Community
OpenEnergyMonitor

Community

Can I measure just Amps in 3 Phase installation

Tags: #<Tag:0x00007f6e00c9d1d0>

I have ordered and IoTaWatt and would just like to measure currents for individual circuits in a house that has a three phase supply. That is there are different circuits that I want to monitor on different phases.

I have seen a post that explains how I can provide 3 voltage inputs (each phase) and then measure watts for each of the circuits however it will be difficult for me to get access to a voltage input for each phase (or any phase at all).

The reduced accuracy of not using voltage is OK as I am just doing an approximate survey of a house electricity use to see where and when the big power usage is happening so I can target appliances for more detailed analysis.

I am thinking I need to modify the code to just insert a fixed 240V constant voltage but it may be more complicated than I think with the zero crossing timing code and the three phases.

Any ideas or comments that will guide me would be appreciated.

Paul

I’m in your situation, Paul. I have an IotaWatt with only 1 voltage reference, a 3 phase house board and no need for high accuracy. I too would like to get IotaWatt to monitor current, multiply by a fixed voltage and publish a power value. I’d prefer to eliminate the VT altogether.

It is actually a much easier measurement to calculate current only and I have used the openenergymonitor code to just measure current without a voltage reference. A number of the commercial units only measure current and multiply by a fixed voltage to estimate power. I have been using an Efergy unit that works like this but I like the iotawatt as I can measure each circuit.

Anyway, maybe an opportunity for a future version of software but for now I will have to run an extension lead to the meter board to get a voltage reference.

It is possible to simulate the other 2 phases from a single voltage reference which is useful – only need one extension lead.

Hi Paul! Can you expand on your setup?

Did you manage to monitor your circuits which are on a variety of phases? How did you ‘simulate’ the other 2 phases. Did you have to modify the IotaWatt firmware?

To clarify my context, I have a voltage reference (on phase B only) on input #1 . I would like monitor all power circuits, which are on a variety of phases. Currently, as I expected, I’m getting good values for phase B circuits, and irregular values for phase A and C circuits.

(I note there is detailed discussion of 3 phase issues in this thread, but this appears to use 3 voltage references and additional circuitry)

Hi Brett,

I am still waiting for my iotawatt to arrive, hopefully next week. The
link below provides information on how to ‘simulate’ the other two phase
voltages.

Essentially you just need to add a 120 and 240 degree phase shift for
the CT that is on the other phases that you are not measuring. Actually
its a 123 and 243 phase shift to allow for the additional phase shift of
the CT and measurement time.

Not sure how to easily work out which phase gets the 123 and which gets
the 243 yet. If you are actually measuring A phase with a VT then a good
guess is that B phase is 123 and C phase is 243.

Let us know how it goes.

Paul

Thanks Paul, that’s just what I needed. I’ll configure that at my 3 phase site and let you know how it goes. I’ll report the results at the link you provided. From the context of that link you’re in NZ, so our results should match - I’m in Australia.
Regarding the shift, my VT is fixed on Phase B. I figure the “B” is arbitrary but that the phases are ordered (ie. ABCABCABC…) so will try:
B phases unshifted (total 3)
C phases with CT config set to phase +120 (total 123)
A phases set to +240 (total 243), or perhaps -120 (total -117).

Brett

Brett,

Phases should be in order if labelled correctly. I think +123 for C and + 243 for A looks logical. Unless the loads are unusual the power factor will be a good indication that you got it right, should be something around 0.9 to 1 or around there anyway. Otherwise just try +123 for A and + 243 for C.

I added to somebody else post (from NZ) - I am actually in regional NSW.

Hope it works.

There is an undocumented API that you may find useful,
Iotawatt.local/command?vtphase=n

Where n is the input channel. This will return the measured phase difference between that channel and the VT at that instant.

  • Configure the CT from the tables as if it is on the phase of the VT. Don’t modify the phase yet.

  • Insure that the circuit has a significant load with a reasonably high PF.

  • Issue the api request.

  • The returned value will be the measured phase difference.

  • I think the returned value should be close to +/- 120.

  • If it’s -120, that should mean this phase is 240.

I haven’t tried this on an actual three phase setup, but it does recognize reversed CTs as ~180.

Be sure you CTs are not reversed, I suspect that would yield a result of +/- 60, but I could easily be wrong about that.

