Computing stuff tied to the physical world

Setting up a test power source

To test the Micro Power Snitch, we need a consistent “test input”, i.e. a fixed load current with which to feed the circuit, as we try out different components and code logic strategies.

Ironically, though the CT output is “micro power”, the input side most definitely isn’t!

Let’s not aim too high at first, we can always optimise and improve later. Therefore, let’s use an initial load of say 110W, or about 0.5A @ 230V.

We could of course simply hook up a few lamps to AC mains, and clip the CT over one of the power wires. But there are a few drawbacks:

  • we’ll need to keep those lamps on, generating lots of heat all the time
  • plus, this means we’d be dealing with AC mains, even though it’s isolated

Fortunately, there’s a better solution, since all we care about is the AC current through the wire: we can use a normal step-down transformer to lower the voltage and then use a few resistors for the desired current load. This creates a low-voltage wire for the CT to clip on.

The lower the voltage the better, in fact. And given that ferrite-based toroidial transformers are slightly better at reproducing the AC mains waveform, we could use one of these:

DSC 4963

This particular unit is 93 x 23 mm and is rated at 2x 6 VAC @ 4.17A, i.e. 50 Watt of power. That’s over 8A when placed in parallel, which would correspond to a hefty 1.9 kW load at 230 VAC. Note that in this limiting case we’ll need to deal with “only” 50W of heat.

As it turns out, when measured, the output of this unit was closer to 2x 7 VAC (RMS).

The one remaining issue is how to set up a convenient load, i.e. how to dissipate all those up-to-50W (of heat!). There are lab instruments called programmable loads, but they are fairly expensive, especially for alternating currents (a diode bridge would introduce non-linearities). All we need really, are a few resistors and a heat sink – such as this concoction:

DSC 4961

Three fat 5Ω resistors, bolted to an old PC CPU cooling fin. Each resistor is rated at 50W. A nice benefit of this setup is that cooling is 100% passive – no fans, extra circuitry, or noise!

Let’s do the math: 5Ω @ 7V will draw 1.4A. Some combinations we can set up with this:

15   Ω @  7V => 0.47A ≈ 110W on 230V
10   Ω @  7V => 0.70A ≈ 160W on 230V
15   Ω @ 14V => 0.93A ≈ 215W on 230V
 5   Ω @  7V => 1.4A  ≈ 320W on 230V
 2.5 Ω @  7V => 2.8A  ≈ 640W on 230V
 1.67Ω @  7V => 4.2A  ≈ 970W on 230V

Higher currents require paralleling up the two secondary windings. This requires some care: the windings must be in phase, and the number of turns on each secondary must be identical. Toroids and other high-quality transformers will normally be ok.

A “DPDT” (Double-Pole Double-Throw) switch can be added to switch between 6V @ 8A and 12V @ 4A from the transformer. Plus “OFF”, if it has a centre-open middle position:

Screen Shot 2015 02 23 at 17 48 59

Note that the switch must handle heavy duty AC currents (over 4A), albeit at low voltages.

With additional resistors the current can also be lowered further. This is quite easy to do, since the extra resistors won’t need to be as high-power. For example: an additional 15Ω in series @ 7V = 0.24A (54W on 230V), and that extra 15Ω only needs to dissipate 2W of heat.

The result of all this is a flexible setup which can safely be used for checking out the operation of the Micro Power Snitch at various power levels. It only gets hand-warm.

[Back to article index]