Turns out Casio offers a handful of wristwatches with "10 year battery life". The claim is that thanks to "an advanced technology" the battery life in those watches is extended to ten years.

Now if you look at different models you see that they are rather complicated hence likely energy consuming - for example, AW-80-1AV model has both a liquid crystal display and hands and also it has LED illumination and a sound alarm.

I first thought that maybe the battery is the key. Model AW-80-1AV runs on CR2025. Energizer CR2025 datasheet specifies that this battery has nominal output voltage of 3 volts and nominal capacity of 163 mAh, so it stores 0,489 volt-ampere-hours of energy.

For comparison, typical basic model of Swatch run about three years on Renata silver oxide 390 (SR1130SW) battery that has nominal output voltage of 1,55 volts and nominal capacity of 60 mAh and so stores 0,093 volt-ampere-hours of energy.

So CR2025 stores about five times more energy, but the basic model of Swatch only has hands - no digital display, no illumination, no alarm, so it likely consumes less energy.

There clearly must be something more than a bigger battery that makes 10 years battery life possible.

How is 10 years battery life possible in a rather energy consuming wristwatch?

  • 2
    \$\begingroup\$ Mine lasted 10 years with the original battery. After that, I replaced it with generic VARTA cells from eBay and only achieve about 1-2 months. Gotta be a special battery ... \$\endgroup\$
    – arney
    Commented Oct 15, 2015 at 16:52

3 Answers 3


10 years =~ 87650 hours.
1 uA drain will require 87.75 mAh in 10 years.
With som shelf life degradation that's close enough to
= 10 mAh / uA / year or
= 100 mAh / uA / 10 years

So your cited 163 mAh battery will supply 1.63 uA mean.
Pushing technology, size and luck may get you to say 5 uA mean.

There are 86400 seconds/day. There are 1440 minutes/day.

You will find that eg alarm use is much restricted in the allowable use to get 10 years. If 1 uA of the drain is for alarm use then you get 24 uA.hr/day or 86400 uA.seconds or 86 mA.seconds. That's about 240 mW seconds at 3 V. Or say 5 x 50 mW x 1 second burst/day.

An LED can provide ample lighting at 1 mA. Use it 5 times/day x 1 second = 5 mA.sec = 5000 uA.sec or "only" 5000/86400 = 0.06 uA mean drain. Increase as desired and allowed.

Can you run a time keeping IC on say 1 uA?
Probably yes.

So overall it all falls in the area of "notionally possible if really really really clever and careful".
Casio can be expected to be quite clever by now.

Note that if any sort of energy harvesting is being used then all bets are on. Harvesting a uA or few sounds doable.


There are many others.

In September 2012 user Hli commented:

An EFM32, which is an ARM Cortex M3 MCU, can run on about 1.45µA while driving a LCD (550nA for the LCD, and 900nA for running the RTC and keeping its RAM). So a chip keeping only time should be capable to run on much less than that

The link he then provided is now broken, so:

EFM32 "Gecko" family are M0+, M3, M4 ARm Cortex microcontrollers from Silabs

Silabs EFM32 search

Wonder Gecko

  • EFM32™ Wonder Gecko 32-bit ARM® Cortex®-M4 Microcontroller Silicon Labs’ EFM32™ Wonder Gecko 32-bit microcontroller (MCU) family includes 60 devices based on the ARM® Cortex®-M4 core, which provides a full DSP instruction set and includes a hardware FPU for faster computation performance.

    Wonder Gecko MCUs feature up to 256 kB of flash memory, 32 kB of RAM and CPU speeds up to 48 MHz. The MCUs incorporate highly differentiated Gecko technology to minimize energy consumption, including a flexible range of standby and sleep modes, intelligent peripherals that allow designers to implement many functions without CPU wake-up and ultra-low standby current. With the lowest active and standby power consumption, the Wonder Gecko is the world's most energy friendly Cortex-M4 MCU.

Other xxx-Gecko variants M0+, M3, M4

Digikey listings of "Gecko" - legion

Lowest cost in 100's with LCD EFM32TG822F32-QFP48T$US2.03/100 Digikey

Lowest power useful mode with RTC running - EM2 - deep sleep

In EM2 the high frequency oscillator is turned off, but with the 32.768 kHz oscillator running, selected low energy peripherals (LCD, RTC, LETIMER, PCNT, LEUART, I 2C, LESENSE, OPAMP, WDOG and ACMP) are still available. This gives a high degree of autonomous operation with a current consumption as low as 1.0 µA with RTC enabled. Power-on Reset, Brown-out Detection and full RAM and CPU retention is also included.

EM1 - sleep

In EM1, the CPU is sleeping and the power consumption is only 51 µA/MHz. All peripherals, including DMA, PRS and memory system, are still available

EM0 - running

In EM0, the CPU is running and consuming as little as 150 µA/MHz, when running code from flash. All peripherals can be active.

So running in EM0 for 1 ms/s adds 0.15 uA to the EM2 standby load.

Overall, operating in EM2 at around 1 uA mean plus EM0 as required would allow the 10 years / 163 mAh example target to be met.


