Newbie - what sensors do I need?

Hello everyone.

I’ve just managed to get hold of a bare emonTx v3 & power supply secondhand. Now looking to add the necessary peripherals to monitor my home (single phase) setup.

Specifically I’d like to figure out how much of the PV-generated power we use domestically and how much is exported to grid. In the hope of ultimately sizing a home battery once we can afford it.

We have a solar edge inverter, a MyEnergi Eddi diverter and two separate SunAmp heat batteries which store excess PV generation for hot water use.

I have a couple of years’ worth of data from the Eddi and SolarEdge which tells me how much we are diverting into each battery and how much PV energy we generate overall. But most of the rest (e.g. where the non-hot-water electricity goes) is extrapolation and conjecture from our mains meter.

I figure I need at least two CT clamps (100A?), an optical pulse meter for the mains, possibly a voltage meter and I’ll need to either buy or build a compatible emonbase. Any other hardware suggestions would be greatly appreciated.

Welcome, Tom, to the OEM forum.

First, a question - is the emonTx power supply a 5 V d.c. USB one or a 9 V a.c. adapter? Because the latter can also function as a power supply, and having one will mean you can know the direction of power flow as measured by the c.t’s. You don’t know this without the a.c. adapter, its prime function is to measure the voltage, both to know the direction of power but also for accuracy, as our UK mains voltage varies quite a lot.

The second, maybe obvious, point is the optical pulse reader is of limited use on your grid meter, because you only get pulses on import – you know nothing about the energy exported. It is however very useful as an aid to calibration.

As a minimum, I suggest you need a c.t. (probably 100 A) on your main grid connection, and one - possibly a smaller one for better accuracy (we can recommend some, but they might mean adding or removing a resistor inside the emonTx). You’ve then got 2 channels left - if you have room inside your consumer unit you could consider putting a bundle of ‘similar’ circuits through each to give you a better idea of where “the rest” goes.
If you have a pulse output on your generation meter, you could use the optical input on that (the power consumed 'dark" is normally below about 15 W, so barely significant in the overall scheme of things) and free the channel for something else.

An emonBase is a Raspberry Pi - running on an identical SD card to the emonPi - but equipped with a radio receiver. But, if it can be close to the emonTx, you don’t even need the radio and you can use a direct serial connection.

As a matter of some importance when you’re looking for a sketch for your emonTx, which V3 is it - V3.2 or V3.4? The V3.2 has the processor on a sub-board (RFµ328) piggy-backed on the main p.c.b, with the radio module on that, while the V3.4 has the processor directly on the main p.c.b.


Thanks for the detailed reply. To answer your questions/points in order:

  1. The power supply is labelled 9V AC, so it sounds as though this can be used to measure the voltage

  2. We have two meters, one is for mains grid import, the other is for total PV generated at source (e.g. not a measure of how much is exported). It sounds like using the pulse reader on the generation meter is the best option.

  3. I’ll definitely get a 100A CT for the mains. Please can you suggest an appropriate size for the second clamp?

Due to the way the Eddi diverter works, I already have 2 CT’s (1 on the mains grid and 1 on the PV supply) so that it can prioritise using surplus PV energy for the hot water. Assuming that CT’s don’t interfere with each other, I had planned on putting the emonTx clamps next to the existing ones.

I guess that I could just use the remaining two slots for CTs on the supply cables to each of the heat batteries so that it can measure what amount of the total has been diverted to these. The remainder would give an overall domestic use but nothing finer grained. Would that be sensible? Diagram below, slightly edited to reflect my setup.

  1. I have a spare Pi Zero W but I’m assuming this is not capable of reliably being used for emonBase. Can you advise what model of Raspberry Pi I’d need to get hold of to build one that is compatible with my version of the emonTx?

  2. On that note, I probably jumped the gun with this post - the V3 won’t get delivered until later this week so I can’t check the exact version yet! I’ll update this thread once I know more…

Yes, it will automatically be doing this.

Agreed - but you can also use it to help calibrate your c.t’s (temporarily, before transferring them to their intended position).

What is the maximum current you’ll be measuring? Your c.t. wants to be not less than this, with a small margin above for safety. Bear in mind (this probably won’t apply to your P.V. infeed, certainly not to your water heating) inrush currents and power factor if you’re thinking of things like fridges etc with induction motors, or a lot of fluorescent or LED lamps.

Input 4 of the V3 (both 3.2 & 3.4) has a high sensitivity input intended for a P.V. infeed when used with the standard (OEM Shop) YHDC SCT-013-000 c.t., so you can use this. For the thermal store, you probably will want one or two of the Talema range from RS Components (why two? - if they are energised separately, put the c.t. inside the Eddi with both line conductors going through, this inherently gives you the total)
My note says:

UK c.t’s 25 A / 40 A / 50 A: or /7754903 or /7754912 Secondary 50 mA, good to 1.6 V. (Ignore ±400 A Output, it’s wrong.) 50A:50mA (good to 1.6 V) 25A:50mA (good to 1.6 V)

With a stereo 3.5 mm plug fitted (extending the wires if needs be), they will directly replace the YHDC one, all that’s needed is a change to the calibration.

They can interfere if they’re touching, but 20 - 30 mm gap should be fine.

@borpin - Can you say yea or nay to this please?

Thanks again for responses.

The thermal batteries are separated physically (one supplies main bathroom & kitchen, one does en-suite), but they are also fed by separate outputs from the block. Please see lower right ‘HEATERS’ in image below ‘borrowed’ from the internet. So I think I have to use separate sensors unless there’s a clever way to do a double-clamp just after they have left the diverter?

Each cable is 16A rated. So 25A CT for each?

Since I definitely want to know what proportion of our usage is PV, I think CT clamp 2 has to be on the PV infeed. This cable is 16A also.

It should work OK as long as you don’t expect too rapid a response if pulling lots of data into a chart. For simple analysis and collection of data it’ll be fine to start with.

I thought I’d told you this - you might need to strip the cables back a bit, but it’s perfectly in order to put 2 line (brown) wires through one c.t., and it will measure the vector sum of the currents. As both are on the same phase, and both are resistive loads, that should give the total power.

Don’t be tempted to use one line and one neutral, the neutral will read the power but negated, so you’ll read the difference - and if the energy is shared equally - little or nothing!

Hello again. So I now have the emonTx and can remove the cover. The board is labelled v3.4.2.

Can you tell me what this affects? The type of sensors I can use, or something more fundamental?

Only the sketches you use. It’s only the emonTx4 that uses different sensors, all the previous versions (V2, V3.2 & V3.4) use the same 9 V a.c. adapter for voltage and the same YHDC SCT-013-000 current transformer (or the Talema range I mentioned above).