I am trying to create a tiny little watch that is supposed to run off a coin cell (because it is so tiny) and basically just a microcontroller with a 30ppm 1Mhz oscillator that counts upwards. The whole thing is then supposed to be controlled by a single button that lets me change the time by flipping bits. The idea was to connect 11 LEDs to the MCU (5 for hours 0-24 and 6 for minutes 0-60) and mount the battery to the back and the mcu+osc and the led on the front, so it looks a bit techy. The problem is, that low power LEDs that would run for a sufficient time >6 months without battery replacement, are barely visible in daylight and those that would be needed too much energy. So here's my question: Is there anything that is sort of similar to an LED in that they are simple to control by a tiny microcontroller and that have a clearly visible state change, eat little energy and readily available and fit the purpose? (I am planning to add a second button that just connects an LED to the battery as a way to use the watch at night).

I was thinking of just using one of those calculator B/W LCDs with a fixed display, but for a binary watch, I would need to have a special one printed which is not very easy.

I was also thinking about just disconnecting the LEDs by default and having to press a button to show the time, but I have something in my right hand already during most of the day.

Does anyone have any idea about this?

  • \$\begingroup\$ If the requirement is to be visible in daylight, consider using LCD (Liquid Crystal Display) instead of LED. Since a reflective LCD can have good contrast in sunlight / ambient light, this greatly reduces the power supply requirement. \$\endgroup\$
    – MarkU
    Apr 12, 2019 at 7:57
  • \$\begingroup\$ Yeah I guess thats gonna be the only way. But I would need to have them printed \$\endgroup\$ Apr 12, 2019 at 8:01
  • \$\begingroup\$ Can I buy single LCD pixels (like just a black block that turns on with current)? That seems like the ideal solution \$\endgroup\$ Apr 12, 2019 at 8:06
  • \$\begingroup\$ A small e-paper display? Just noticed someone has already suggested. \$\endgroup\$
    – HandyHowie
    Apr 12, 2019 at 8:51
  • \$\begingroup\$ Could you use bright LEDs with a sunlight sensor so they only shine bright when the sensor detects light above your choice of brightness? \$\endgroup\$
    – C. K.
    Apr 12, 2019 at 9:10

3 Answers 3



E-paper is as low power a technology as possible. because it does not need power to maintain its display. If you do not display the seconds on your clock, you could have your microcontroller sleep for a vast majority of the time, extending battery life drastically.

you could also add the LEDs for night-time display, but have them run from a second battery to avoid compromising the main battery's life.

  • \$\begingroup\$ thats actually pretty interesting, even though pretty overkill. Although I have trouble finding one small enough to fit a watch \$\endgroup\$ Apr 12, 2019 at 8:24
  • \$\begingroup\$ this one, but it is a bit pricy shopkits.eink.com/product/… \$\endgroup\$
    – Sclrx
    Apr 12, 2019 at 8:28
  • \$\begingroup\$ thats definietly out of my price range, but thanks anyway \$\endgroup\$ Apr 12, 2019 at 8:29
  • \$\begingroup\$ Yuo might want to look at waveshare then, the price is much more reasonable, however the display is not round, said that, it would't be the first rectangular display watch. waveshare.com/product/modules/oleds-lcds/e-paper.htm \$\endgroup\$
    – Elmesito
    Apr 12, 2019 at 8:56
  • \$\begingroup\$ Good Dispaly has a 1.02" panel and it costs less than $5 \$\endgroup\$
    – Gilad
    Apr 12, 2019 at 9:19

Interesting "bare metal" watch project. I think the hardest tasks will be sourcing a suitable display and managing the power budget.

Back in the mid 1970's, my father was involved in one of the early attempts at using LEDs to make a digital watch. (This was back when red was the only available color.) These suffered from low contrast in daylight viewing, and they required the user to press a button to momentarily turn on the display. Otherwise the LEDs would quickly drain the coin cell battery.

Digital watches became practical with the development of LCD displays. These are made with a thin layer of "liquid crystal" material sandwiched between two panes of glass, with clear conductive traces printed on the glass itself. Driving an LCD is fundamentally different from driving an LED: the drive polarity needs to be switched back and forth (AC not DC), otherwise the liquid crystal material gets an image permanently "burned in".

