Extending Cables from a CT to EmonPi

I have just extended the CT cables from the meter box on the outside wall of my house to the consumer unit where the EmonPi is located. I have followed the advice on which cables to buy and where to connect the braid to earth etc found at Learn | OpenEnergyMonitor. I have used mono jack plugs rather than the stereo suggested but I’ve read elsewhere that mono is OK too. I have found though that the readings are about -300W when the CT should be reading zero as the CT is not clipped around the main live cable. I have checked the CT elsewhere (without the extension cable and it reads zero correctly) and I have tried a second CT and that does exactly the same so I don’t think its the CT. I also have a second cable installed (for a power diverter) and this cable does exactly the same either of the CTs.

Any advice on what I’ve done wrong or is this an expected result that I just need to calibrate out ?

How long are the cables? And do they run near to the meter tails? Where have you earthed the screen?

Check that there isn’t a connection from the plug body to anything other than the c.t. secondary winding, because the plug body is actually the input connection, the plug tip is an a.c. ground (but 1.65 V d.c.).

What do you read if you short the cores at the c.t. end?

A mono plug is OK because the ring of the stereo plug is not used.

Hi Robert,

Thanks for replying so quickly and for your advice

The cables are about 20m in length.

I’m pretty sure that the meter tails run through the loft but the electrician did a very good job of hiding them ! They come into the consumer unit from above (I think without taking most of the consumer unit apart to see properly.) so that would tally with the idea that they go through the loft. My new CT extension cables go in a completely different route under the gravel drive and then into the consumer unit through the same loft. It is possible they pass reasonably close at this point but only for a short length.

I have screened the cable at the CT end by connecting the screen via a short length of cable to the Earth connector of a normal 230v plug.

I’m fairly sure that I have good joints between the extension cable and the CT and also at the EmonPi end with the connector. In fact I borrowed one of Robin Emley’s ideas and cut the CT cable in half so I didnt need to make the joint on the plug end and then I just have two connectors at either end. I have used Wagos and then I used some chocolate strip and then I soldered then (3 different attempts to make sure the connections were good enough) and these 3 attempts made no difference to the answer.

I then made sure that the Emonpi wasn’t the issue and connected the CT without the cable extension but using Wago connectors to reattach it to the plug and that read zero (+/- a few Watts). So the CT is working and I haven’t programmed the EmonPi to do something strange.

I then reconnected the extension cable and shorted out the far end (with and without a CT attached). I also measured the resistance of the cable from the connector along the cable and back to the connector.

No CT -320W and MegaOhms
No CT and shorted -312W and 1.3 Ohms
CT and shorted -354W and 1 Ohm
CT and not shorted 101 Ohms

Before I dig up the cable again I thought I’d see if you had any further thoughts.

Should have said that CT and not shorted was -360W

I presume you haven’t put the c.t inside the consumer unit because there isn’t room?

Try removing the screen earth at the c.t. end and instead earth it at the c.u. end by the emonPi. Although in theory that should be less good, it’s the next (and at the moment, only) suggestion, while I think about the way the noise pickup - because that’s what it seems to be - is getting in.

There is a second consumer unit in my garage to charge the car so I wanted to pick up all the power before the tails split.

I think I see what you’re saying now. Your infeed and meter is in the garage, where you have Henley blocks. One feed goes to the garage c.u., the second makes its way via an unknown route into your loft and hence to the house c.u.
I couldn’t quite understand how you’d got a socket by the meter - that’s what most people don’t have.

And you want the total of both, i.e what the meter reads.

Exactly right.

A few more details.

I have just made sure that I hadn’t corrupted the Emonhub config file. I dont think I have - is there anywhere that can change an input ?

nodename = emonpi
names = power1,power2,power1pluspower2,vrms,t1,t2,t3,t4,t5,t6,pulsecount
datacodes = h, h, h, h, h, h, h, h, h, h, L
scales = 1,1,1,0.01,0.1,0.1,0.1,0.1,0.1,0.1,1
units = W,W,W,V,C,C,C,C,C,C,p

I have dug up the cable (it wasn’t a big job) and laid it so that it was away from the house and got the same results. I don’t think it is the route of the cable.

