All CT's are rated too high

I started filling my shopping cart to purchase and noticed that the smallest CT’s they sell are 50 amps!

Those would work great for my off-grid and solar feed breakers, but most of the loads in my home are on 15 amp breakers.

The problem is that, upon looking up the specs on these things, they lose accuracy below 10% of their rated current. So a 50 amp CT would be less accurate below 5 amps. That’s 500+ watts! Most of the loads in my home are in the 20 to 100 watt range, which is 5x to 25x below the accuracy threshold of the 50 amp CT’s Iotawatt sells.

Why don’t they sell the 20 amp units?

Is there a recommended alternative? I can use either split core or solid, makes no difference to me to pull a wire out of a breaker to slip a solid core CT on it.


While I’m familiar with this school of thought, IoTaWatt does not seem to suffer from this problem. The 50A CTs are very high quality and are rated at 0.5% accuracy.

While it’s probably true that other energy monitors get better results with limited range CTs, it should be pointed out that those monitors have 10 bit ADCs and sample rates of 50-60 samples per cycle where IoTaWatt has 12 bit ADCs and samples at about 640 / cycle (60Hz). So 4x ADC resolution and 10x sampling.

The 20A-25A CTs that are generally available do not have the power to drive modest burden resistors at fill rated capacity and begin to show distortion and increased phase shift below rated primary current with a 20 Ohm burden resistor. They have retrofit cables that are not UL Listed, and when all is said and done (shipping and 25% China tariffs) are not significantly cheaper.

Using a standard 50A AccuCT on my test bench, with a 40W incandescent light bulb, I get about 335mA draw with my Fluke meter.

The IoTaWatt averages 338 mA +/- 1 mA over 10 minutes.

That’s about 0.9% difference between the meter and IoTaWatt/AccuCT. But the Fluke meter accuracy on the 400mA scale is +/- 1.5% of reading, so there is no significant difference.

How low can you go?

Realizing that near perfect accuracy at 0.7% of scale may not be convincing, I dug out an incandescent nightlight.

Fluke 58.7 mA
IoTaWatt 58 mA

That’s at .1% of scale!


Ok, good enough for me.

Its not like I can’t swap them out at a later date right?

When we switch to off grid mode, the battery’s BMS will tell me if there are any serious inaccuracies…