After skipping April, I decided to pick up a pretty old project that I’ve had lying around for years. A long time ago, I wanted to get into working with electronics and microcontrollers, and bought a bunch of base Arduino kits. I tried to make a simple clock with them, had the time desync after a day or two, then just gave up. (I wasn’t even aware modern RTC modules could be so inaccurate)

This month, I still didn’t have a much better idea of what I wanted to do, but I thought I’d try to at least get something working. So I set the following goals:

• Make an Arduino based wake light. When the time approaches my typical wake time, it should slowly turn on an LED strip.
• Make so it can automatically fetch the current time, so it doesn’t desync even with bad RTCs.
• Give it an e-ink display to show a calendar for the new day.
• Try and make it power efficient

I think that sounded pretty reasonable, but, spoiler alert, I didn’t get it all to work.

The repo for this month is here.

Power struggles

When trying to power the LED strip, I ran into an issue that the software world just doesn’t prepare you for. I was using the USB connection to power the Arduino, and the Arduino’s pins directly to power the LED strip. This worked fine for simple testing, but when I tried to color the whole strip in white, an issue occurred.

As you can see, the LEDs are turning red. This is because the power supply is not supplying enough power. As I understand, red LEDs can run on a lower voltage than other LEDs. So those stay lit while all others shut off.

So I decided to use an external power supply. Instead of using the USB connection. It still tints orange, but it’s not as bad. Also, I intend to use a pure blue color for the wake light, so it should use less power than the white one.

I thought about maybe using one of those rgb LED strips that have their own power supply, and a remote to pick the color. Maybe I could hijack the control of that to power control the LED strip without much power draw on the Arduino. But it worked well enough from here that I didn’t bother.

The clock

I used an Arduino Mega R3 compatible board, since I wanted to use the many pins for the e-ink display later. It doesn’t have a real time clock, so I had to use an external RTC module. (Visible on the breadboard in the videos above)

I let it run for a day and then checked the output. It was already starting to drift, which was extremely disappointing. I did have the idea to maybe use some long term measurements to figure out how much the clock drifts over time, so I could maybe compensate for it. But I wanted something useful at the end, and connecting my alarm clock to my computer regularly to check time drift wasn’t very appealing.

So I ended up switching to an Arduino UNO R4 WiFi compatible. That one has a build in RTC, and thanks to the Wi-Fi, it can use the Network Time Protocol to get the time from the internet. The idea was for it to update every day, maybe once it gets close to the wake time, but looking at the code, it seems I took that out at some point.

There is still a serial command to make it re-fetch the time manually. But the build-in RTC seems to be more reliable than the external one. Still, I tried to write the code in a way that it supports both methods of time keeping.

The display

The e-ink display wasn’t too hard to get running. Although I found the documentation pretty bad, I did get the examples to work and could work off that.

WARNING: Video with rapidly flashing e-ink screen

Obviously, most of what’s shown here are just static images. But the basics are there.

Debugging hurdles

Something that ate most of my time on this project was trying to figure out issues, which I didn’t really expect.

Basically, since the Arduino is running in my bedroom without any PC, I don’t have any serial logs. And since most of the code was triggered by a specific time, they are difficult to fully debug.

Most of the issues I had was that I woke up, and the lights weren’t on. If I triggered them manually, they would work fine. Of course, you can just change the alarm time while debugging, but having to wait for the alarm to go off was still very annoying.

Calender request

This was the part where I really got stuck. For the calendar, I wanted to query Google Calendar for all upcoming events, and display them on the e-ink display. And at the time I couldn’t wrap my head around how to do that a limited memory budget.

Essentially, I needed to send and receive data over http, and then parse the JSON response. With preferably as little dynamic allocation as possible. And that problem essentially stunlocked me for the rest of the month.

So since I never had data for the calendar to display, that also never got finished.

Conclusion

At the month’s end, all I had was an LED strip that tries to wake me up. The core of the project was done, but nothing more. I might continue this in the future. There’s definitely a lot that could be done. But for now, I have some other projects in mind.