Electronics Production

20 September 2017 | By Casey Evans

“Be a pineapple: Stand tall, wear a crown, and be sweet on the inside.”

This is the schematic from Brian's programmer


  • Group Assignment: Characterize the specifications of your PCB production process.
  • Individual Assignment: Make an in-circuit programmer by milling the PCB, then optionally trying other processes.
  • Class Notes

    Other People

    We did "show and tell" at the beginning of class and it was pretty cool.


    I'll have to come prepared next time. This is an additional incentive to actually finish something on time. (sigh) Or maybe I'll just let my substandard creations live alone. I don't get how people knew this was happening. Now the room smells pleasantly of burnt cardboard though (think wood fire but slightly artificial-more like Christmas). Ooh one person mad a skyscraper light, some windows are open (vector cut) and others are closed (rastered). Hmm looking at these maybe a better way to parameterize is to somehow make the material thickness relate to a simple scale factor.

    Called out! Lol. Don't make an external site. Will start working on that. Also make really small files...maybe just these text blogs with a link to the wix site for more pictures as long as it lasts? 100KB images at web resolution. Check out Jake's site, they have tons of pictures but they are compressed. PNGs are great but JPEGs are easier to compress. Git can have branches--test versions and such that you can eventually merge into the final version. So no individual repositories. We'd have to work with everyone's stuff on a local spot. I'm not 100% sure how I feel about that. Just cuz my laptop is so old and sad that someone's accident could cause problems. Files, issues, wiki and mattermost make messaging possible on git. Issues are meant as the mailing list. You can use a custom notification setting to only receive notifications on issues. "Soft science and soft art, cuz it's all kinda bull" -Liz, new student. Fun honeycomb helmets Basata (cut, fold and link) by Mohammed, new student. Aluminum honeycombs in planes. Pinar has a great webpage. It also has great work throughs/how to use equipment and software. She did a really cool octopus jewelry stand. She also made a neat shirt. "Lasercutters are trying really hard to catch on fire" -Neil. So use low settings for fine details (it won't get up to good speeds). Masks for screen printing or heat transfer to do tshirts and stuff. To improve smoothness, scale up image (upsample) then use a blur filter. Illustrator and Inkscape both have image tracing. Dick Blick has screen printing stuff. Inkscape can also separate by color. Eleanor also has a nice page. Vinyl cutter tip: rolls v sheets need little rollers to be in specific areas. Optical sensing is there when you first put in the sheet (like a printer) so maybe try it again. Niki has a cool pinball machine. Margolus billiard ball universal theory paper. Chris Carter mad e a fancy bike treadmill. On a neat cardboard electronics press-fit thing: wipe and slide is good switching contact so compression of joint holds them together. The "City Light" is actually just modular city pieces. Still neat. Not sure who that was. Cardboard flexures are also cool. Music composition vs patents. Warning! It is very possible to get called out by Neil for not having stuff done before class if it shows up on your page. Ahh. Also be ready to explain why things aren't working for you. Also don't work the night before. Ever. Just get it done by the weekend I think is the best bet. API - allows automation of git functions. Document as you work -- check. You know what...I think I need to take each week as a "learning time" and not worry too much about its perfection. Set up a tracking page for the final project. Update each week with applications to the final project.

