Computing stuff tied to the physical world

The first MPS circuit

Now that we have a predictable power load and some sample code to test operation, it is time to try out a first harvesting circuit. All powered from just the Current Transformer:

Screen Shot 2015 02 24 at 10 10 54

The two red LEDs in series are a way to limit the voltage levels from the CT to a safe 3.4..3.6V, i.e. something that won’t fry the LPC810 µC.

And indeed, when we hook it all up and flip the switch it will actually work!

We can tell, because the LEDs will light up, indicating that there is some power coming in. The tricky bit is that the µC code is not doing anything we can detect. Or so it would seem.

The sample code was constructed in such a way that it spends most of its time in deep power-down mode, consuming less than 1 µA. But there is a small tell-tale sign: once every 3 seconds, the two LEDs will dim slightly – once very briefly, and once slightly longer.

And with the 10 Ω resistor in the ground return lead of the µC, we can actually see the current consumption of that quick power-up-and-go-to-sleep with an oscilloscope, as well as the dip in power supply voltage (Vdd) – due to the brief increased load:

SCR58 4

Celebration time: that little demo is running off energy harvested out of thin air!

The voltage drop may seem insignificant, but that’s because the µC goes back to sleep within a mere 250 µS. Even a constant 2 mA load (e.g. keeping the µC on) would make the power supply voltage gradually collapse. There simply is hardly any energy for us to use!

But there’s a far bigger problem hiding in there…

[Back to article index]