EmonTx readings

Getting Started
1

EmonTx readings

I’ve just bought an EmonTx with 4 CT sensors and UK AC PSU from the shop. The plan is to connect it to a mini PC and read the values over serial. Whilst testing out the unit to get some basic understanding on how it works I noticed one of the CT ports seems to read higher than the others.

As you can see from the readings below, power on CT1 is generally 4-5 higher than CT2 and CT3.

MSG:550,Vrms:244.38,P1:190,P2:186,P3:185,E1:304,E2:297,E3:297,pulse:1
MSG:551,Vrms:244.38,P1:185,P2:180,P3:180,E1:305,E2:298,E3:297,pulse:1
MSG:552,Vrms:244.22,P1:251,P2:246,P3:246,E1:305,E2:298,E3:298,pulse:1
MSG:553,Vrms:243.75,P1:210,P2:205,P3:205,E1:306,E2:299,E3:299,pulse:1
MSG:554,Vrms:243.77,P1:195,P2:191,P3:190,E1:307,E2:299,E3:299,pulse:1
MSG:555,Vrms:243.78,P1:191,P2:187,P3:186,E1:307,E2:300,E3:300,pulse:1
MSG:556,Vrms:243.78,P1:175,P2:171,P3:171,E1:308,E2:300,E3:300,pulse:1
MSG:557,Vrms:243.65,P1:172,P2:167,P3:167,E1:308,E2:301,E3:301,pulse:1
MSG:558,Vrms:243.51,P1:180,P2:176,P3:176,E1:309,E2:301,E3:301,pulse:1
MSG:559,Vrms:243.52,P1:185,P2:180,P3:179,E1:309,E2:302,E3:302,pulse:1
MSG:560,Vrms:243.41,P1:210,P2:206,P3:205,E1:310,E2:302,E3:302,pulse:1

I’ve swapped the CT sensors between ports and the higher reading stays with CT1.

With the CT sensors connected but not clamped round any cable I see the below:

MSG:1,Vrms:240.90,P1:1,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:2,Vrms:240.95,P1:0,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:3,Vrms:241.35,P1:0,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:4,Vrms:241.64,P1:0,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:5,Vrms:240.78,P1:1,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:6,Vrms:240.23,P1:1,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:7,Vrms:240.14,P1:1,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:8,Vrms:239.97,P1:1,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:9,Vrms:240.25,P1:2,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:10,Vrms:240.27,P1:2,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:11,Vrms:239.98,P1:3,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:12,Vrms:240.46,P1:4,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:13,Vrms:240.80,P1:4,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:14,Vrms:241.11,P1:4,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:15,Vrms:241.58,P1:5,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:16,Vrms:241.54,P1:5,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:17,Vrms:240.98,P1:5,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:18,Vrms:241.28,P1:5,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:19,Vrms:241.06,P1:5,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
MSG:20,Vrms:241.09,P1:5,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:1
…
MSG:163,Vrms:243.77,P1:7,P2:0,P3:0,E1:2,E2:0,E3:0,pulse:1
MSG:164,Vrms:243.86,P1:7,P2:0,P3:0,E1:2,E2:0,E3:0,pulse:1
MSG:165,Vrms:243.87,P1:6,P2:0,P3:0,E1:2,E2:0,E3:0,pulse:1
MSG:166,Vrms:243.91,P1:7,P2:0,P3:0,E1:2,E2:0,E3:0,pulse:1
MSG:167,Vrms:243.03,P1:6,P2:0,P3:0,E1:2,E2:0,E3:0,pulse:1

I also discovered that when running the EmonTx using a separate DC PSU (in addition to the AC for voltage measurement) the offset on CT1 increases.

MSG:11,Vrms:242.31,P1:11,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:0
MSG:12,Vrms:242.52,P1:11,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:0
MSG:13,Vrms:243.36,P1:10,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:0
MSG:14,Vrms:243.76,P1:10,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:0
MSG:15,Vrms:243.38,P1:11,P2:0,P3:0,E1:0,E2:0,E3:0,pulse:0

This is with stock firmware and no calibration performed as yet.

Is this to be expected due to manufacturing tolerances? It seems a bit odd that only 1 CT port is off.

Would calibration help here?

Appreciate any advice or experience other users have.

Thanks.
Paul

2

Welcome, Paul to the OEM forum.

Don’t compose your message offline - when you do that and then copy the text into the edit window here, the system thinks you’re a spammer and your post is trapped awaiting moderation. This is why it didn’t appear straight away.

That’s exactly what is happening. Actually, you’re probably lucky to have the other 2 inputs so close together. Calibration - at a sensible current, bearing in mind the full scale current is 100 A for inputs 1 - 3, should bring CT1 input into line with the others.

You’ll always see a small power - it is electrical noise that’s being measured and appears as a power. When you use the d.c. input, one source of electrical noise could be that power supply, so this will explain the slightly higher ‘noise’ power. Bear in mind your maximum power will be around 24 kW, so 10 W in 24 kW isn’t a great deal in absolute terms. Calibration won’t change that. What will change it, if your wiring allows, is to use a multi-turn primary winding for your c.t. (Pass the wire through a number of times.) This makes 1 A in the wire appear as n A, n being the number of passes through the c.t. Then change the calibration by dividing by n. That will also divide the noise it measures by n. Of course, you can’t do this with a meter tail!

There’s an article in ‘Learn’ about the component tolerances, if you feel in investigative mode.

Incidentally, NEVER clamp or wedge the c.t. onto a cable, unless the mounting is specifically designed to cater for this. The ferrite core is brittle and will snap with very little force applied to it, which will effectively destroy it beyond repair. It’s quite OK to let it hang loosely on the cable, where the cable is inside the opening makes a barely detectable difference to the reading. All you need do is make sure the mating faces of the core are clean, then clip the two halves together.

3

Hi Robert, thank you for your reply.

Don’t compose your message offline - when you do that and then copy the text into the edit window here, the system thinks you’re a spammer and your post is trapped awaiting moderation. This is why it didn’t appear straight away.

Noted. It’s a force of habit to compose offline as in the past I’ve been bitten more than once by losing a post when submitting it for one reason or another.

Incidentally, NEVER clamp or wedge the c.t. onto a cable, unless the mounting is specifically designed to cater for this. The ferrite core is brittle and will snap with very little force applied to it, which will effectively destroy it beyond repair. It’s quite OK to let it hang loosely on the cable, where the cable is inside the opening makes a barely detectable difference to the reading. All you need do is make sure the mating faces of the core are clean, then clip the two halves together.

I’ve read the CT article on ‘Learn’ so am aware of this. What I actually meant was that the CT sensor was connected to the EmonTx but didn’t have a wire going through the CT. Therefore the expectation would be to see zero power used.

Like you say, 10W in 24kW is a very small inaccuracy overall. What threw me was that fact the other two CT ports almost match each other and when no power is monitored they measured zero. I’ll count myself lucky on those appearing to be so accurate. :slight_smile:

You’ve confirmed my thoughts on the calibration point so thanks for that.

I have an oil filled radiator that I think is 2kW, using that and the trick with the multi-turn winding through the CT should help me test with higher loads.

Thanks for the advice.

4

Discourse - the software we use - is pretty robust. A part-written post will appear in another browser even, and survive restarting your browser.

That’s how I’ve tested the SCT-013-000 up to 250 At (Ampere-turns, not plain amps, 5 A & 50 turns - but I have to be careful with not heating the coil too much.)

5

It’ll survive a reboot as well. (I rebooted in the middle of composing this post to confirm)