IoTaWatt three phase configuration

Tags: #<Tag:0x00007f6e156beee8>

I live in France and my house is electrical supply with 3 phases.
I have got an electrical car (Renault ZOE) with a 22 Kw charger (three phases).
I would like to monitor my electrical consumption in different locations :
Total consumption, Car charger consumption, First floor consumption and so on.
I bought a IoTaWat with three AC-AC Voltage Sensor Adapter (VT) and nine current sensor SCT-013-000.
I thought it was possible to plug this three VT sensors to take voltage reference on each phase but I did not get that the plug is not the same as the current sensor…

Please could you tell me if there is a place where I can find some information about using IoTaWatt in a three phase environment ?

Do I need three IoTaWatt box for monitoring each phases ?


Hi Thierry,

You will need adapters to connect two of the VTs to the stereo jack inputs. The adapters add a 288 ohm resistor in series. I am working on supplying OEM with some, hopefully by the end of January.

I’ll try to get some documentation in the wiki before then.

Thank you for your answer.
I will wait for those adapters.
Let me advise when available.


First a great thank for such a good product as IOTAWATT for me is.
Bought 2 pieces .
Here in Austria we have also 3phase system.
Same problem as gth has.
Are any new about the adapters ??
Need to buy one.
If necessary i can then compare a “real” 3 phase IOATWATT measurement with the derived 3 phase measurement and lokk how much difference is in the praxis.

Sorry, I have all the parts and boards, just haven’t made a batch. I’ve had such positive feedback about the derived method that I’m wondering if this is even necessary. Nevertheless, I’ll make it a point to put some adapters together over the next few days and get them off to the OEM shop.

EDIT: Neglected to mention that the newer adapters are different from the prototype adapters that a few users have. They will require that the burden resistors be removed from the input channels that they will be used on. This change is especially important if using multiple IoTaWatt with splitters on the VTs. They reduce the load on the VTs, and more importantly uncouple the DC bias. Bottom line is that you can use splitters to share the same set of AC adapters across any number of IoTaWatt.

BTW/ I’m more reliably at

The reason why i will run a comparison is that our voltage is relative instable compared to grids in urban cities.
We have up an down from 215 up to 240 in shorttime. So im intersted who accurate the IOTAWATT works in such a environment. We also have 50 HZ and you calibrate for 60 HZ. As i understood the difference is not so big but as more as the voltage is instabil i think the failure becomes bigger.

An other point i am personally intersted in is how much differs the three phase voltages together and with 3 VTs i get it easy in a graph.Is a diiference between peak and offpeak hours and weekend and so on ???..

I am also very positive that the accurancy for most of the measurement which you are normally have are really good enough and only for special applications you should have it in mind that is also a 3 phase measurement thith measuring all 3 voltages is possible and give (how much?) better results.

I don’t think the overall variation is As much a problem as phase-to-phase voltage variation. You would be able to measure that with three vts.

I have compared the 60Hz calibration vs 50Hz and found that although there is a difference in phase shift, it is largely proportional across the VT and CT and so the error seems to be pretty small when the two have similar shift. It would be greater with a high shift VT and a say a solid core CT with very low shift.

The adapters are done and should go to OEM Friday or Monday.

Hi Bob,
I‘me a bit new to all of this. If I buy a iowatt how do I request these adaptors?
Many Thanks

Hello Bob,
You helped me a lot with setting up my three-phase system with solar and energy with IoTaWatt. I have tested the system for 2 months and now come back to the forum. I’ve noticed that things have been developed in terms of the three-phase system. It is now possible to set up a three-phase system with ‘Derived Reference’. If I set up the system with ‘Derived Reference’, does that come out the same as the system according to the attached picture below?


Here is a feedback: I compared the values ​​of IoTaWatt with the meter from the power company. The values ​​agree very exactly with the solar part. In the energy part I compared the values ​​of 2 months. The values ​​of IoTaWatt are too large by a factor of 1.54. For both months, this value is exactly the same. I suspect that this is related to the indirect measurement. A current transformer reduces the current by a factor of 40. I will now adapt this factor and look again at how the result comes out.
Thank you for your great work.