I have tried connecting the earth at each of the cable and the same result. I did notice that the input read -176W without the earth connected but -340W with it connected.

I have now connected a completely different cable (an old reel of telephone wire I had). I left the cable on the reel and connected a CT to one end and the Emonpi to the other. This was significantly better at just 27W out - I suspect I wasn’t overly careful about the connections.

I am therefore questioning the cable now. This is the spec:


Conductor Material – Pure Copper
Area, conductor CSA – 0.22mm²
Conductor make-up – 7/0.2mm
Shield braid make-up – 96/0.1mm
Filler – Cotton Yarn
External diameter – 5.15mm
Impedance – 75Ω
Capacitance Core To Core - 80pF/m
Capacitance Core To Screen - 215pF/m
Cable type – 2 Core Microphone
Insulation Material – PVC
Sheath Material – Polyvinyl Chloride PVC
Colour – Grey
Operating Temperature – -20ºC to +60ºC

A highly flexible outer PVC sheath and cotton in-fills next to the conductors makes this cable extremely resistant to damage that may be caused by bending or twisting and so it will stand up to heavy use. The pure copper conductors ensure the signal quality is excellent, the dense shielding surrounding the conductors keeps out any unwanted interference.

Is there anything about this that would concern you ?


This is weird, that’s what I would call a much better cable than the - probably unscreened - telephone cable, and very close to what I’ve used in the past.

If the telephone cable gives you 27 W of noise/pickup, I’d go with it.

In emonhub.conf - the “scales = …” line does apply a multiplier to the input, e.g. Vrms is multiplier by 100 inside the front end, sent as an integer and multiplied by 0.01 in emonHub to put it right again. The 6 temperatures are similarly multiplied by 10.

When I ran the telephone cable out to the meter box I got something a bit bigger but you are probably right that I should choose the cable with the lower value.

Is there anywhere else that modifies the input or is it just the scales= line ?

Do you have any views on the fact that earthing the screen makes it worse ?

As you say all a bit weird and In some regards I’m glad that it wasn’t a silly mistake that I’ve made but something a bit more challenging - what else would I been doing in Lockdown ??


My thinking goes along these lines:

You’re seeing (say) 333 W of “noise” power in your c.t.
The full scale power is 100 A × 240 V = 24 kW.
At the ADC input, that’s 50 mA @ 1.1 V.
So at the ADC, you have 700 µA of noise current in the burden resistor, and 15 mV that the ADC is seeing.
I’m beginning to wonder if there’s a significant potential difference between the house earth and the garage earth?
Did you try earthing the cable screen only at the house end?

Earthing the circuit earth of the emonPi could pose a problem if you connect (say) a laptop that’s also earthed and there are circulating earth currents, so that’s not generally advisable.

But it might be worth earthing the case, via one of the screw heads for the end plate.

The maths all makes sense. I have tried earthing at either end of the cable but I’ll do it again just to make sure.

When you say earthing the case via one of the screw heads are you referring to these star screws ? And I connect that to the same earth as the cable screen?

Yes to both. (“Torx” is the name, the size is T20). The anodising on the extrusions prevents a good contact, but those self-tapping screws have bitten into bare metal.

Thats solved it !

It was earthing the EmonPi case was the answer. Just earthing the sheath at the consumer unit end was the same as earthing it at the meter box end but earthing the case, via the Torx (something else Ive learnt!) was the answer. I now get about 20W error with a 5W noise on top. I have added an offset line in the app so that I get a roughly zero W without the CT clipped on to compensate.

Thanks so much for all your advice and know-how. I really appreciate it.

1 Like

Out of interest, I fired up my emonPi with 2 c.t’s (a reasonable distance from any cables) but only with their fitted leads (no extensions) and ran it for about 2 hours. Part of the time, the case was earthed. I’ve graphed the two powers separately, as they overlap rather a lot. Without giving the game away, I’d say it’s very difficult to discern when the earth is on and when it is not.

This is not of course inferring that I don’t believe you - I do, I know just how unpredictable noise and suchlike can be - but it rather illustrates why nobody has worried about it much.