I am using an ATtiny85 to sleep for a long period of time, collect some data, send it over radio, and go back to sleep, after disabling everything possible. The ATtiny is currently powered by a coin cell. I can get quite some time off that coin cell, but it's simply not enough, and the coin cell cannot be changed once the board is deployed.

My aim here is to find an alternative power source for my board. Some good news is that the boards are never more than one or two feet from a receiver, so rf-like power transmission may be possible. Some other options may be motion or heat (it's a wearable device). The amount of power could be pretty small, as I only need it at peak periods, so I could just charge up a supercapacitor for a while and drain it when power is needed after the Attiny wakes up for less than a second.

The ideal situation would be to have a means of power transmission or generation that provides enough current to continuously power the microcontroller without the need for sleep, but I recognize that this is probably not feasible. However, due to the limited range it might be. Another concern is radiation to the wearer.

If anyone has any experience in this kind of power transmission or generation or just any ideas I would love to hear it. Thanks!

  • \$\begingroup\$ Peak current, and average draw? \$\endgroup\$ – Passerby Jan 12 '16 at 1:02
  • \$\begingroup\$ Sorry, should have included that. Peak is approx. 15mA @ 3v and average is around 5uA. Peak could be decreased to about 10 mA if need be. \$\endgroup\$ – Alex Wulff Jan 12 '16 at 1:03
  • \$\begingroup\$ What is the geometry? Can you put a 4' coil around the receiver? Could you put the receiver centered between 2 plates that are 4' apart? \$\endgroup\$ – bigjosh Jan 12 '16 at 1:39
  • \$\begingroup\$ The controller of the boards that receives the data could fit that coil, I think. The actual boards themselves need to be as small as possible - about 1x2 in. \$\endgroup\$ – Alex Wulff Jan 12 '16 at 1:42
  • \$\begingroup\$ Does the board get any sunshine? A small solar cell could charge a supercap or small NiMH cell. \$\endgroup\$ – Brian Drummond Jan 12 '16 at 10:33
  1. Some deep-space vehicles use radio-active thermal radiation + thermo-couples to produce electricity. Obviously, you can't use a radio-active source, but you might want to research how much energy could be generated using body heat + ambient heat for your system. If there is sufficient energy, then you may be able to power your tiny system

  2. Another possibility is harvesting kinetic energy generated by walking--however, that would require more hardware than simply using a high-energy cell in your system--so unless weight & cost aren't majors factors, then that idea probably won't be useful. Nevertheless, check this out. Here is another reference that may be useful to you.

Note: I imagine your best choice would probably be a rechargeable li-ion cell though, with respect to overall cost & weight factors.

  • \$\begingroup\$ The only thing I've found that has a reasonable size/price that can generate anything from heat is a Peltier device, however that'll work for about 10 seconds until the temperature gradient is gone. Have you heard of anything else? Thanks for the info about a LiOn cell - I haven't really researched whether or not a super capacitor would fit my size constraints (about 1x2 in) \$\endgroup\$ – Alex Wulff Jan 12 '16 at 1:16
  • \$\begingroup\$ Yes. This: powerchips.gi/index.shtml is a remarkable device (if it actually works). The manufacturer claims it works, however they won't release any samples unless you're prepared to invest in their plans. That device (or something like it) might be useful with ice as the cold side & body or ambient heat as the other side. hth \$\endgroup\$ – DIYser Jan 12 '16 at 5:23
  • \$\begingroup\$ Thanks @DIYser, I'll look into that chip and see if I can get one \$\endgroup\$ – Alex Wulff Jan 12 '16 at 15:51

Enocean have been making energy-harvesting radio light switches for several years. That converts the pressure of switching a light on or off to power a radio transmitter. It allows light switches to be put anywhere without the need for wiring.

Seiko have been making 'self-winding' 'Seiko Kinetic' electrically-powered watches for many years. They claim "1986 [...] was the first watch in the world to convert kinetic movement into electrical energy". It might be practical to do some experiments with that mechanism.

Pavegen make a flooring that can harvest energy of footsteps. Maybe similar technology could be put into shoes to harvest energy for the wearer.

I searched for self powered shoes and energy harvesting shoes, and get several interesting links.

Also, I have seen watches with solar cells, which can generate enough power from the ambient light they receive to keep them running.

Combined with a super-capacitor, or long-life rechargeable battery, should power something for many years.

As Seiko watches are profitable, desirable, products, it is feasible that another company has developed a variant of similar technology. If your Chinese is good, or you have access to someone who could help, you might be able to track down a third-party product. Failing that, try to get one cheap second-hand, and try to make a prototype. If your product is interesting enough Seiko might be interested too.

IIRC, the EnOcean micro-generator was developed by a third-party company, I think, German. They might have other products.

Lots of companies are working on energy-harvesting shoes.

Small side note: I think there are microcontroller familiess which are lower energy than ATtiny. So you might want to look at other MCU's, especially ones combined with radio, which may reduce the power requirements.

  • \$\begingroup\$ All very interesting links, thanks. Ill check them out. If only Seiko would provide their designs... \$\endgroup\$ – Alex Wulff Jan 12 '16 at 3:08
  • \$\begingroup\$ @AlexWulff - As Seiko watches are high-profit, desirable, products, it is feasible that another company has developed a variant of similar technology. If your Chinese is good, or you have access to someone who could help, you might be able to track down a third-party product. Failing that, try to get one cheap second-hand. If your product is interesting enough Seiko might be interested too. Also IIRC, the EnOcean micro-generator was developed by a third-party company, I think, German. They might have other products. \$\endgroup\$ – gbulmer Jan 12 '16 at 3:14
  • \$\begingroup\$ Wow, never seen those watches before! Very cool! More info here. \$\endgroup\$ – bigjosh Jan 12 '16 at 16:16

You mentioned a "wearer", so assuming this powered board will be mounted to a living, moving person?

