Archived Forum

Hi, I'm no expert with 3-phase, but someone will no doubt correct me if get this wrong ...

A single emonTx can only measure voltage on one phase.  It can, however, measure current on 3 phases, by using 3 CTs.  To calculate the real power on three phases, the sketch would need to extrapolate the voltage signals for the two phases for which voltage is not being measured.  This process will introduce some inaccuracy, but may be OK for your needs.

With 6 analogue inputs being available, your best approach may be to use three for voltage and three for current.  A PCB is available from another forum member for this purpose.  Having populated this board as shown, you would just need to provide 3 CTs and an Arduino or similar.  I presume that a suitable sketch will be available from the same source.  Otherwise, the 2-phase code that I posted here on 17/10/2013 (for a New Zealand project) could be easily adapted for this purpose.

The above PCB appears to have a single 3-phase input connector (L1, L2, L3 & N).  If you can supply this combined feed from a single source, then the only other power outlet you'd need would be for the processor's DC power supply.

For accurate calculation of real power, phase-shift effects should be taken into account.  My 'New Zealand' code, as above, includes this feature as I noted at the time of posting:

This code has been optimised for monitoring real power on several channels.  PowerCal and phaseCal can be individually set for each CT.  I think the Uno platform is well suited for this particular project, but similar code could be run on an emonTx with additional front-end circuitry as required.  I hope that it will be of use to anyone who has need of such a solution.  

It's interesting how many queries there have been about 3 phase monitoring recently. It's almost unheard of for a domestic property to get 3 phase in the UK, so the emonTX hardware reflects this.

However, in Europe, especially in Scandinavian countries where electric heating is the norm, 3 phase is common. There's an important distinction to be made between domestic and commercial 3-phase usage. In commercial applications, the size of the loads means you can more easily balance the load between the phases, and a significant proportion of load is often made up from 3-phase motors (used for heating, cooling, ventilation, pumping). Domestic properties aren't like this, the phases can be really unbalanced.

For that reason, monitoring the voltage on all three phases can become much more important!

"For that reason, monitoring the voltage on all three phases can become much more important!" Indeed. And don't forget that if the neutral voltage is significant, the angles between the phases will shift too - and it would be almost impossible to compensate for that dynamically. It all gets very interesting.

An alternative to Robin's suggestion: MartinR's "Full Fat 3-phase monitor". It uses 3 stacked emonTx's - Version 2 (V3s could be used but the means of interconnecting them is not so straightforward). That will give you a lot more processing power too.

[I had never come across 3-phase for a domestic supply in the UK before coming to this site, and as far as I know we have two: Noah and MartinR. Noah's is a converted industrial premises, which explains it in his case. I don't know about Martin's.]

To hopefully clarify Robin's answer a little: "To calculate the real power on three phases, the sketch would need to extrapolate the voltage signals for the two phases for which voltage is not being measured.  This process will introduce some inaccuracy, but may be OK for your needs."  That's not totally accurate. The published 3-phase sketch measures a single voltage and records it, then "plays it back" 120° and 240° later to use it to calculate the real power etc for the 2nd and 3rd phases following on behind. The underlying assumption that leads to a possible inaccuracy is that the 2nd and 3rd phases are exactly the same shape and amplitude as the 1st phase. Hopefully, if all the loads and supply impedances etc are balanced, that will be true. In practice, it won't be, and how far out they will be will depend on your individual circumstances.

Robert,

I'm also going 3 phase. I live on a farm, but don't do much farming and currently have single phase. I have a little bit of 3 phase stuff, compressor, welder and a motor but it was cheaper to just use a generator for that. Our local DNO is replacing the overhead grid around here so I took the opportunity of going 3 phase It still wasn't cheap and I have to have it installed in their maintenance schedule which seems to be stuck!!.

It has 2 main advantages-

1 I can use my 3 phase equipment without running a generator.

2 I can install more solar panels.

A friend (another Tinbum!) who also lives on a farm and doesn't farm has single phase on the same single-phase final transformer as his neighbour who does. Neighbour's milking machine has allegedly blown up a laptop, mains-borne network, a portable tv and another neighbour's central heating. I installed a large surge suppressor for him, so far no more trouble. Except thieves came a month ago in the night and stole his copper overhead feeder. He's on Ali now.

RW:  That's not totally accurate. That's not totally accurate. That's not totally accurate.

Sorry Robert, I thought I was making just the same point as you have done.  Maybe I should have said that the process of extrapolating, or "playing the voltage samples back later" may introduce some inaccuracy, rather than it necessarily will do.  As you've already mentioned, if all three phases are nicely in balance, both approaches should give the same results.

[Edit - But that process is not extrapolation, that is what I was saying. RW]

OK, Chief, so I used the wrong word.  Is there a right word for this operation?

When I first moved here 11 years ago I couldn't get my single phase welder to work. It went back under warranty twice but they said it was fine. Some time later I was checking one of my meter on the mains and noticed it read 263v. It turns out  that our 5kva !!! transformer was totally on it last legs. It had had quite a few lightening strikes as well they said. It was replaced within 3 hours by the  DNO.

