Word Clock

Daily Journal

10/30 — 12/2

I spent the week of 10/30 learning about neopixels and ATTINY microcontrollers. I learned how to communicate with neopixels via libraries and how ATTINY412 and ATTINY1614 are sufficient for the project, yet much cheaper and more efficient than using an Arduino, as no excess features are being wasted. I also laser-cut the wood onto which I will mount my neopixels for the clock.

12/5 —12/9 & Weekend (12/10)

I spent this entire week soldering my neopixels (32 of them!) At the beginning, it took me ~20min per neopixel, but after adjusting to the process, I became much quicker. I found that putting solder on the neopixels, then stripping the wires and holding them to the solder as I touch it with the iron works the best. Every time after soldering another pixel, I plugged it into a test Arduino and made sure that there were no shorts/solder bridges anywhere. Then, I covered the connections in hot glue to make sure they stay in place. Also, I left a lot of extra wire coming off of the first neopixel so that I had room to correct errors in soldering when attaching it to a circuit. I ended up re-soldering this 3 times (once tip of wire broke, second two wires touched and the ends of the wires melted because of a short + too much current, third time with a heat shrink). There were several challenges I encountered throughout this process:

• Once ~8 neopixels have been soldered, they draw too much current for the Arduino to supply and the results become inconsistent. To solve this, I plugged an external power supply into the Vcc pin of the neopixels

• In the design, there is one section where there are not enough holes for the wires to pass through. To solve this problem, I simply connected those two neopixels directly without the wire passing through the wooden mount. I made this choice because I didn't have enough time to redo all of the soldering. Later, when I placed the clock face on the neopixels, this wiring got in the way of the cover, but I still managed to secure it in place.

• When I soldered, sometimes the outside of the wires would get too hot to hold by hand and melt. My solution was to hold the wires with tweezers and quickly melt solder then remove the iron so the wires don't melt.

• In the beginning, sometime hot glue would come off of the neopixel connections, so my solution was to simply add more hot glue so there was more surface area to adhere to. This solution worked excellently.

12/12

Today, I tested different combinations of wooden layer between the neopixels and the face of the clock to see what gave the best results, trying to maximize the brightness of the words while minimizing bleeding through to other words. I soon encountered a problem: the individual pixels shining were visible through the acryllic and words. To solve this, and after much testing, I concluded that two wooden layers with a piece of paer in between, acting as a diffuser, returns the best results.

12/13

Today, I soldered a ATTINY1614 microcontroller and tested it by making an LED blink. All of the soldering, suprisingly, worked on the first try, but when uploading the code, I got an error message. It turned out the problem was that the Arduino IDE has two different settings for ATTINY1614 board, one of which says "Optiboot" on the end. The one with Optiboot would not upload to the board, and after trying the other one, the blink code worked right away.