Energy Harvesting Applications: The New Frontier

Among the many new applications where integrated analog devices are showing up, we find energy harvesting applications in the forefront. For these applications, sensors are generally involved and power supply (or power supply support) circuitry is needed. So combining all onto an IC makes perfect — and probably elegant — sense.

One of the recent papers published as part of the 2013 IEEE International Solid-State Circuits Conference (ISSCC), is “A 1μW-to-1mW Energy-Aware Interface IC for Piezoelectric Harvesting with 40nA Quiescent Current and Zero-Bias Active Rectifiers.” The authors are Chris van Liempd, Stefano Stanzione, Younis Allasasmeh, and Chris van Hoof.

What these folks have suggested is to use an active bridge rectifier made up of small insulated gate N- and P-channel FETs. Since the end usage is for smart sensors, putting the sensor interface circuitry on the same IC as the FETs is the next logical step.

Why do you need active diodes? Because standard PN diodes will have an excessive voltage drop with even low forward current of several hundred millivolts. That ruins whatever possibility you had hoped for regarding high efficiency. Also, in some cases, your voltage source may be on the same order of magnitude as that diode-drop.

You would use the active bridge rectifier with a very low-power AC source such as a piezo-bender used to harvest energy. Here's a simplified model for piezo device and the active bridge power supply:

Equivalent circuit of the piezo-bender and the active bridge power supply. This and the following images are from the ISSCC paper cited above.

Equivalent circuit of the piezo-bender and the active bridge power supply.
This and the following images are from the ISSCC paper cited above.

As the title of the ISSCC paper makes clear, this is for extremely low power levels — 1μW to 1mW. This is the order of magnitude for power levels you could expect from impact sensors used in freight shipments or in human-powered devices. In this case, “human-powered” means low levels of movement from your daily life or (for implanted devices) muscle movement. At these power levels, the power supply circuitry must be very efficient. That's why you need diodes with such a low forward voltage drop.

The equivalent circuit of diodes in the bridge

The equivalent circuit of diodes in the bridge

The authors claim voltage efficiency of 94 percent at a power draw of 1μW and 99 percent at 1mW. They further claim a forward voltage drop of around 20mV.

If you fabricate the “diodes” from MOSFETs on an IC, the next thing to put on that IC is the rest of the power supply circuitry, the microcontroller unit (MCU), the radio (e.g., Bluetooth), and any sensor conditioning circuitry for whatever the end product is. That would look something like this, without the sensor circuitry shown:

Sensor circuitry would typically consist of a few op-amps or instrumentation amplifiers connected to the ADC inputs of the MCU. Again, this can be included on the IC — more integrated analog functionality.

This gives you an idea of the chip size in which to squeeze the amount of circuitry shown. Have you worked on any energy harvesting projects? Do you have any suggestions for such projects or products?

11 comments on “Energy Harvesting Applications: The New Frontier

  1. jkvasan
    June 30, 2013


    I didn't have an opportunity to work on energy harvesting yet am fascinated about the whole technology. In years to come, this can change the way remote sensing is done and probably healthcare delivery.

    Clever idea – using those small IG FETs  for the bridge rectifying function. Avoids getting limited by forward voltage of standard diodes.

  2. David Maciel Silva
    June 30, 2013


    I have not had chance too, by which I analyzed, this is a complete integrated control. These new devices are ideal for application rectifiers due to direct bias voltage to be smaller.

    We can discuss the best options for capture and control of energy.

    What do you think?

  3. SunitaT
    June 30, 2013

    @Brad, thanks for the post. I haven't had the opportunity to work on energy harvesting applications. Interesting to note that active diodes are used to avoid  excessive voltage drop with even low forward current of several hundred millivolts. 

  4. Brad Albing
    June 30, 2013

    @JK – That's what I think too – energy harvesting for everything from electronics that we wear to electronics that's used in industrial/transportation applications. From a few of my other blogs (past and future) we can see that this is an area to get involved in – that's where the money is/will be.

  5. Brad Albing
    June 30, 2013

    Yep – even a few 100 mV is too much.

  6. jkvasan
    July 1, 2013


    I would not be surprised if children's toys could use energy harvesting. Though it is a wild thought, I could not help thinking this technology being used in  street lighting.

  7. eafpres
    July 1, 2013

    In my work on sensors and sensor networks, there is constant discussion and innovation around energy harvesting for remote wireless sensors.  This makes good sense–even a sensor node that can run for a year on a coin cell becomes unmanegable when there are thousands of them and a year later they are randomly dieing.  A lot of good progress has been made in lowering the power requirements for sensors, and especially lowering the power requirement for wireless communication.

    But when you move from sensing into applications, like generating enough power to charge your phone, or a thing I saw recently to collect energy from sidewalks and store it to run street lamps, I start to think this is over the line.  In those applications I think harvesting becomes generation and storage, and things like solar panels or thermoelectric generators make more sense.

  8. Brad Albing
    July 1, 2013

    Energy harvesting for sensor apps – such as the railraod apps I discussed in my Squeaky Wheel blog – makes perfect sense. But you're right about streetlights – best to use solar panels, not sidewalks. What was the premise? Harvest energy from the thumping of foot-falls; or from the heat from the sun beating down?

  9. eafpres
    July 1, 2013

    Hi Brad–the energy is “harvested” from footsteps.  There are many applications proposed of this sort.  As a Chemical Engineer, I was taught to do an energy balance around the system.  The issue I have with this particular idea is that making the sidewalk move a little bit costs the human energy.  That energy has to come from food.  So the people who walk on said sidewalk have to eat a tiny bit more to make up for the tiny bit of energy produced.

    Unfortuneately, people are LESS efficient producers of energy than, say a coal fired power plant.  Roughly, not accounting for the carbon footprint of mining coal or the carbon footprint of producing food, people generate 25% more carbon per kWh than coal fired plants.  In general, producing food generates more carbon than mining coal, especially as the animal content goes up.  So the real differential is probably more than 25%.

    By the way, the article I read was in none other than Design News:

    Smart-City Technology Harvests Energy From Footsteps

  10. Brad Albing
    July 1, 2013

    Uh-oh. Sorry….

  11. RedDerek
    July 3, 2013

    One engery harvesting that has been around commercially are shoes that light up. But droping down to the 1uW range is interesting. Looking at the diode equivalent in the article, I see this being a good application for the Low Gate Threshold FET that Brad talked about

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.