Archived Forum

I'm not a meter expert, but from what I've been able to glean over time:

A Ferraris 3-phase meter has 3 discs mounted on a common shaft and 3 pairs of voltage and current coils. The torques from the 3 phases add mechanically and so record the arithmetic sum of energy across the 3 phases. Neither the angle between phases, nor phase rotation, counts.

I would expect a 3-phase digital meter to be designed to closely emulate this. However, I believe some are sensitive to phase rotation, but other than flagging it as an error, what this actually means in terms of how the operation is affected I don't know. I can't see that the meter should object to the two phases being 180° apart - if you had a really, really unbalanced load you could get close to that condition in real life.

How have you wired the meter? Presumably:

So on balance, what I think you are seeing is probably correct, and it's an important contribution to our knowledge of digital meters. 

Thanks for the reply Robert. Yes, I wired it up exactly as in your diagram.

The meter has a reverse energy register, which I can read, but in this test I didn't put enough power through it to trip the anti-creep (I momentarily put a kettle on to trigger the forward anti-creep). I'll have another go later with a bigger load.

A bit more info on this...

I replaced the forward load with a kettle controlled by a triac/emonTx and put a CT across both cables, so the normal diverter test but with forward and reverse current on different phases.

I left the system running in balance for more than an hour and nothing registered on the meter display (which is reading to 3 decimal places i.e. 1W/hour)

I then tried turning the kettle off, letting the energy bucket fill and then turning it back on, so "importing" a burst of 1800 Joules. I did this lots of times and still nothing registered on the display.

I then increased the energy bucket in the sketch so that more than 3600 Joules was imported at each kettle switch on and this time the display incremented by 0.001 each time.

So it seems that the 3 phase meter works much like a single phase one, with one energy bucket of about 3600 Joules shared by the 3 phases.

This means it should be possible to fill the bucket from one phase and empty it into a different one, but I think it also means that if you're importing on one phase and exporting on another then it may not be a good idea to divert the exported energy.

I've also concluded that a 3-phase meter makes a useful test bed for a single-phase system :)

Kudos to you, sir! I think your conclusions are spot on. The 'energy controller' needs to be extended across all 3 phases!

(I think it was always understood that a 3-phase meter would work properly on a single phase, but it's nice to know that the energy bucket is the same size at 1 Wh).

Just a quick update on this...

I'm now using a single energy bucket across the 3 phases and in today's sunshine with the immersion heaters fully on most of the time I'm seeing import on 2 phases and export on the 3rd phase (that doesn't have an immersion connected). The total load though, as shown on emonGLCD, is sitting nicely around 0 +/- a few watts.

I checked the house meter and the 1Wh LED isn't flashing at all, so I'm pretty convinced that this theory is valid.

Robert, I don't think I quite got my point over when I said "a 3-phase meter makes a useful test bed for a single-phase system". What I meant was that a 3-phase meter is better than a single phase one for testing a single phase system because with it wired as in my test you can verify that your system is not importing or exporting even without any solar power present.