace going from the output of my filter to an analog input of my microcontoller, for I fear that this was having an impact. But alas, that will have to wait for another time!

Testing my PCB

Acknowledgements

So many people to thank this week!

Return to Home Page