Archived Forum

I suggest the maximum number of primary turns possible, then adjust the value of the burden resistor as necessary. That way, the errors inside the CT should be minimised. However, a word of caution regarding the wire gauge: you will now have multiple bunched conductors and the wire needs to be derated. If the bunched length is small, and the rest of the coil is fanned out to allow convection cooling, then you may be OK. What I would do, apart from checking the numbers, is once in service, check the condition of the insulation regularly at the CT where it is likely to be warmest. If any degradation is apparent, you must replace the wire.

Thanks Robert, I'll update this thread once I've got some sensible results. As I've only just ordered the second EMONTX it will be a week or 2

I've decided to go with 8 turns of 2.5mm conduit wire for the 500w circuits and 4 turns for the 1kw circuits. It fits through the CT albeit a little snuggly with 8 turns. I will standardise on the high sensitivity burden resistance. A short test with 7 filament bulbs drawing 470'w suggests there isn't any noticeable heat generation with 1.5mm cable so 2.5mm should be even less loaded. I'll keep an eye on it once I've put it inside its plastic box.

When you know the maximum power that could be flowing in each circuit, you can arrange for the output from each CT (using any number of turns) to span most of the ADC's input range thereby giving optimum resolution. 

I would be most surprised if your measurements are significantly improved when using multiple turns around your CTs rather than just one.   I look forward to your results.

Multiple turns is a useful technique for simulating the effects of larger currents.  But in a properly scaled system, I have yet to see any benefit of measuring current/power in this way.