Archived Forum

I've been trying to work out what I need.

I think I need:

• 8 no. CT sensors - 3 to go on the line feeds from the solar PV inverter and one (somewhere else)? And similarly 3 to go on the main meter in feeds and again one more (somewhere else)?
• 3 no. emonTx V3 - Electricity Monitoring Transmitter Unit - 433MHz to collect the data, one for each phase.
• 3 no AC-AC Power Supply Adapter - AC voltage sensor (UK Plug) - one for each phase on grid in. Do I need another 3 for the solar pv inverter feeds?
• 1 no. RFM69Pi 433Mhz Raspberry Pi Base Station Receiver Board, which I can attach to my existing Raspberry Pi.  Does anyone know the maximum distance that can exist between the emonTx and the base station?
• 1 no. Programmer - USB to serial UART to program the emonTX's with three phase sketch.

If I'm going awry here, I'd be grateful for pointers. Many thanks.

The idea of one emonTx per phase is a good one as it allows the most accurate measurements. So per emonTx, for the main PV and grid measurements, you need as you say 2 CTs and one ac adapter. The voltage on the grid is the voltage on the grid, so you don't need and cannot use a separate ac adapter for the PV infeed. You can add a further 2 CTs per emonTx for loads on the same phase. Note that by changing a resistor, the emonTx input sensitivity can be customised. You must NOT have the three-phase sketch, as each emonTx is dealing with a single phase, but I recommend you keep the programmer as you're bound to need to adjust something sooner or later. Also note that the 3 emonTx's will operate independently, so will report to your Pi at different times - and the reporting intervals might also drift due to crystal tolerances. This shouldn't be a major problem but if two try to transmit at the same instant, both messages will be lost. There is no firm maximum distance over which the RFM69CW will operate as so much depends on obstacles in the path and local sources of interference. You might find some indications over on the JeeLabs website.

You might not need the emonTx's - if you have suitable Arduino boards, you could convert the suitable ones (don't ask me - I don't know the Arduino range, hopefully someone knowledgeable will chip in) by using the emonTx Shield instead. You should also look at MartinR's full-fat 3-phase monitor, which stacked the old emonTx V2 (no longer available) and used a wired connection between the three.

And of course read the Building-Blocks articles about diversion.

Hi rtX,  I have done exactly what you suggested, using 3 single phase EmonTX's as Robert mentioned.   The resulting hardware looks like the attached photo.

The dashboard we built looks like this:

http://data.peakmonitor.com/test01

(although it looks like the site lost internet connection on Saturday as the data has stopped then).

Works great!  10 second resolution on each phase.  If you zoom in on the graphs (click and drag to zoom) you can see the differences in power usage on each phase.  Also check out the voltage history on the "subpanels" page.

Very interesting to see this, but at the same time worrying because I am now looking at having to install monitoring on a few three-phase systems and certainly had not anticipated having to use anything more than an emon Pi and perhaps a single Tx.

Is there absolutely no way it can be done using a single box? I have ethernet sockets available so don't need wireless - although I guess that doesn't get me far given the Tx doesn't have ethernet either.

Chris.

I have developed a PCB for continuously monitoring real power at one point in a 3-phase system.  It uses the standard Atmel 328P processor, so programming is the same as for the emonTx.  If connected at the grid supply point, the associated code can control the diversion of any surplus energy to one or more dump-loads.  This board has an integral power supply so no DC or AC adapters are required; the phases are fed directly to three separate transformers on the board.  To measure the flow of energy at a second location, a second board would be required.  Each board can send its data by RF to an emonBase unit just as the emonTx does.  Temperature sensing is not yet supported for this application.

Full details at http://mk2pvrouter.co.uk/61101.html   and there's an email address for enquiries on my Shop page 

Hello  First post so please don't be unkind....

I can see what Chris is getting at .....With the EmonPi only having 2 CT inputs, the Emon_TX has 4, so....

EmonTX_1 (on Node 9) measuring current on the three phases with CT1, CT2 and CT3

The AC voltage reference will be between nutral and a single Phase (i.e. 240V)

The second Emon_tx (on node 10)will look at the three phase loads ...

So the hardware requirements will be 1 off Emon_pi, 2 off Emon_Tx and 6 off CT units ...

