Brand new here, need some advice

[Faze (Dale)]

Hi OEM community
Brand new here today, so excuse my lack of knowledge.

Recently, as a simple project for myself, I tried to build my own energy monitor and it didn’t work out as well as I had hoped. (I wont go into details). however, I really want a whole house energy monitor with the eventual goal of providing power usage trending of customizable specific time periods.

So looking around I found this group. After some time looking at the products here at OEM, I’m considering purchasing one of these emonPI energy monitors. However I have some concerns that I need validated or corrected and I just need general advice before I fork out my cash.

My power box looks like this and you can see my pi project with a daughter board that is hosting 2x CT sensors (Plan is to replace the current pi with an emonPI):

The CT sensors I purchased were these because they fit (snug) on my cable diameter:

AC Current Transformer 100A/100mA Smallest Split Base AC Current...

Summary CT = 16mm inner diameter, 42 ohms internal resistance, 100A/100mA 1000:1 ratio

Doing a little research, I was reading this article about suitable transformers and couldn’t find one that matched the specs of my CT:
Edit: as a new user I cannot post more that 2 url links, but there is an article on this site about suitable current transforms that is very technical:

I would like to avoid buying new CTs, if possible but will do so if I have to.
So my questions,

1. Will my current CTs work with the emonPI? And if not, can anyone make a recommendation on a set of CTs that will work on my cable diameter and help me achieve a setup where Im trying to measure whole house power.
2. In laymen’s terms, if I need to make modifications to the emonPI which I noticed was mentioned here: Modifications may be necessary, what would I need to change or do because much of the technical language outlined in the link above about suitable transforms is slightly over my level of understanding. i.e I dont know if I need to include additional hardware, or its physical modification to the emonPI or just software configuration changes to the emonPI.

Really appreciate anyone’s advice and patience

[Faze (Dale)]

Technical article about suitable technical CTs:
https://learn.openenergymonitor.org/electricity-monitoring/ac-power-theory/use-in-north-america

[Robert.Wall]

Welcome, Dale, to the OEM forum.

@Faze

as a new user I cannot post more that 2 url links,

That’s not true any more - I’ve bumped your privileges up a notch.

@Faze

Will my current CTs work with the emonPI? And if not, can anyone make a recommendation on a set of CTs that will work on my cable diameter and help me achieve a setup where Im trying to measure whole house power.

Yes they will, but you’ll have to change the burden resistor inside the emonPi - or add another externally. All you need is another 22 Ω in parallel with the one that’s fitted. If you’ve already built your own energy monitor, that shouldn’t present a problem. Doing that modification means you won’t need to change the calibration in the software. (And if you don’t do it, you’ll only be able to read up to 50 A on each leg of your supply.)

Do you intend to put the emonPi in there? Two possible problems: is there room for two 120 V sockets - one for the emonPi’s power supply and one for an a.c. adapter to measure a sample of the line voltage? and second, is there a way for the Wi-Fi to get out, i.e. is the door wood, or is there a metal cover you’ve removed that will pretty much keep it in? - the alternative to that is an Ethernet cable to your router.

[Faze (Dale)]

Thanks for the quick response Robert.

Firstly, with regard to the parallel resistance. Is there any documentation you can link to me on replacing the burden resistor inside the emonPI or how you go about installing an external resister. I actually dont have any external resister installed on my current setup. So not sure how to do this.

To answer your questions, yes there is room, I have a small gap at the bottom of the switch board where cables can access the 120v sockets on the wall. One for the pi and one for the sample line. Its how I have my current pi setup.

The switchboard is very close to one of my ceiling mounted APs for wireless ethernet, signal is very good. Works just fine with my current pi on wifi, so no issues with network access or power point access.

[Bill.Thomson]

@Faze

My power box looks like this and you can see my pi project with a daughter board

@Robert.Wall

Do you intend to put the emonPi in there?

Hi Dale,

I’m in the US too. A couple of things to be aware of…
If you want to keep your load center (breaker box) wiring code compliant, the only other items allowed inside a load center other than circuit breakers and bus bars, are CTs.

