Anomalous power reading

I have a solar power installation with a 50A AccuCT on each of the two phases. I’ve noticed that the one phase consistently shows a larger power output than the other. This difference even persists at night, when the one phase indicates 0W, as expected, and the other one indicates around 5W. This morning I moved the CT from the phase exhibiting this strange behavior onto the other phase (i.e. both CTs are now on the same wire), and found that the difference in power measurements persists.

Can you think of any possible cause other than the CT being bad? I’d appreciate suggestions for troubleshooting.

Yes, there are other possibilities.

Putting the two CTs on the same wire was a good starting diagnostic. Can you confirm that the 0W/5W situation exists with both on the same wire? If so, what do they read if you remove both and close them with no wire?

Next, do both cords follow the same route to the IoTaWatt, and do either pass close or loop around conductors that may have high current?

Last, if you swap inputs used by the plugs, does the problem move with the plug or stay with the input?

Yes, the 0W/5W reading remained the same when I put them on the same wire. I only removed the 5W CT since the other one was difficult to get into place, and removing it seems to have much less diagnostic value. The CT with the strange behavior reads 0W with no wire inside it.

Yes, they follow the same route, and they don’t pass close to any high-current conductors. After removing the one to take a reading without a wire, I put it back right next to the other CT, on the same wire, and made sure that the signal cables run next to each other to the IoTaWatt. The readings are 1923W for the one that is 0W at night, and 1845W for the other.

I forgot to mention in my initial post that I had already tried this. The problem remains with the CT rather than the IoTaWatt input channel.

So just to be clear, with both wires on the same conductor, at night, one (lets call that one A) reads 0W and the other (B) reads 5W. When you remove B from the wire, it goes to 0W. Is that correct?

Yes, correct. Also worth mentioning is that B makes a buzzing noise and vibrates depending on how it’s arranged on the wire it’s attached around. I haven’t noticed any of my other CTs doing this.

I don’t have a problem with replacing the CT, I’m just not comfortable with the reported symptoms indicating that. This would be the first time I have seen this failure mode.

The CT is producing output when clamped on a wire that the other says is 0W, yet when it’s not on any wire it says 0. Then, if I’m reading this correctly, the CT B reads substantially lower than A when there is 15-16 Amps in the wire. So it’s higher at very low current, and lower at moderate current.

Just to close the book on this, can you substitute another CT for B, using the same input as B and clamp it alongside A to verify that they then agree?

I understand completely. The behavior is really strange, so it’s perfectly reasonable to explore all possible alternative explanations before assuming that the CT us bad.

Will do. I’ll report back when I have the results.

Last night I confirmed that a different CT shows 0W when connected to the same wire as CT A. Repeating the test this morning, I am seeing some differences between CT A and the 3rd test CT (e.g. 1852W vs 1845W), but these are much smaller than the differences between CT A and CT B (e.g. 1967W vs 1835W).

I also noticed that on CT B, unlike on the other CTs, the top part of the core is loose in its housing, which no doubt explains the intermittent vibration and buzzing sound it emits. Could this also explain its inaccurate readings?

That’s better.

That’s a 0.4% difference and well within the margin of error when comparing results of a varying load using the status display. The status display is showing a damped average, with the contributing samples taken several hundred milliseconds apart. The best way to compare two inputs for agreement is to plot them and compare the total Wh. Even a 10 minute window will show much better how the compare.

Sounds plausible, I’ll send you another 50A replacement.

Thanks! Before you do that, though, would it be worth me trying to glue the loose half core in place to see if that resolves the problem?

That’s up to you, I’ll send the replacement anyway.

Thanks, much appreciated. And thanks for the great support in figuring out the cause of this problem.