That’s shot down my first thoughts as to the cause of the problem.
That surprises me. The link you gave to the cable you bought was not exactly forthcoming when it came to details, but it looked very much as if the cable was a straight 4-core, and not twisted pairs, which is what you should have. On the assumption that the power reading comes from magnetic fields inducing currents in the cable, the twisted pairs of the CAT 6 (assuming you’d used it correctly with a twisted pair for each c.t.) should have performed better than the 4-core cable. That leads me to suspect that magnetic fields might not be your problem.
Earthing the cable screen at both ends was definitely not the best move - you could well have created an “earth loop” and had quite high currents flowing in the cable screen.
I don’t believe the wire thickness (0.4 mm diameter, 0.129 mm²) will have any effect whatsoever - after all, what is the current it’s carrying? Almost nothing: 0.35 µA.
Where is the cable run? Is it adjacent and parallel to cables carrying a heavy current, or an inverter drive, or something like that?
Knowing now that you have an emonPi2, and before you spend more money, I’d suggest connecting the cable screen only to the emonPi2 earth. Connect this to the plug sleeve, if that’s practical. The schematics don’t show a connection between GND and the case, so alternatively, in the absence of a 3-pole green plug (pulse/analogue input), the next best place to connect the cable screen is the GND connection in the emonVs - again assuming you have one. I don’t have an emonPi2, so I can’t check this. I do have an emonTx4 (which is essentially the same input circuit), and some long microphone cables, so tomorrow (if I have time) I’ll try a c.t. on the end of 30 m or so of cable and see what I get.
This is how you should connect the cable to an emonTx4/emonPi2:
For my cable testing I am running it through my front door, down the hall, and into the attic where the emon is (my wife keeps saying “you can’t leave these cables lying there” lol!!) so there is not any electrical cables interfering.
I saw that image and the notes - but it mentions nothing about the lay-up of the cable. Is the whole lot twisted or not? I’ve mentioned “twisted” several times - believe it or not, this is important.
I have connected a voltage-output c.t. from The Shop to an emonTx4, using an emonVs as its power supply. The c.t. is on my c.t. test rig, set to give about 4.6 kW indicated with 1 A indicated (10 At in the c.t.).
First test: Connected via a “short” (3 m) microphone cable and a 3.5mm to ¼" adapter cable. Result: Zero power for zero current, 4.6 kW at 1 A.
Second test: add a 24 m microphone extension lead: Result: Zero power for zero current, 4.6 kW at 1 A.
Third test: add a 18 m microphone extension lead: Result: Zero power for zero current, 4.6 kW at 1 A.
Fourth test: add a 16 m microphone extension lead (total: 61 m): Result: -1, 0 or +1 W recorded for zero current, 4.6 kW at 1 A.
When the 3,5mm plug is inserted into the emon unit, the red and green lines I drew are the points of contact - correct?
What I circled in yellow suggests that the shielding connects to the green line, which ultimately goes to the ground. I am not using an emon-VS, so I was thinking of connecting the shield to the emonPi2 case, but the case does not appear to be connected to any ground circuitry. Can I use the GND pin circled in the picture below?
Lastly, below are the connectors that I am using at both ends of the LAN cable to connect the 3.5mm plug and CT (basically I cut the existing CT cable and reconnected them to both ends of the LAN cable). I connected the shielding to the 3rd position (shown with the line).
Yes, correct. This is how jack plugs & sockets are drawn.
Yes - that is what I had in mind.
Conventionally, pin 3 of your 270° 5-pin DIN connector is ground, so that’s correct.
Go ahead and connect to the GND pin on the green socket.
(I’m using a combination of 3-pole 3.5 mm and 2-pole ¼" jack plugs and sockets and 3-pin XLR connectors, in all 5 mating pairs of connectors. That ought to have provoked a problem - but didn’t.)
This rings alarm bells. Experience has taught us that not all 5 V power supplies are equal - some churn out remarkable amounts of noise. My mention of Rupert above relates to a detailed explanation he sent to us privately about 2 years ago concerning a r.f. suppression component commonly found in switched mode power supplies, which in combination with stray capacitance added by an extension cable produced in the emonTx V3 and emonPi1 exactly the fault you have. What are you using for your power supply? It might be worth trying a completely different one if that’s possible.
When I moved the emonPi2 outside directly onto the outdoor AC unit and connected it with standard-length CT cables everything worked as it should. In this case, I plugged the emonPi2 5VDC into a regular wall wort (110vac to 5vdc adapter). I need to try this same adapter first to see if the power strip is the source of my problems.