The NEC says low voltage wiring (the CT leads) and high voltage wiring (mains) are not supposed to be in the same conduit. i.e. the low voltage wiring needs to be in its own conduit.

If you put the electronics in a separate enclosure, i.e. one that’s not a load center, that would keep
things in line with the code.

[Robert.Wall]

@Faze

Firstly, with regard to the parallel resistance. Is there any documentation you can link to me on replacing the burden resistor inside the emonPI or how you go about installing an external resister.

As far as I know, there’s nothing written specifically for the emonPi. However, the “emon” part of the emonPi is a cut-down version of the emonTx, and adding a second parallel resistor to that has been written about many times.

First question - do you want to buy an emonPi, or would you want to buy the emonPi Shield and the case & display, and use your RPi?

The better, safer method is to have the parallel burden resistor external to the emonPi. Your c.t’s don’t have a protective zener diode, which means without a burden, they’re at risk of damage if unplugged while current is flowing in the cables they are clipped around. The resistor can be anywhere on the cable - where you extended the leads looks to be a good place. It is of course connected in parallel with the c.t. and the plug. (A ¼ W resistor is plenty - the power in it will be 220 mW worst case (i.e. with 100 A flowing, c.t unplugged, normal use will be 55 mW max).

If you don’t fancy doing that, and want the resistor inside, and if you do buy the assembled emonPi, you’ll first need to disassemble it. With nothing connected, take off the acrylic end (nearest the pushbutton, a T20 screwdriver is required), and slide out the display and the Pi plus the ‘emon’ part together just an inch or so, taking care not to strain the push button and display connections.

If you now look down behind the two 3.5 mm sockets, you’ll see the burden resistors - the big ones labelled “22R0” - with a hole for a wire-ended resistor at each end. That’s where you solder in the second parallel 22 Ω resistors.

If you’re not confident you can do that with the two parts still assembled, you’ll need to continue with the disassembly. Pull everything out of the case. Unplug the connectors to the display and pushbutton at the PCB end, noting where they went. Next, remove the three screws holding the two boards together via the spacers, then you’ll probably need a flat-bladed screwdriver to lever the connector apart to split the boards.

(Hint: With the boards separated, you can plug in a programmer and the USB power, and use the Shield like an emonTx - so you can check it works and gives the correct current/power after you’ve added the resistors.)

Reassembly is fairly straightforward - if the connector feels as if it hasn’t mated adequately, you’ll need to adjust where the extender sits between the boards. The ‘emon’ PCB only fits properly in one pair of slots, the narrowest.

[Bill.Thomson]

and slide out the display and the Pi plus the ‘emon’ part together

Snapping a picture or two of it at that point might not be a bad idea.

[Faze (Dale)]

@Bill.Thomson
I did not know this, Thanks for the info. I’ll just move and mount the emonPi to a small rack on the outside and below the breaker box, I have a shelf it can sit on and the low voltage wiring can be routed up into the breaker box. cheers.

@Robert.Wall
Thank you so much for the detailed description and pictures! Really appreciate the feed back and I’ll probably be attempting installing the additional 22 Ω resister internally.
To answer your questions. I after the feed back I got here, I went and purchased an emonPi and so I’ll be replacing my Pi and just reallocate it for some other project. I had fun with it it…, built a simple python script which took line reads from a daughter board on the pi and injected that data into an influx DB. I then had grafana plot the data into a graph which I would display in openHAB. But now I wanted more trending data and it looks like the emonPI has much more maturity and capabilities and so this will save me alot of time working all that out.

Cheers!

[Robert.Wall]

Bill is a big fan of Grafana - I expect he’ll comment, and tell you how to use it on your emonPi.

[Faze (Dale)]

Hi @Robert.Wall
Finally got my emonPI in the mail.
I didnt have the resistors at the time so just set it up, as is, just to get a baseline of what the numbers would look like with an unmodified box. And I think that the emonPI is giving me some over the top numbers, which I can only assume is because I don’t have the additional load resistor installed. I’m pretty sure Im not using 12.5kw in such a small town house.