    This Week

    Building a programmer. Don't need to know how it works yet, thank goodness. Unfortunately that's the only part I really already know. PCBs are usually made by etching. Not very ecofriendly, but high resolution and a parallel process so you can do more than one board at a time. SDS - safety data sheets. We'll be machining boards. Mount stock on machine and then cut it away. Don't break endmills (tip of the drill tool). The bed is moved by the machine. There is a frame on the bed, then a sacrificial layer, then the piece. Including the tool creates a force loop. Underlay is right under the board. You want to make sure everything is clean. What are "shoplifting tags." You can sew with conducting thread! Endmill can't machine glass (fr4). Use fr1 instead. Not as good at high temperatures but it machines "beautifully". Kapton for flex circuits. Vinyl cut electrical tape as an insulating layer with vias between layers. 5-100$ for a two-sided board with 20ish boards. 3PCB.com makes 5 boards for 5$ from Shenzhen. Gold Phoenix is a wraparound Shenzhen. Trace width - how skinny they can be. 5/1000s of an inch it starts to have issues. "High density interconnect" is small and close (high density) and expensive. 1.5 = one sided with 0 ohm resistors as jumps. Solder masks limit where solder can go (it's not something you wear). Blind vias stop on an internal layer. Buried vias are between internal layers. Atting45 is the processor we will use (check out digikey). CSP - chip scale packaging. They just have pads, not legs. Neil hates breadboards and wants us to hate them too. I certainly hate them but probably not for the reasons he wants me to. Nevermind, he's talking about wires moving around, tracing wires to check them and stuff. Definitely an uncool feature of breadboards. Solder is a 'eutechtic'. To clean the iron, put solder on it and clean it off with a sponge. Heat both sides of the joint (pad and leg, for instance). Avoid cold soldering: heat joint - add solder - remove solder - let solder flow - remove iron. You can use solder as a temporary glue until you can solder it properly. Start at low things in the middle of the board to help with access. Oils from your finders over time will etch your board so you need to wash it. Don't mangle tweezers. 10kx10k image resolution limit. PNGs are good. Double stick tape to attach the PCB.

    Training Summary

    Trained by Gavin. More to follow. Notes taken hardcopy.


    Individual Assignments

    So this project was actually fairly simple. I decided to go with the recommended guide from Brian (TA) to make my programmer. With a bit of help from Alexandre I was able to get the printer to appropriately connect to the driver. But not without the printer deciding to go on a rampage and kill the bit by spinning up, flying across and down into the surface before stopping and trying to go deeper. Not so good but easy enough to fix. Apparently I was the first person to break one. I guess that makes me experienced. Here's what it looks like so you can be experienced by proxy:


    The first round it didn't cut deeply enough and talking with Alexandre it apparently has to do with the screw jigging it up. I tried again focusing on holding the end mill down and adjusting the depth to 0.0045" and it was still not deep enough so I changed it to 0.005" deep. That worked well.


    I was following Brian's page and part of that involved removing the tip of the programmer, which I did with a razor. I'm also going to recommend adding something on the bottom to make it slightly taller in a USB port.


    Once I did that I soldered the pieces according to Brian's page.


    And finally I programmed the programmer.


    It was really nice to see it work the first time. I was lazy so I used an ammeter to connect VCC and VPROG. It was very easy at the time but I skipped the rstdisbl step to check it worked and such (several week break) and when I came back to it I had much less success with that method so I caved in and used solder. Removing it wasn't nearly as terrible as I'd expected. I used a heated solder sucking gun (you held it to the solder, which melted, then you pressed a button and it sucked up solder). I'm still no good at the solder wick thing. I feel like I don't have enough hands to do it properly. Or else I'm still doing it wrong. Anyways, here is the successful command prompts from the Ubantu computer in the lab, which is what I used to program my programmer (more on that in Week 7).


    Just to cred Alexandre again, here is his page. I used his LED figures when I was making my Week 5 board to see if some unlabelled LEDs were red or green. He also did the group project print for EECS.

    Alexandre Akaspar's page

    Another useful link is the FabLab Inventory page. It includes the specs on all the parts I used.

    Group Assignments

    Alexandre did the line test shown below. It was able to do really well on carving lines inside its path but obviously it was limited by its own width in the inverse. Something I also noticed on my own design was that burrs popped up in the final product. Maybe it is in how the machine disengages from the material or in how the material reacts to the spinning based on humidity/atmosphere. I didn't notice any real damage to the circuit integrity, at least for the applications I think we will be pursuing. If we were dealing with high frequencies I might be more concerned.