They are now going to be putting in a 100 KVa (I'd asked for 50 KVA,the smallest they do). I've also got quite a few surge arresters that I shall be putting in - too many electronics. I'm also thinking of putting them on the DC from the PV panels as well.

We've had quite a few undesirables here as well but thankfully I've spotted them all!!

Power costs in Australia have soared over recent years.  They have gone up about 70% in 5 years, and are forecast to keep going up.  Given 3-phase is used in Australia (presumably for swimming pools and large air conditioners that are not so common in the UK) then there may be a lot more 3-phase questions coming your way.    The problem is that the EmonTx page says you can use an EmonTx for either 4 single phase lines or 1 3-phase, but it doesn't provide any information about 3-phase past that point.

I am no electrician, but I am interested in the Raspberry Pi and also my energy consumption.  I thought this project was a great way for me to learn about the RP and also meet my energy interests.  So the question now is "what is the best way forward for me?"  Cost is not the main factor, but I don't want to get too technical on the power board side.  I can buy one or 3 EmonTx devices, and I can clip the CTs where I am told, but not much past that. 

1.  Is there a way to measure apparent power for 3 phase?  Is it just a case of clipping 3 CTs (one per phase)?  If so, maybe that would be a good way to start.  I could set it up and compare my findings with my power bills and see what happens.

2.  Would a retail unit like this http://steplight.com.au/monitor/ew4030-watts-clever-smart-energy-monitor/ work for 3-phase power as an alternative?

3.  If I go the whole hog, I assume I will need to do the following (please confirm).

• have an electrician install 3 power outlets on my power board (one connected to each phase
• install 3 x EmonTx, one for each phase.  Attach a single CT to each phase line and plug the AC plug into the associated outlet.

1. If your system voltage is reasonably constant, you can get an indication of apparent power by just measuring the three currents and supplying a fixed voltage (this is how the simplest commercial units operate). Over time and with careful adjustment of the voltage, this can tally reasonably well with your meter. A better way is to measure the voltage of one phase and the currents of three phases. With the "three-phase" sketch, this will give you a good measure of real power in one phase and a good estimate of real power in the other two. Finally, the best way is to actually measure the three voltages and the three currents, which will give you the best accuracy and both real power and apparent power.

2. From a cursory glance at the web page and manual, that unit in essence conveys the reading of your meter into your living room. So it will give you the same information that the flashing LED on your meter gives you - the total energy (from the number of flashes) and power (from the flash rate). The advantage is that by definition it is accurate (unless it misses a pulse), the disadvantage is the information available to you may lack detail - for example you won't know the energy in each phase, and at best it will increment in units of one pulse of your meter - usually but not necessarily 1 Wh.

3. That is certainly one way to go.

Also living in Australia... I had three phases installed in the previous house; 18kW commercial cook top, pool pumps, solar hot water pump, etc.
I moved to the country side a few moths ago... same thing... 3-phase power.

Before living in AU, I lived in Germany... had three phase power... basically all my lif (in numerous places).

My point, it may not be common in the UK, but certainly elsewhere. My critique for OEM was the design is UK focussed and not world focussed, which is what I would expect from an 'Open' system, which implicitly addresses the world. :)

Hi MallyCat, I would be rather looking at this set-up www.ekmmetering.com/ekm-metering-products/electric-meters-kwh-meters.html (rather then Steplight) -- I am currently thinking as to whether I build an OEM with 3 stacked emonTX v2 or get this one. You might consider DIN rail power points in case you are strapped for space in your meterbox.

Yes, power prices will continue to go up given the continued drive for privatisation of public utilities and an energy policy which is ludicrous at best. Which has me also researching the use of a battery bank during night time, rather than exclusive 'dumping' energy to a heater. Thought I understand why it is done in the UK.

MaxG: My critique for OEM was the design is UK focussed and not world focussed, which is what I would expect from an 'Open' system, which implicitly addresses the world. :)

Sure, but the world is a big place, and you have to start somewhere!  Its good to see so many contributions now appearing from further afield.  The 'Open' in the title is presumably intended to reflect the open-source nature of the enterprise. 

"World focussed" sounds like something out of Star Wars :D

An obvious expansion option would be to have "single phase" and "three phase" emonTX designs. One would have three AC inputs, three CT inputs, and however much else there is IO for on a 328, and the other would be the current v3 except it has 1 100A input and 3 30A inputs, rather than the reverse. 

The other thing to keep in mind with that pulse-counting solution above is that if you have PV,  the meter pulse will only fire when you're importing.  I have one of those EM1000 meters shown in the photos in the link above.  It typically stops pulsing once the sun comes up and doesn't start again until the sun goes down.

I'm on a very generous feed-in tariff, so monitoring my day time consumption (sourced from the panels) is at least as important as monitoring night time consumption (sourced from the grid).