If so, you could use an eccentric weight like those used in self winding watches to generate a small amount of power.

enter image description here

The idea would be to have a magnet that rotated past or though a coil. Each pass would generate a current in the coil, which you would use to charge an on board capacitor.

The power levels generated would be extremely low, so one of the many fine energy harvesting chips would be helpful.

  • \$\begingroup\$ Now this is truly an interesting concept. I bet there are some digital ones out there that do this too. I’ll research this for a bit and see what I come up with. \$\endgroup\$ – Alex Wulff Jan 12 '16 at 2:17
  • \$\begingroup\$ Depending on application and type of motion, it might also be possible to use a much simpler linear resonant mass - basically a magnetic BB inside a soda straw with wire wrapped around it. \$\endgroup\$ – bigjosh Jan 12 '16 at 2:19
  • \$\begingroup\$ Could you give a rough estimate of how much voltage that would put out? I assume that it would be minuscule. \$\endgroup\$ – Alex Wulff Jan 12 '16 at 2:21
  • \$\begingroup\$ Depends on frequency and amplitude of motion. \$\endgroup\$ – bigjosh Jan 12 '16 at 2:22
  • \$\begingroup\$ Another example... kjmagnetics.com/blog.asp?p=shake-flashlight \$\endgroup\$ – bigjosh Jan 12 '16 at 2:22

3 high quality AAA batteries in series directly connected to Vcc should be able to power an ATTINY85 running at 10Mhz with an average draw of 5uA for at least a decade of run time. It is probably a good idea to add a large capacitor in parallel with the battery pack to smooth out the high current periods since the batteries will last longer with a smooth, low current draw.

enter image description here

If you use the low voltage ATTINY85V part running at 4Mhz, then you could even get away with 2 AAA batteries and stay within the chip's operating voltage range.

An AAA battery typically has a capacity of >1000mA hours if run down to to less than 1 volt slowly. Newer high quality cells can physically last 10+ years before they start to fall apart...


Normal ATTINY85 @ 10Mhz has a voltage range of 3V-5.5V.

Low voltage ATTINY85V @ 4Mhz has a voltage rangge of 1.8V-5.5V.

enter image description here


You could meet your 1"x2" size requirement with 2 N size batteries size-by-side and still have at least a decade run time if you use the ATTINY85V part @ 4Mhz.

  • \$\begingroup\$ Actually, I’ve got it running at 1MHz, so I could go even lower! N batteries I think may be my last resort if I can’t come up with anything else. \$\endgroup\$ – Alex Wulff Jan 12 '16 at 2:14
  • \$\begingroup\$ You need at least 1.8 volts no matter how slow you go, so might be better to run at 4Mhz so you can get your job done as quickly as possible and go back to sleep since wake time will dominate your power usage, but all depends on actual application. \$\endgroup\$ – bigjosh Jan 12 '16 at 2:18

I can't comment yet, but got a question: for how long would you plan to have the device running?

In terms of battery I had excellent results with the 3V Lithium CR123A (not rechargeable) or the CR2540, which offers more than double of CR2032. They have 10 years shelf life and very low self-discharge.

Depending on the life expectation for your product, you might find better to combine battery with energy harvesting, which you could combine both using reasonable capacitors.

Another point is how often do you send data over RF? Is that possible to buffer the data to reduce the radio usage? Maybe have you radio to be powered by a capacitor charged by the harvesting, so you can keep collecting data using the battery and only transmit it when there's enough charge on the capacitor.

Also, depending where your device will be used the Piezoelectric Energy Harvesting might be another option.

Good luck!

  • \$\begingroup\$ Thanks, that cr123a looks ideal to replace the coin cell I’m cutely currently using. 1.5 mAH is a lot. Piezoelectric might not work for my application unfortunately. \$\endgroup\$ – Alex Wulff Jan 12 '16 at 12:22

You can get a Lithium rechargeable cell, the same size as a cr2032 or any primary button or coin cell. The lir2032 is the same size but a higher full charge voltage. These can be trickle charged, or you can use a proper charging circuit. These are used in those cheap solar powered flashlight Keychain.

As for charging, a solar cell would work well. You wouldn't need much light for your purpose.

Otherwise you could use rf coils for power. While in use might be difficult, so a dedicated charger might be quicker.

A direct charge is simplest though. So it would need a charging port, which can be as minimal as two pads on the outside.

  • \$\begingroup\$ Rechargable batteries would be ideal, but unfortunately once deployed these boards cannot be accessed to direct charge. Solar power would also be good but these boards also don’t have access to light. I think that the lir battery you mentioned would be a perfect one to store accumulated energy in however. \$\endgroup\$ – Alex Wulff Jan 12 '16 at 2:12
  • 1
    \$\begingroup\$ @alex the keywords you should be looking for on Google are ulp (ultra low power) and energy harvesting. Tons of white papers. \$\endgroup\$ – Passerby Jan 12 '16 at 2:52

How about solar? Well why not ? Very tiny input, but it sounds like your device will spend far more time idle/charging than running. Trickle a few fractions of a mA over many hours into a cap?

  • \$\begingroup\$ Unfortunately solar is out of the question for this one :( \$\endgroup\$ – Alex Wulff Jan 12 '16 at 1:33
  • \$\begingroup\$ Just a thought. Sorry! \$\endgroup\$ – adengle Jan 12 '16 at 1:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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