Computing stuff tied to the physical world

Voltage sources

A lot of the work at JeeLabs is about wireless communication. When the signalling doesn’t use wires, then the most practical option for power is to also provide it without wires.

There are a number of options:

  1. batteries, both single-use and rechargeable
  2. super-capacitors, i.e. very large capacitors
  3. solar energy, i.e. photovoltaic cells (yep, same Volta as before)
  4. energy harvesting, i.e. tapping some ambient source of energy

DSC 1949


There are many types of batteries. Let’s focus on the small ones first, since the aim here is to power small wireless nodes, spread out in an around the house:

  • AA and AAA: single use alkaline provides 1.5V, and rechargeable NiMh around 1.2V, so in both cases we’ll need two or three of them in series to produce 2.4..3.6V, roughly. These can be used either through a regulator with fairly little voltage drop / loss, or as is if the voltage range is known to be acceptable across the battery’s entire lifetime.
  • Coin cell: the most common one is the CR2032 (20 mm round, 3.2 mm thick), and delivers a nice 3.0V – the disadvantage is that most coin cells have limited power (200 mAh) and are single-use, leading to recurring replacement and waste disposal.
  • LiPo (Lithium Polymer): these rechargeable cells supply 3.7..4.2V and have a very high energy density – advantage: available in a wide variety of sizes/capacities (also as cell phone packs) – disadvantage: needs some care, since the internal resistance is very low (a short can produce high currents on “unprotected” cells).

When making battery choices, it’s also important to consider what voltage range all parts of the circuit will support. The LPC8xx series works from 1.8..3.6V and so will the RFM12/69 radio modules, for example. Many sensors work well in the 3.0..3.3V range, some support a larger range – as long as all components are used in their operating range, we’ll be fine.

More power

Sometimes, you may need much higher voltages, i.e. 12V to drive a motor or relay, and much higher currents of 1A or more. The hobby R/C model airplane/boat/car world has embraced “LiPo packs”, which look like this:

63379 m

This particular one is described as “4S1P”, which means 4 LiPo cells in series, and just one of each (none in parallel). Its nominal voltage is 14.8V, which is really the minimum value once the battery is nearly discharged – when fully charged up, it’ll be 4x 4.2V = 16.8V. The capacity is 1300 mAh, i.e. it can supply 1.3A for 1 hour, 0.65A for two hours, etc. before needing a full recharge. This is actually a relatively small unit in the hobby R/C world.

But here’s the thing to watch: the discharge rate of this particular unit is specified as “40C”, which means that this battery can deliver a current of 40 times its capacity – i.e. 52A! Although it would only be able to do so for 90 seconds, since even at that rate the capacity is still the same. But what this means is that this harmless-looking battery of just 77x34x31 mm can pump a whopping 16.8 x 52 = 870 Watt (≈ 1 horsepower!) through a circuit with low enough resistance. Enough to melt just about any piece of metal connected to it.

LiPo batteries are fantastic power sources, and can keep their charge for many years if only a tiny current is drawn from them. But unconstrained ones can also be very dangerous.

Note: the extra connectors shown above are typical for series-connected batteries, they are used to make sure the voltage is evenly spread across each one of them while charging up.

Super capacitors

These are – as the name indicates – capacitors with a very large capacity. They are not necessarily physically large: most “supercaps” only tolerate a maximum voltage of 2.7V or 5.5V. Here is a 0.47 Farad @ 5V unit, it’s 17×14 mm, i.e. comparable to a small battery:

PB SERIES 13 0H 16 8L

These units are great for storing a little energy, but they are still capacitors: when discharging, the voltage will drop exponentially, as with any capacitor. And their self-discharge due to internal leakage is still in the order of 100 µA usually, so they may not even last through the night, regardless how low-power the rest of the circuit is.

