You have to be pretty ambitious to modify a clock to run for 50 years on a single battery. You also should probably be pretty young if you think you’re going to verify your power estimates, at least in person. According to [Josh EJ], this modified quartz analog clock, which ticks off the date rather than the time, is one of those “The March of Time” projects that’s intended to
terrify incentivize you by showing how much of the year is left.
Making a regular clock movement slow down so that what normally takes an hour takes a month without making any mechanical changes requires some clever hacks. [Josh] decided to use an Arduino to send digital pulses to the quartz movement to advance the minute hand, rather than let it run free. Two pulses a day would be perfect for making a 30-day month fit into a 60-minute hour, but that only works for four months out of the year. [Josh]’s solution was to mark the first 28 even-numbered minutes, cram 29, 30, and 31 into the last four minutes of the hour, and sort the details out in code.
As for the low-power mods, there’s some cool wizardry involved with that, like flashing the Arduino Pro Mini with a new bootloader that reduces the clock speed to 1 MHz. This allows the microcontroller and RTC module to run from the clock movement’s 1.5 V AA battery. [Josh] estimates a current draw of about 6 μA per day, which works out to about 50 years from a single cell. That’s to be taken with a huge grain of salt, of course, but we expect the battery will last a long, long time.
[Josh] built this clock as part of the Low-Power Challenge contest, which wrapped up this week. We’re looking forward to the results of the contest — good luck to all the entrants!
24 thoughts on “Low-Power Challenge: Making An Analog Clock Into A Calendar With A 50-Year Life”
You probably could use less power with a cpld instead of a micro
The calculation is missing the time it takes to tick at the 55ua current. That can then be used to convert the 6 pulses to an average current, then it can be used with the avrage current of the arduino to calculate the battery life.
Try this calc in advanced mode
Microamps per day make my brain hurt.
Yeah… Especially since the definition of an amp is coulombs per second. Microamps per day would have units of coulombs per second per day. Just as bad as measuring speed in mph per day.
… and in which unit you measure the capacity of a battery, tell me ?
People measure battery capacity in (amp)(hours) or Ah or mAh. For example, what is the capacity of your phone battery? Almost certainly it has a mAh rating. That is current times time, not a current per time.
Mph per day would be a measurement of acceleration. Does such a thing as current-acceleration exist?
Yes: In op-amps, that would be voltage slew rate, where you have a resistive load.
Or of range
The project documentation is much clearer than this confused write-up. 6.2 uA average which translates to 46-55 years on a 2500-3000 mAh battery. The original author is also very clear that this is theoretical and even mentions battery shelf life of 10-15 years.
I recall back in the day when National Semi came out with the LM3909 low voltage LED flasher, that in some cases the battery would last longer flashing the LED than it’s static shelf life. Somehow the little current pulses going through the battery in series with the cap would reverse the effects of self discharge to some extent.
That would be microamp-hours per day. 6uAh per day over 50 years is only 110mAh. The current draw is about 6uA continuously while the controller is running, and the time the hands are moving is so small overall that the 10x increase in current adds just about nothing to the total energy usage. This is just over 2600mAh over 50 years (I always like to check the numbers).
I must say I’m a little surprised the controller runs at all on only 1.5V! I’m not sure why you would still run it at 1MHz though, instead of just using the 32767Hz oscillator that also runs the RTC?
Is that 6uAh per day?
I know this isn’t really the point but it also kind of is.
The clock doesn’t have a 50 year run time because the batteries it’s designed to run on will self discharge in about half that.
I’ve no idea what kind of battery could really last 50 years. And I’ve been googling for a while and can’t work out what could.
Only thing I can think of is a dry-pile battery, but no idea if it can power this. One has been ringing a bell for almost 200 years
Supposedly the bell draws 1 nano Amp of current.
So, in theory, a 9000 V dry-pile, stepped down to 1.5 V could run this clock. Maybe 12kV to overcome losses.
As for how to step it down, one idea would be to a set up like the above bell, with the clapper connected to one side of a transformer and the other side connected to the far end of the dry-pile. This would, in theory, create pulses at the secondary. Then use a capacitor to smooth the current.
An AA cell won’t last 50 years, but a small PV panel and supercapacitor might. Solar powered calculators have no trouble running on just room light. I have some calculators with them that are about 30 years old and still work.
Given the time scale, you could also run the micro’s clock a *lot* slower. Digital watches usually run at 32.768 KHz, and can run for years on a tiny coin cell.
NiFe might be a reasonable option, though I think 20 years is a typical lifetime expectancy
Let’s clear up the hype around the infamous phrase “6 μA per day.”
Continuous consumption of 6μA for 24 hours equates to a daily requirement of 148.8μAh or 0.15mAh (note that the actual value in the documentation is 6.2μA and not just 6μA).
So, we are now back to a typical power requirement unit.
When it comes to a wall clock, there is plenty of space for a small solar panel like those used for garden lights.
This solar panel can fit inside the wall clock and can power a super capacitor coin cell, for example.
A small solar panel available on Amazon has a rating of 6V/1W, which means a peak power output of 167mA. Let’s assume 150mA.
So, the small solar panel only needs 0.001 hours or 1.5 minutes of solar light to charge a super capacitor coin cell. Indoor, we can assume a pessimistic efficiency of 25%, so with just 6 minutes of indoor light, we are done.
The smallest coin cell super capacitor stores 22mF @ 5.5V, which is tens of thousands of mAh (slightly less than 30kmAh).
A super capacitor is a much better solution than an alkaline battery for long-term use. It is also maintenance-free.
I hope my calculations are correct. Please let me know.
Sorry, the smallest coin cell super capacitor stores 22mF @ 5.5V, which is 589 mAh.
Good ideas. However, most PV panels are optimized for sunlight (high light levels). They deliver essentially nothing in room light.
But there are different models that are optimized for room light (low light levels).
They are not optimized for sunlight, since sunlight is mostly “green”, while silicon PV cells are most sensitive to near-infrared.
Please be kind and respectful to help make the comments section excellent. (Comment Policy)