Origin Story

Many years ago, I came across the Time Since Launch clock capsule from CW&T. I remember being instantly amazed by the simplicity of the concept and the elegance of the design. I am not a particularily nostlagic person but the idea of having a discrete and simple reminder of a happy moment from my life spoke to me. I also remember the moment I saw the price of the clock and being outraged. How can they justified 600 USD for a clock? I was much younger at the time but I was further enougth along on my electronics education to know that keeping time is among the most basic thing any electronic system can do. I kept the link to the clock in one of my bookmark folder and moved on.

Fast forward a few years and I am now in the fourth year of my Electrical and Computer Engineering PhD. I come across this bookmark again and I go through the exact same two steps: “Yea I remember this thing, this is such a cool idea!”, immediatly followed by “WTF 600 USD for that clock!? Thats around 1k CAD!”. Except this time I did not stop there. For my PhD I learned to design PCB circuits. Before, I knew the concepts involved but I never actually designed one from begining to end myself. After learning (and failing many times), I can now better appreciated the efforts involved in manufacturing a battery-powered system that is meant to work continuously for 100+ years. I am not qualified to determine if the price of the Time Since Launch clock is fully justified but at least I understand that is takes some efforts to build it.

Now with my newfound knowledge of PCB design, an obvious questions immediatly came to mind: Could I built it myself? I mean, surely I can build a crappy version that would keep the time and update a display, thats not even a challenge, but could a make a decent enougth version that I would not be ashamed to show to other people? I would have to cut some corners and adapt the design but I think I could do it. Let’s find out!

The Plan

Ok let’s lists the components involved in this type of product. The goald it not to build a one-to-one replica of their product (as if I could…) but to get a similar function. I want to identify the functional blocks, not the exact components.

• Source of energy: We need power to come from somewhere. The original capsule seems to use two cylindrical batteries. I don’t know the specifics but I suppose they are some kind of long-life batteries that would cost me an arm and a leg to acquire.
• Time keeping circuit: Some kind of clock. Either a dedicated Real-Time Clock (RTC) chip or something integrated in a microcontroller. The precision is not the main focus here because I do not care if there are a few minutes/hours of drift per year. All clocks are wrong and I am fine with mine being a bit more wrong than the other.
• Display: The original capsule seems to use a segmented LCD display. This is a good choice for both simplicty of interfacing and low energy consumption at rest. However these are not the most popular technology anymore so sourcing this type of display is not easy and the choice is limited. Other types of display might be better suited for me.
• Design: CW&T put a lot of attention on the design of their products. Their if-I-see-color-I-die-in-real-life apporach to desing is not what I prefer but there is no denying that it looks sleek and discrete. Maybe too discrete for my taste.
• Trigger mechanism: They nailed it with their pin-pull trigger idea. It looks very satisfying to pull and there is the hidden expectation that something will explode. I like this idea but there might be other cool and satisfying mechanisms that fits this project.
• Some kind of enclosure: I would like my clock to run as long as possible and enclosing it in a secure case seems like a good idea. 3D print a box and throw a couple desiccant packets in there is a good reduction to practice but eventually I would like to make it better. Maybe a metal and acrylic enclosure with a seal to hold a negative pressure could work well. In any case, there need to be something to hold the components and protect them.

We now have the core functions, let’s look into actual components.

For the power, I will start with two AA batteries. I have no idea of the shelf-life of these batteries but for now I want something simple and inexpensive. Two batteries in series provide 3.3V, an easy voltage to use, and should contain around 4000mAh. If we want to power this device for 100 years, we need to limit the average current to around 5mA. We can do a lot of things with 5mA!

With that constraint in mind, I started looking for microcontrollers and displays. For the microcontroller, I don’t need much. If it includes an RTC, it only need to connect with the display. I mostly want something that consumes very little power. My personnal favorite, the Raspberry Pi Pico 2, with its 15-100mA of consumption when running and 40 pins is vastly over powered. The ATtiny85 seems more reasonable with only 8 pins and 4mA of consumption. I never worked with it but I found a nice little breakout board from Adafruit. It adds a USB ports for easy flashing and some voltage regulation.

Now for the display, I reduced the choice to e-ink panels or segmented LCD. While I like the look of 7-segments displays, sourcing them seems to be difficult and they require a LOT of pins to control (at least one per digit plus seven for the segments). That mean they require a driver and while I like the idea of using shift registers, that seems like a lot of trouble for a very rigid setup. The second option, e-ink displays, is appealing but not without drawbacks. One one side, it does not use any power when not is use and they are available in a variety of sizes and sometime colors. They also require a driver but these are also easy to find. One the other, they are very slow to refresh (compared to other displays). A small binary display takes at least one second to refresh. With that in mind, I chose to not display seconds. Removing seconds enable the use of e-ink and provide ample room for putting all the components in sleep mode for 59/60 seconds. I would then display minutes, hours, days and years when the origin clock displayed seconds, minutes, hours and days (who has time for converting days to years by head…). I selected the 1.54’’ (4mm in normal units)e-ink display from Adafruit because it includes the driver, can buffer frames (usefull because the ATtiny85 is pretty limited in memory), has an enable pin that can disable the extra circuitry in, and is available from Adafruit so i can place a single order.

Summary

Two AA batteries provide 3.3V to an ATtiny85 microcontroller and an e-ink display with its driver. The microcontroller displays the time, and then goes into deep sleep for ~59s with the RTC providing the wake-up signal. When waking up, the time is increased by one minute, the display calculated, and the cycle repeats.

The parts I chose are only for prototyping and if all goes well I will make my own PCB to integrate everything nicely. Hopefully Everything works together and I can focus on achieving the lowest possible power consumption. Next part when I receive the components.