New developments have even led to ultra-capacitors, with capacities of up to 5000 Farad (compare that to the 0.1 microfarad capacitors used for digital circuit decoupling!):

ESHSP 5000C0 002R7

But that’s a 16x6x6 cm unit, costing around €200. Not really what we’re after, usually…

Solar energy

Solar energy is “in”. Lots of people have solar panels on the roof – free energy, as long as the sun is shining, right? Not so fast, there are some hurdles when used inside the house…

Screen shot 2010 03 05 at 23 1 03 57

The big disapointment you will run into when trying out solar cells indoors, is that the sun doesn’t really shine into the house all that strongly. Yes, we could collect some power right next to a south-facing window, but it’ll be a lot less than outside, and who wants a big panel on each sensor node anyway…

Expect to see 1000x less solar energy in many parts of the house inside, versus outside.

Also, you’ll need to carefully investigate what type of material is used in your solar setup. The mono-/polycrystalline types are less suited for indoor use than amorphouse ones.

Getting by on solar energy on a long dark week in winter, somewhere inside the house is still going to be a tough proposition. Hopefully better technologies will arrive one day.

Energy harvesting

This is an area where a lot of development is taking place. Wouldn’t it be great if you could pull some energy literally out of thin air? There are usually lots of electromagnetic fields around the house, so why not pick them up and convert them to a power supply trickle?

The chance of collecting enough energy and finding a source which is available often enough is still slim for now. Here are some sources you could look into:

  • radio waves from a nearby transmitter or a WiFi access point (think nanowatts)
  • magnetic pickup from AC mains (i.e. transformer), only works when power is drawn
  • a piezo crystal being pushed (as in lighters): delivers a very brief high-voltage peak
  • heat differences can be converted by a Peltier element (think: inverted thermocouple)
  • sound can be picked up by a microphone (but where do you get sound all the time?)
  • as a variation of this, you could consider vibration, such as from a fridge motor
  • esoteric, but not unthinkable: the human body (it generates up to 100 W, as heat)

And of course all sorts of ways to convert mechanical energy / motion into electricity:

  • spinning wheels, dropping weights, chains & pulleys, doors & windows
  • water flowing through a pipe (rainwater from the roof?)
  • wind energy, air motion from convection


For most intermittent sources, our best bet is probably to combine it with a rechargeable battery of some kind. That way, we can collect/harvest energy when it’s available, and save it up for hard times. The most likely candidates for such intermittent storage are probably the LiPo battery and the supercap.

The issue here is to plan for worst-case energy scenario’s: a week of cold & dark weather during the winter will strain even the best solar collection setup.

The alternative is to surrender to external variations and take a very different approach: power up when the energy is available, and start measuring, sending, etc. And then just die and wait for the next energy-rich period to power up again.

Although this strategy looks simple on paper, it can be surprisingly tricky to implement. Many electrical circuits have a very inconvenient startup current “ridge”: until fully powered up, so they can start doing clever things to save energy, many chips tend to draw quite a bit of current (relatively speaking). That current draw can easily prevent the power source from ever reaching the minimum voltage needed for proper operation. Catch 22!

Now what?

The conclusion of all this is a resounding “it depends”. The simplest power source is an external one, i.e. a battery with known voltage and capacity, which then gets replaced (or recharged) from time to time. None of the alternatives has the same level of predictability.

Battery life can be prolonged either by making the circuit more energy efficient, or by using some alternative source(s) to “top up” the available power.

For outdoor use, say a plant-watering monitor in the garden, solar w/ a small LiPo battery is probably quite a good option. Or if restarts are ok: just a solar cell + reservoir capacitor.

In each of these cases, you’ll need to determine what voltage levels are involved, and how to get things to match up with the circuit demands. A few NiMh’s or a LiPo, in combination with a linear regulator will often be fine. Switchers can manage energy more efficiently, but they also consume a small-but-permanent quiescent current which cannot be ignored.

In the world of electricity and power, there is no such thing as a free lunch.

[Back to article index]