Yes, it should simplify the setup. It looks like you are using SCT013 CTs, so with the Derived Reference specification, you can specify those and the phase correction for the CT will be included in the calculations.

I doubt it. I can’t think of any scenario where the power would be overstated. Improperly adjusting the phase should always understate the values. Can you break that down and post the numbers? I am curious what the meter actually reports.

Why is that? My microwave oven in standby (i.e. just being a clock) uses about 3W, but if I introduce a big enough phase error in the measurement, I can make it read anything up to about 80W.


We can continue this discussion at The IoTaWatt forum

What kind of numbers?

Sorry, why? Is it a private forum here? How to switch?

I’m the only contributor providing support for IoTaWatt here, and the bulk of the activity here concerns the other OpenEnergyMonitor products. The IoTaWatt forum is exclusively IoTaWatt, and better leverages my time and effort and builds a body of knowledge specific to IoTaWatt for all IoTaWatt users, not just the OEM community.

The IoTaWatt forum uses the same software as this forum and works the same way. You need only go there and register.

It’s also a place where the theory of making accurate AC power measurements is freely discussed, and since that theory is based on the laws of physics, it transcends any particular implementation. I’ve never seen an IotaWatt or any of the other OEM products.

Here are two for your consideration:

The first is my microwave oven being a clock, as discussed above. The second is my pool pump when it’s sucking on air, i.e. it’s a 1.5hp synchronous AC motor with almost zero load on it.

Those are the actual signals on the wire, they’re not indicative of any measurement errors. In both cases the current signal is distorted (i.e. it’s not a sinewave), but they’re close enough that you could at least imagine what the 50Hz component would look like were you to extract it. In the Microwave case, the current leads the voltage by almost 90° and in the Pool case the current lags the voltage by about 50°.

Now imagine what happens if you introduce a 5° phase error (say) to both of those measurements. If your 5° error shifts the current to the right (i.e. delays it) it will increase the reading for the Microwave and decrease the reading for the Pool. If your 5° error shifts the current to the left (i.e. advances it) it will increase the reading for the Pool and decrease the reading for the Microwave (so much so it will likely make the Microwave appear as a generator rather than a consumer). So without knowing the nature of the load, it’s impossible to predict whether an introduced phase error will cause an increase or decrease in power reading.

When posting these examples it would be a lot easier to knock up something with a signal generator or a spreadsheet, but to avoid acsusations that I’m talking about some theoretical situation that will never happen in real life, I always try to use real measurements from real loads as found around my house: washing machines, hair dryers, microwave ovens, pool pumps, CFLs etc. These loads exist, you can’t just wish them away, and the more your monitor drills down to a per-circuit measurement the more likely you are to have such a load totally dominate the measurement you’re trying to make.


Bob, don’t you think that if the question has been asked here in this forum, then it should also be answered here?
If other openenergymonitor members in the future have a similar question, they shouldn’t really have to switch forums to follow the conversation…


Thanks @dBC thats a really informative post, I was discussing this question of phase error with Robert day before yesterday as I was confused how you could read 80W on a microwave standby load of 3W with a phase adjustment. I assumed the microwave standby would have been a switched mode power supply with a current draw peak in-phase with the voltage. Robert mentioned that the power supply used in your case could be a capacitive dropper which I was not familiar with (
which would result in the out of phase current waveform you show there and the behaviour you mention in terms of phase adjustment.

1 Like

Its probably worth clarifying for your benefit @hph @townlema @MIA @gth and anyone else reading this, While most hardware on the OpenEnergyMonitor forum such as the EmonTx, EmonPi & EmonTH are developments of @glyn.hudson and myself (we run the OpenEnergyMonitor shop, @Gwil is also part of the shop team). The IotaWatt is Bob Lemaire’s development and Bob is an independent self-funded developer, while we have collaborated with Bob on its development, the bulk of the work is Bob’s own. Our arrangement with Bob Lemaire for the IotaWatt is as a distributor of the IotaWatt in much the same way as we are now starting to distribute the OpenEVSE charging station (which also has its own forum).

Bob Lemaire (@overeasy) has setup a dedicated forum for IotaWatt support and development which can be found here IotaWatt forum. Otherwise Glyn and I will try our best to help with IotaWatt questions asked here in future.