Getting a custom LCD glass is a bit costly because you have to pay the Non-Recurring Engineering costs (NRE), but for a watch application there should be suitable units already available in stock. Additionally many LCD display use triplex multiplexing. The logic to drive a triplex display is a bit complicated, but it's included in the custom timekeeping ASIC that the watch vendor would use. If you've never seen a digital watch teardown it's worth looking at; the ASIC is mounted on a PCB and "zebra strip" is used to contact the LCD glass directly. The whole thing is held together by the outer case. So if you're truly going this route, some level of mechanical engineering design is needed. It is usually also possible to purchase LCD glass that have metal contacts bonded to the glass, which can then be soldered to the PCB or plugged into socket strip.

You could try using an LCD Graphic Display Module, which has an integrated display controller included -- so all you have to do is send command codes. A quick search at mouser.com for LCD Graphic Display Module turns up this one from Newhaven display, needs 3V@1mA and also a 6V supply for contrast -- so you'd need a charge pump doubler, but could probably drive it from a 3V lithium coin cell for awhile... though read the CR2032 coin cell datasheet too. For a CR2032, 1mA drain is possible but kind of a lot. May be OK when active but display should be off during sleep. This kind of display has an integrated display controller that expects to receive command codes from a microcontroller, so it might not be right for your project.

Another technology worth looking into would be E-Ink Display (or E-Paper), which would seem ideal for daylight readability. Such displays are available through Adafruit and similar distributors, but they look a bit pricey. Also e-ink displays were originally developed for the Amazon Kindle e-book reader, so the displays tend to be more like the size of a tablet rather than the size of a watch. And, like the graphic LCD, you're more likely to find it as a whole integrated controller + display module, rather than just the bare display part. So again this might not be quite right for a true bare metal project.

Given the requirements of this project, I'd suggest this approach:

Research the internet for LCD display modules, LED displays, E-paper, etc. Most likely you won't find exactly what you need, but you can get an idea about cost budget, and which companies manufacture these units. Most important, see if you can find datasheets. (Must. Have. Datasheets.) Typical modern datasheets include web links to the manufacturer's website. Check the manufacturer's website, see if they have any way to contact them for applications support. (Note: I am an applications engineer at a semiconductor company, we love hearing what our potential customers are trying to do.)

It couldn't hurt to ask these manufacturers what it would take to make a custom glass, what the minimum lot size is, how much the setup and per-unit cost are... and ask if they may have something close to what you want, that you could use for a prototype. Only they can tell you what's possible for them and how much it costs.

You might also try contacting the folks at Sparkfun or Adafruit (who are more like hobbyists who became electronics distributors out of necessity), since this could be an attractive type of project (low-power, coin cell watch battery form-factor, simple/primitive/bare-metal display, maybe somehow "IoT" related...?). I think if you get the display design right, you could enable a lot of other similar projects.


LEDs are currently the lowest power, most efficient technology we have. They are also cheap, readily available and easy to control. Organic LEDs might be a bit more efficient (or so the manufactures would have you believe), but are harder to find and more expensive. If you need to to be brighter, you need more to use more energy. A bigger battery or replace it more often are your only practical options.

That being said, you could use the brighter LEDs and PWM them so that they aren't quite so bright, so use less power. You could have some input to set the brightness of the LEDs (varying the duty cycle of the PWM), that way you can easily tune the brightness of the LEDs for the environment they're being viewed in.

The better solution is, use a larger battery pack. A pair of AAA could store more energy than a coin cell (depending on which coin cell you've chosen), and aren't that much bigger.

EDIT: Forgot to mention, if you look at LED binary watches which are available at the moment, most only have the LEDs on when you press a button, the rest of the time the watch face is dark. Clearly this means that a lot less power is used, as the LEDs are only drawing power for 5 or 10 seconds at a time.

  • \$\begingroup\$ Aas dont fit in a watch. But adapting brightness based on time is pretty interesting. Although the MCU would need to stay awake the entire time to drive the PWM instead of me paying able to put it to sleep for 2 seconds most of the time. \$\endgroup\$ Apr 12, 2019 at 8:11
  • \$\begingroup\$ Read my post again, an led button is not a solution for me, or I would have bought such a watch \$\endgroup\$ Apr 12, 2019 at 8:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.