So Ive now got a bunch of new resistors and I’m about to install them internally but wanted to confirm a couple of things before I break out the soldering iron.

The CTs state in their product guide (link above) that they have an internal resistance of 42 ohms. quote " Internal resistance: 42Ω "
I believe the emonPI comes with small 22 ohm resistors already installed in parallel. I can see from the pic you uploaded above.
And you’re suggesting that I need to install an additional 22 ohm resistor onto the CT lines in parallel, one for each line.

Am I correct in thinking that in order to adjust the circuit to accommodate my CTs and power environment, we are looking at 42+22+22 = 68 ohms in parallel is where I need to get to?

cheers

[Robert.Wall]

@Faze

And I think that the emonPI is giving me some over the top numbers, which I can only assume is because I don’t have the additional load resistor installed. I’m pretty sure Im not using 12.5kw in such a small town house.

That’s exactly right, and as you should have expected. Two times too much, to be precise. And your maximum current will show as 100 A, but will be only 50 A.

@Faze

The CTs state in their product guide (link above) that they have an internal resistance of 42 ohms. quote " Internal resistance: 42Ω "

That’s irrelevant for the calculations, but useful to know when you’re testing for a fault.

@Faze

Am I correct in thinking that in order to adjust the circuit to accommodate my CTs and power environment, we are looking at 42+22+22 = 68 ohms in parallel is where I need to get to?

No. That’s wrong on two counts.
First, the 42 Ω is irrelevant.
Second, resistors in series add up like that, resistors in parallel add as the reciprocal of the sum of the reciprocals, so 1 ÷ ( (1÷22) + (1÷22) ) = 11 Ω.

Those resistors are ½ W, so they are probably too big to fit between the holes in the PCB, you’ll need to either stand them on end, or (safer - you can unplug the c.t. without risking damaging high voltages) connect them across the white and black wires from the c.t. where you’ve extended them to the jack plug.

It’s safer because a c.t. is a current source. It will try to generate whatever voltage is necessary to drive the current it wants to into the load (or burden). You can always safely short-circuit a c.t., and it’s actually then working under no load. If you leave it open-circuit, it will generate a high voltage trying to force a current that can’t flow, and may well damage the insulation inside and destroy itself. If it doesn’t damage itself, it will probably need demagnetising to restore accuracy.

Homework question: With the resistor fitted and plugged in, what resistance do you expect to see measured between the black and white wires? What about when it’s unplugged from the emonPi? (The correct answer depends on where you add the resistor - in the emonPi or in the lead.)

[Faze (Dale)]

Ok bit of an update.
I was able to get the resistors soldered in ok.
I had to do as you suggested and put them end up.
Looks like a hack job but seems to have worked.

Software wise, I think I have it working correctly, I’m just not sure how I can verify that I have set it up correctly.
I have outputted the information to my openhab display in the kitchen. Looks somewhat ok,… wish I could get the cost graph to be the same colour as the other graphs.

This issue I have now is trying to verify the information its outputting is accurate.
Both the watts Im using and the way it calculates the energy costs. I’m definitely not paying as much as it states in the pic.

[Robert.Wall]

@Faze

This issue I have now is trying to verify the information its outputting is accurate.

What measuring instruments do you have? A multimeter, especially one that can measure up to 10 A a.c. will be very useful.

And are you running emonCMS? If you have the latest version, I think you should have a “Serial Monitor” option on the “Admin” page, and you can use that to adjust the calibration of the emonPi itself. (My emonPi is at least one version ahead of most people, so I’m not sure if you will have that.)

The emonPi is in essence a 2-channel version of the emonTx, so everything written about calibration for that (in the ‘Learn’ section here) applies more or less exactly to the emonPi.

Your first step is to verify the voltage. Measure your mains with a multimeter and check that you get the same value on the inputs page of emonCMS.

Do you have a resistive load (a fan-less room heater or a simple electric kettle, for example) that you can measure the current of? (You’ll also need to put your meter lead through the c.t. so that you can see what the emonPi is reading.)

Let me know what you are able to do, then I’ll tell you what to measure and what to adjust; because otherwise there are too many variables to consider.