First of all, it is interesting to see that that picture has been deleted. A few hours after my comment. I don't know why, but i don't want to speculate. I will look for a copy. Filename was R_123517565_2.jpg Perhaps Google finds something. If I find it, i will tell it here.
@Alan: you measure the tension or between two phases (not between two pairs of phases as you say, excuse me), or between the phases and neutral. An open input line (usually with high impedance in the megaohm range) with a bit of cable leads always to noisy signals in the mV or even V range. Test it youself with a normal electronic digital voltmeter with an unconnected cable (I mean unattached to the 240 V AC tension) attached to the input. Perhaps in the middle of the Gobi desert or on the moon you will not catch noise, but inside a house or laboratory you will see noise. I repeat: Rossi said clearly that the measurement were made "before" the control box. (see his blog JONP, i think it was yesterday) This rules out any strange phase shift between the AC tensions leading to false tension measurements made between to phases. Furthermore: if there was such an exotic phase shift between two phases, you would expect to see also exotic tensions between the other lines. Don't you agree with me? At this is not the case. So, the hypothesis of an open tension input is not confuted by your comment. Of course i do not question this instrument PCE830. But: with a "near zero" tension, the power calculated will result in a very small value, here only 39 Watt. A user asked me a question. This is my answer: There are clamp meters only for AC, they are in fact transformers with a primary coil of "on winding" or loop. DC currents cannot be detected with these AC clamp meters, heavy DC currents may disturb or falsify measurements by saturation of the core material (unlikely here). These AC clamp meters are usually cheap. Then there are DC clamp meters with hall effect sensors. They are usually a bit more expensive and less precise and may have an offset value to take into account. In many cases they would not show you a AC current. Some clamp meters can measure AC AND DC currents. In fact, a hall sensor can easily detect AC currents, if the sampling computer makes enough measurements per second to compute the true RMS value for a period longer than the frequency used.