When the compressor runs, my emonPi2 is higher than what I measure with an ammeter.
emon shows the fan at 1.55 vs 1.19 on the ammeter
emon shows the compressor at 8.79 vs 8.63 on the ammeter
The differences aren’t huge when the unit is running, but the non-zero values when it isn’t running do present a problem. When I look at the ports that do not have CTs attached, the typical value I see is 0.05. Any thoughts as to what is causing values around 1 amp when nothing is drawing current?
It’s picking up a voltage from somewhere, we need to find out where.
The cable you’re using now is the CAT 6, using two twisted pairs inside the same overall earthed screen?
If you temporarily short-circuit the c.t. (it won’t harm it), what do you see then?
You are using one pair per c.t?
Is there another current-carrying cable near the c.t. or near the cable?
The numbers I reported earlier were with the cable in coils indoors. When I laid the cable outside, on the ground, I saw slightly higher readings, but I was reading power and not just current – about 16 W with a full-scale of ~24 kW. Unfortunately, a direct comparison with current isn’t possible.
Yes, I’ve used two twisted pairs inside the same CAT6 cable (white/blue, and white/green).
Yes, just one pair per CT.
Yes, see the pictures below. The purple wire is for the compressor fan and the yellow wire is for the compressor. Both wires are coming off of the capacitor, and the 220VAC feed is right below to the left. I could try to relocate the CTs or is there some form of shielding I could use?
My CAT6 cable is on the ground and then goes up the exterior wall adjacent to a cable TV COAX cable. Then through the wall into the attic (no power cables nearby) and over to the emonPi2 that is attached to the furnace. The CAT6 cable is 65’ long (approx 20 meters).
I have not attempted this yet as I will have to open up my splice outside and it is 115F right now (I’ll melt!!).
I think the official name for that is a mess. I simply cannot see from a photo what is what amongst that tangle, the only way would be to dive in, disconnect things and untangle it. But having said that, it may or may not be the problem, or even part of it. Unfortunately, the c.t. works by sensing the magnetic field around a cable, so by the very nature of this, it will pick up any magnetic field that’s there, even though it shouldn’t be significant unless the wire causing the trouble is within an inch or so.
Does the emonPi have a good earth, and what do you mean by “furnace”?
Before you do that, unclip the c.t’s and just move them as far from the other stuff as you conveniently can, and see what the effect is.
Is there a wiring diagram on the inside of the cover you removed?
If there is, could you take a picture of it and post it here?
It might help Robert with pinning down your issue.
I’m in the US too, so your picture is similar to the condensers I have.
Hence my question about the wiring diagram.
OT
115… You must be in Arizona or New Mexico.
Brings back memories of when I lived in Alamogordo.
Not quite as hot as where you are, but 106 was a daily summertime high when I was
there. (2002 - 2007)
And that goes for the diagrams too! After a bit of unpicking the details and redrawing, here’s what I think:
Both the compressor motor and the fan motor are capacitor-start, capacitor-run induction motors. (Clue: the ‘R’ & ‘S’ connections refer to the ends of the Run & Start winding, ‘C’ is the common ends of both windings.) The purpose of the ‘start’ winding in series with the capacitor is to provide a phase-shifted second phase in the motor stator so as to get a rotating magnetic field. This winding generally carries a significantly lower current than the main (‘run’) winding. The bad news is, the start winding for the compressor motor is the yellow wire, and the start winding for the fan motor is the purple wire. So you’re not measuring the total motor current of either. You need to put the c.t’s on the black wires coming off the contactor terminal K1-1. I can’t see where the second black wire terminates, so you’ll need to get hands-on in there and find it. If you can’t find those, the other end of the compressor motor is the red and yellow wires (so put both through the c.t. in the same direction!), and the other end of the fan motor is the purple and orange wires (both of those through the c.t. likewise).
Both compressor and fan are controlled by the same contactor, so this begs the question, what do you gain from measuring them independently?
Also, bear in mind that the power factor for the compressor will be quite a bit less than 1, and the fan motor probably even lower, so the current you measure times voltage will not indicate the true power.
Robert, I will tackle this sometime today and let you know what I can accomplish.
The reason for measuring each independently is for diagnostic and alerting purposes. If the fan stops working the compressor will overheat, so we look for zero current draw on the fan. We monitor the compressor for run cycles, constant running, and over-drawing the maximum rating.
Report what you see when you’ve moved the c.t’s, and we’ll take it from there. But while you were only measuring part of the motor current, it didn’t make sense continuing like that. True, you’d get running/failed from it, but that’s about all.