Energy harvesting:

Vibration and motion may well be possible energy sources.

A silicon solar PV/solar panel seems viable.
Very roughly power available is 150 Watts/m^2 at 1 sun = 100,000 lux.
A 10mm x 10mm "panel" at 10 lux at those ratings would provide ~= 150 Watt x (0.01m x 0.01m) x 10lux/100000lux = 15 microWatt.

10 lux is dim roomlight - at the level where colour fades into monochrome. Dim!
If that level of sensitivity can be maintained at such low light levels (as it quite possibly can with other 'chemistries') the light powering looks viable.

  • 1
    \$\begingroup\$ It's not only timekeeping that works continuously - there's also a display and hands that work 24/7/365. \$\endgroup\$
    – sharptooth
    Commented Sep 18, 2012 at 8:16
  • 1
    \$\begingroup\$ Energy harvesting sounds likely. LCD can be lowish power. Hands are annoying. If they step at 1/second and if we assign them 1 uA then they get 1 uA.sec.sec = not much. \$\endgroup\$
    – Russell McMahon
    Commented Sep 18, 2012 at 8:19
  • 12
    \$\begingroup\$ An EFM32, which is an ARM Cortex M3 MCU, can run on about 1.45µA while driving a LCD (550nA for the LCD, and 900nA for running the RTC and keeping its RAM). So a chip keeping only time should be capable to run on much less than that. \$\endgroup\$
    – hli
    Commented Sep 18, 2012 at 9:32
  • 3
    \$\begingroup\$ "Can you run a time keeping IC on say 1 uA? Probably yes." The PCF2123 can run at 100 nA according to the datasheet. So, yes. \$\endgroup\$
    – Renan
    Commented Nov 13, 2014 at 20:21
  • 1
    \$\begingroup\$ Curious that you didn't mention the Citizen Eco-drive watches that have a solar panel embedded in the face and advertise no need to ever replace a battery. I've had one for well over ten years and it has the main 3 hands (hour/min/sec), four other dials, two small LCDs, a LED, and receives radio signals for syncing with the atomic clocks provided by the government. \$\endgroup\$ Commented Aug 24, 2017 at 22:02

Perhaps those watches use some kind of energy harvesting system to recharge a rechargeable battery?

An automatic quartz watch has an energy harvesting mechanism that, like a mechanical self-winding watch, pulls small amounts of energy from the day-to-day motions of the person wearing it.

A solar-powered watch uses a tiny solar cell to pull energy from ambient light. (Even indoor light, much dimmer than sunlight, is adequate to keep the watch running).

I hear that such watches typically run normally for a day or so when cut off from outside power (taken off the wrist, put in a dark room, etc.). Then they enter low-power state where everything is turned off except internal timing -- the LCD display goes blank, the hands stop. The watch has a power reserve that can keep internal timing going for at least a month; one manufacturer claims it has watches with a 4 year power reserve. Then when you pick it up and shake it, the battery starts to charge up, and the "hands will magically spin to match the current time." ( a ).

Are you looking for detailed information on how it's possible to build electronic devices with extremely low power consumption? Then you might enjoy reading Jeelabs's notes on low power electronics ( b ). One JeeNode has run for over 2 years on a single battery charge (charged on August 21st, 2010; still running and continuing to count on Sep 15, 2012). ( c )

Or are you looking for techniques to keep a battery from failing prematurely? While recharging a battery many times over 10 years is "easier" than trying to get a primary cell to last 10 years, it doesn't make it "easy". I've bought new rechargeable batteries for several consumer electronics devices when the original rechargeable battery went dead in less than 5 years -- not merely drained, but completely dead. (What makes this especially frustrating is when 5-year-old equipment uses batteries that are some special shape that stopped production years ago and are now unavailable, and I suspect all the shiny new oddly-shaped batteries will likewise be unavailable in 5 years.). How to keep batteries from failing prematurely would make a good separate question -- I wish I knew the answer.

  • 1
    \$\begingroup\$ If the charger is not optimised for battery lifetime, then taking the battery off charge once charged and not allowing it to deep discharge will together increase lifetime. LiIon batteries have a calendar life regardless if use. NiCd and NimH do not. LiFePO4 may not (not much comment available thereon). \$\endgroup\$
    – Russell McMahon
    Commented Sep 19, 2012 at 17:02
  • \$\begingroup\$ Then it would be innovative-my-bum: Working solar+storage watches with no primary batteries involved have been in that manufacturer's portfolio for probably more than 10 years. \$\endgroup\$ Commented Oct 14, 2016 at 9:10

The key here is that it has no sweep second hand. The minute hand is incrementing only once every 20 seconds and that's the only movement that's occurring to the hands.

  • 1
    \$\begingroup\$ Does that really make so much difference? Don't all those circuits and display elements drain notable currents? \$\endgroup\$
    – sharptooth
    Commented Nov 14, 2014 at 9:15
  • 3
    \$\begingroup\$ @sharptooth LCD elements are voltage driven, so they are operates with small currents. (Having said that, I don't have any numbers for what it takes to switch one segment in a typical watch.) \$\endgroup\$ Commented Dec 23, 2014 at 20:27

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