Accuracy of 2x CT Connected to Single Phase (2 hots and a neutral)


I have 2x CT’s installed in a single phase (US two hots, and a neutral), and ran into my emon PI.

Today they are running at 116 volts, and 117 volts. Some days I may get to 110 on one, and 120 on another, etc.

I have the 9V EmonPI transformer connected to one of the hot legs, but that is it.

If one phase gets very low, or very high, but the transformer is connected to the other phase, what kind of inaccuracy are we talking about here?

What kind of inaccuracy overall if my voltage does not match what the Emon pi is configured for percentage wise? (Calibrate voltage on emonPi)

This is the amperage on a normal day:

That totally depends on the value the a.c. adapter feeds compared to the voltage that the ‘other’ leg of the supply is seeing, when you have a 120 V adapter on one leg.

You need to do the maths. P = V × I. V is measured for one leg and guessed (or assumed the same) for the other, while I is measured separately for the two legs. Therefore if the voltage is 5% higher on the leg that’s not measured compared to the voltage on the one that is measured, that is your error - the actual power will be 5% higher than indicated. This of course assumes you’ve calibrated the a.c. voltage to read correctly the voltage that it’s measuring, and the powers too, remembering that tweaking the voltage in emonhub.conf does not affect the power values, those need to be calibrated separately.

When I’ve written the software for it, an emonTx V4 should solve this problem, because you have 3 voltage inputs available in that with the emonVs.

So it appears that the emonpi is recording the vrms at 111.2 but the actual voltage is 120.2.

So this means that the emonPi is calculating the POWER (kwh) around 9.1 percent low right? So it is actually recording LESS power then is being used?

No, power is measured in kW, it’s energy that’s measured in kWh. But you would be correct in that both will be reported low.

You can “correct” the voltage reported by changing the “scales = …” value in emonHub by 9.1% upwards (i.e. multiply by 120.2 / 111.2 ≈ 1.08), and if the currents are correct, then the same factor for power1 & power2 (because the multiplication of voltage and current is one of the first processes, done a long time before the data arrives in emonHub).
If you suspect the currents are not the same and they probably won’t be exactly the same, put both c.t’s on the same leg and adjust one to match the other. Then, with it back in the right place and in the absence of anything better, over a period of time you can then do small and equal adjustments to both power1 & power2 scales so that their total energy agrees with your supplier’s meter.

It seems to me that you ought to look at split phase US AC with two seperate voltage probes as well as two current transformer probes to avoid the ±8% or so errors reported above. Then P1 = I1*V1 and P2 = I2 * V2 and add P1+P2 or upload them both.

Split phase is different to what I have here in the UK so I can’t test much for you.

You seem to have not noticed that the emonTx upto but not including the V4, and the emonPi, have only a single voltage input, so your suggestion is unworkable with only one device. Those of us who’ve been around here for a while are well aware of the problem. The emonTx V4 will enable that facility for the first time here on OEM with devices available from the OEM Shop & Megni.

What I had noticed was that the V4 is described as having triple independent I and V inputs for three phases, and what I advocated was to use two.

You didn’t mention the emonTx V4.