Archived Forum

As you suggest, your answer will be ring-core c.t's. The problem you will face is the emonTx requires a significantly larger voltage than most c.t's are designed for. Therefore you need to derate your c.t. to keep within its power rating. As your circuits are 16 A, and most small c.t's are designed for a 0.333 V rms output and the emonTx requires approx 1 V rms, you should select a 50 A ( = 1 / 0.333 x 16) ring core c.t. without an internal burden resistor (i.e. "current output"), which should come significantly smaller than the standard 100 A split-core one. You'll also need to recalculate and change the burden resistor in the emonTx.

As an alternative, I think some constructors have used a much more powerful processor from the Arduino range with the required number of analogue inputs, and constructed their own front-end interface based on the emonTx circuit.

This type of CT works well, also from YHDC.  It fits nicely over one of the live cores of a 2.5 T&E cable.

The output signal is very similar to the split-core version.

Thank you both very much for the answer.

The CT that @calypso suggest looks very promising (in terms of size and price for now). Are you using these CT's?

I'll have a closer look when I'm at home. I guess I also should order a emonTx so I can start playing :-)

That one will give you more than enough voltage, so it looks OK to 20 A. It requires a 120 Ω burden so you will need to remove the on-board burden resistor on inputs 1-3 and substitute a 120 Ω (possibly wire-ended - the holes are there) one. Input 4 already has a 120 Ω burden so that's OK. That will give you a maximum current of 18.3 A on all four inputs. The next larger preferred resistor is too big, your maximum current would be 14.6 A.

Thanks very much Robert!

"The next larger preferred resistor" Is this to gain a greater accuracy? Will it be a big difference?

I'm asking because it is most unlikely I will get past 14+ amps on a group.

The CT that @calypso suggest looks very promising (in terms of size and price for now). Are you using these CT's?

Yes, one of these CTs tells me how much surplus energy has been diverted each day.  Not much at this time of year, but that's not it's fault!

The "next largest preferred resistor" (probably 150R) would provide more sensitivity (in ADC-steps per Amp) but over a reduced range.  Hence your ADC may reach its limit at less than the maximum current which you want to measure.  As long as most of the ADC's range is being used for the largest current you expect, that's as good as you can get with a simple linear scheme. 

The recommended maximum output swing of the CT is generally less than the ADC's input range.  Stressing the CT by using too high a burden resistor reduces the quality of its output signal.  It's probably best to use a burden that's slightly on the low side, and accept that not all of the ADC's range will be used.  With the 5V Arduino, my Mk2 system is scaled to reach the ADC's maximum at around 5 kW.  When using a system which operates at 3.3V, such as the emonTx, the match between the ADC and the CT is correspondingly better.

Basically, yes. But the engineering decision I'd go for is 4 inputs the same, as the difference will be marginal.

Thanks to both of you for helping! I kind of have an idea how to proceed.