What am I missing?

Roj

What am I missing ...

Although each of your emonTx units has a CT for each phase, it can't measure the AC voltage on more than one phase.  So you will only be able to calculate apparent power (I * I) rather than real power (V * I).

[You mean apparent power = V_rms × I_rms, whereas real power = 1 / n × Σ (V_n×I_n)  - RW]

Roj, you're missing 2-off ac adapters to measure the alternating voltage, one for each emonTx.

Robin: With the delayed waveform in the 3-phase sketch, you can get a good approximation for real power in the 2nd & 3rd phases. It's better than apparent power but it depends on the phase voltages remaining in balance. Unless you can guarantee that, it's clearly never going to be as accurate as your dedicated 3-phase unit with the voltage of each phase accurately monitored and the real power accurately calculated.

Many thanks. Got it

Roj

OK Robert, I think we've got multiple ideas in this thread as to how to go about monitoring 3-phase power.

With 3 separate emonTx V3 boards, one per phase, accurate measurements of real power can be done at up to four locations on the associated phase.  This will allow calculations for real power (or anything else) to be transmitted for datalogging purposes every few seconds.

With a single emonTx at each of two locations, Robert's 3-phase sketch can provide a good approximation to the above by measuring voltage on just one of the phases, and using delayed versions for calculations on the other two phases.

Neither of the above approaches would be suitable for diverting surplus power in a 3-phase environment.  For that application, voltage and current measurements are needed for all three phases.

Due to the layout of the rather large switch rooms, I am unable to monitor incoming power on each phase - I can only measure PV generated power. In fact its not feasible or even sensible to measure any temperatures at these locations - "outside" is too far away and "inisde" can be absolutely roasting hot due to the presence of heating ducts and even solar themal installations (which I'm hoping to monitor at a later date). Additionally its going to be pushing things asking for more than one or maybe two 13A sockets for the bricks.

If I use an emon Tx and an emon Pi, is there any way I can avoid having to use radio transmission between the boxes?

Chris.

Yes, you can use a serial connection.

Yes, that's a concise summary.

Think this must be the page you're referring to regards a serial connection, Robert?

https://openenergymonitor.org/emon/node/3884

...but it doesn't actually show where I need to connect to on the emon Pi circuit board. I'd like to power both the Pi and Tx off the DC brick that will power the emon Pi. I see there are pads marked 5v and GND next to the RJ45 I/O connector - should solder onto those? And there are Tx and Rx data pins on the header marked 'ATmega328 UART' - would the Rx pin be correct for serial data from the emon Tx3.4?

Once I've got the connections made, what exactly do I need to do to the Pi's code to get it to read the data off the serial connection? Is this included in "Robert's 3-phase sketch" - which I assume I'll have to put on my Pi ?

The emonPi's add-on board is already using the only serial port on the Pi's GPIO. You would be far better off using a usb serial port (e.g. usb programmer). That makes connecting to the emonPi much simpler due to the usb ports being external.

Paul

For a perhaps tidier connection, one of the "flylead" type usb to serial adapters may be better than a full blown usb programmer.

Yes, that page will do - I've edited the link above (copy/paste goof!).

From memory, if you specify NO radio, the 3-phase sketch will default to sending data via the serial port that should be usable (i.e., it sends the raw data as an ASCII string but with no text annotations). I was assuming you'll be using a RPi, not the emonPi, if you want a wired rather than radio connection. And yes, if you've got 3 phases and want anything like real power, that's the sketch that you need. The 'standard' sketches are all single-phase, so will only give apparent power (VA) on a 3-phase system.

[Edit]
It looks as if the version with in-built default to serial output is the one I'm still working on, so you'll need to either add the required print statements inside an appropriate pre-processor directive, or convert the existing print statements to send only what is required by emonHub.

Having a bit of a re-think on the serial connection and wondering if it would be possible to get two or more emon Txes to transmit to a single emon Pi from different locations? It might be that my various switch rooms are all in range so that I only need locate the Txes in them and have the Pi somewhere I can get to more easily.

I think as long as you change the node ID on one of the emon TX so they are not reporting the same node that will be no problem?

"that will be no problem?" Correct. An even better trick is to change the node numbers to ones that are already set up in the emonPi.