Thanks Bob & Paul. Thanks for a great product, Bob. I’ve just updated the IotaWatt in a 3 phase board. Here is the summary:

Setup:

  • The board is 3 phase with individual circuits on a variety of phases. My VT is on phase B.
  • Using SCT-013-000 and SCT-006-000 CTs
  • Australian system (240V 1 phase / 415V 3 phase)

Notes:

  • As Bob says, a reversed CT with the “command?vtphase=n” API returns confusing values, such as 60. This is probably +240 (phase shift) - 180 (reversed) = 60.
  • Seems that for CTs on phases other than the VT phase (B in my case), the Iotawatt cannot know if it’s reversed.

Steps I took:

  • Ensured “Allow negative values” is on under each “Input”
  • Firstly found the ‘correct’ orientation of the CTs on B phase, using trial and error. I thought the arrow and the writing side (on the CT) face the meter. This gave negative values so I reversed all CTs to point to the load. This gave positive power values on the B phase. Either my understanding of the instructions for the CT orientation is wrong, or perhaps my VT is wired backwards.
  • Use Iotawatt.local/command?vtphase=1,2,3,4,5… on some high PF, loaded circuits. I found A phase ~-120 and C phase ~120. I set inputs to “Generic CT”, input the windings (SCT-013-000 = 2000, SCT-006-000 = 800) and set the phase to +243 (for Phase A) and +123 (for Phase C).

The results seen right: power values as expected, solar is (correctly) negative. I don’t know if accuracy is good/bad compared to using 2 (or 3) VTs.

1 Like

This isn’t necessary. There will be a little “reversed” circular arrow before the name of any ct that reads negative. The allow negative should only be needed on a circuit that can actually have reversed net current flow, like a main with solar downstream.

That’s the key. The polarity of a VT is somewhat arbitrary. It should be consistent for a particular brand, but can and does vary. It’s not possible to reverse the VT in 230V polarized plugs, but US VTs typically can be reversed just by reversing at the plug.

This is an excellent example and a great experiment. The accuracy will not match the three VT setups, but it may well be close enough for many that want a simpler three-phase solution. What I would like to see is a reconciliation with the meter. If you could read the meter at regular intervals, like every few days, and record the aggregate kWh measured by IoTaWatt at that point in time. I do that by creating a total power output that adds all of the mains, then graphing the “energy” of that output. The high value is the current reading. If you could post a summary of that over a couple of weeks, I’d have a better idea of tha eccuracy of this method.

If it prove# to be viable to say, better than 5%, It would be worthwhile to simplify the configuration procedure with automatic phase recognition, which I think is possible in most cases.

Thanks for blazing the trail on this.

Update: system is working well.
For this site I haven’t measured incoming mains and cannot read the meter regularly. I’ll try to do that on the next site.
In future I would test the polarity of the vt and rewire to suit - positive for centre pin?
Regarding summing to check accuracy, it would be great to sum all subcircuits and compare to incoming mains and the meter. However to sum subcircuits to check against incoming main, i think we’d need to add real and reactive power individually. This is tricky as no pf is published by iotawatt and emoncms cannot sqrt to get apparent summed power.

Not following you there. I think both the meter and iotawatt are measuring real power.

I have just recently got a setup the same as this going at my house. 3 phase power measuring voltage on a single phase and i also have net solar.
We have been at home most days and it has been hot so we have consumed some power, the latest reading i have is
meter = 391.9 kWh
Iotawatt = 387kWh

So about 1.25% error at the moment.

1 Like

That’s actually much better than I expected, and very encouraging. I have been toying with the user interface to facilitate this approach. I believe I’ve got a reasonable algorithm for automatically identifying the phase shift on each CT, so that it may be possible to simply indicate a “three-phase” system up front and go about configuring your IoTaWatt just like a single phase user. We’ll see. I may be asking you to.test that out.

For the benefit of those just tuning in, this user is doing three-phase with just one voltage reference. I’m calling this “ derived reference” because the voltage/phase reference for all phases is derived from the single VT.

This may turn out to be accurate enough for all but the most discerning users, but it’s still too early to tell.

Thanks for the feedback, keep the updates coming.

Another update.
Meter = 926.3 kWh
IotaWatt = 915.5 kWh
About 1.17% error.

Thanks Andrew. The delta is about 6kWh over about 630 kWh or a little less than 1%. That’s amazing. I would be satisfied with that in a single phase system.

I’ve got the new release ready to go with all the enhancements to make derived reference really easy to configure (as in the Wiki). There’s an unrelated issue with the local graph utility on IOS devices but may put it out to ALPHA anyway.