Archived Forum

A lot of fairly deep questions there. I'm in the UK, Bill Thomson might want to chip in with some local knowledge.

Zeroth question: DON'T get the 0.333 V output CTs. The voltage that the emonTx/Pi is looking for - whether it comes from a burden inside the CT or one inside itself - is about 1.1 V rms, so you're only using ⅓ of the available range. Get 1 V output or, if you want to completely customise the thing, get the burdenless version and fit a burden resistor of your choice to give you the correct scale for your situation. (Either way, you'll need to work inside the device. More on that below.)

First question: An emonPi is an emonPi. I'm not aware of any hardware difference, it's the complete kit of parts that makes the difference. You should only have one Pi, you can add five emonTx's sending data to it by radio.

Measuring the 240 V circuits:
As I tried to explain in the US page, for a true 240 V device (load or feed - the principle is the same) the current is the same in both legs. If the voltage is the same, you can measure half the voltage and the current and double the answer. In your case, the voltage on the two halves is not the same, so you must either measure the two half voltages and the current, then add the voltages and multiply the sum and the current to get the total power, or you must measure and multiply the two halves separately and add to get the total power. The snag is, to get real power, all that has to be done in one unit, and OEM doesn't offer anything with 2 voltage inputs. So you need to treat the 240 V devices as 2 × 110 V that just happen to share a 2-pole breaker.

Is a 100 A CT the right one for all the individual circuits? You will probably find it beneficial to increase the sensitivity by increasing the burden value, alternatively as you're clearly competent to drop wires out of breakers, you could consider cheaper and smaller (and probably more accurate) ring-core CTs for the individual circuits. But that would mean changing the burden resistor inside the emonTx.

Second question: As I hinted above, to measure real power, both voltage and current need to be measured in the same unit. So it's not as simple as the totals need to add up. You must allocate inputs and circuits so that the voltage and currents are all on the same leg. That inevitably means that you'll have the contributions for the 240 V devices in two parts, which you'll need to add in the feed in emonCMS.

Third and final question: Yes. Both the Pi and the emonTx are designed for a current input of 50 mA for full scale. For the best performance with anything different, or if you get 'voltage output' CTs, you will need to do a little customisation inside the devices. It's not hard if you have experience of working on circuit boards. If you haven't, then I'd suggest enlisting the aid of someone who's happy to do that for you.

Fourth  and final question question (FTFY and anyway, experience tells me that's a complete fallacy - there will be more!):
You're missing a programmer and lead. You'll need to edit and reload the sketches in the emonTx's to give them unique identities. And a power supply for your Pi. But you could economise on parts: All the emonTx's and the Pi could use two ac adapters for measuring the voltages (one per leg) IF you provide the emonTx's with a 5 V d.c. supply - and that could be a single one supplying all five and the Pi. So instead of 6 a.c. adapters and one 5 V d.c. USB adapter for the Pi, you could have two a.c. plus the one 5 V d.c. adapter. You'll need a handful of plugs and leads if you do that.
And you'll need RJ45 connectors etc for your water meter.

Robert - 

Thank you for your reply. And you are correct, your answers have now generated additional questions :-) Please bear with me, a lot of this stuff is new to me and I want to make sure I understand it all before proceeding. 

On my 240V main incoming feed, the voltage on each leg is the same, it is just the current that is unbalanced just due to the way the 120V circuits are distributed across each leg. This only applies to my main incoming feed, not any branch 240V circuits which (I assume) should (by their nature) be balanced. 

So to measure just my MAIN incoming feed, I would need two CTs, one for each hot leg since they are not balanced as far as current is concerned. I would also need to measure voltage across one of the legs since the voltage is the same on both legs. Is that correct or do I need to measure voltage across both of the legs? I noticed that you said I needed two AC adaptors (one per leg) and was just wondering if the voltages on each leg are the same, why do we need to measure both of them?

I also understand with the solar monitoring I think I may need to use another voltage monitor on the solar input as a way to measure the actual current flow. Is this correct?

You mentioned ring-core CTs. I am very adept at pulling circuits from my main panel so if the ring-core CTs are more accurate, I would like to go with those, but I have several questions:

1) You mention having to change the burden resistor in the emonTx units. I am not new to soldering, but I am not super good at it. How difficult is it to change this resistor and how many do I need to change per unit? Are there 4 burden resistors inside each unit since they will handle 4 CTs or do they share a single resistor? How would I know what value resistor to install in it's place?

2) Considering that I have a range of circuits from 15A to 50A, 120 and 240V, do I get 100A ring-core CTs for each circuit or different CT values for the 15A circuits as opposed to the 50A circuits?

3) I have 4 x 240V loads on my panel. My Air Conditioner, range, pool and clothes dryer) but all of those loads are balanced, so I assume that I need one CT per each of those circuits and then I just double the value. Is that correct?

4) Do you have any recommendation as to the size of the CTs I would need as well as a specific manufacturer that you know woks well with the system?

Thank you again for your help!

Hi rjsears

I have a somewhat similar service here in the USA --- 200 Amp Service and a 11,400 watt Grid Tie Inverter.

For the 2 Main feeders, I recommend you take a look at the Wattcore WC3 Split-core 100A:0.1A  (1,000 Turn Ratio) .

It will easily fit around the 4/0 feeders and handle 100 Amps at 240 V or 24,000 Watt loads which should be sufficient unless you have some really high loads all on at the same time.

This CT does not have an internal burden resistor, but does have a safety feature of  Zener Diode clamping  at 6.5V RMS or 18.4V Pk-Pk. --- sells for about $39.

http://byramlabs.com/store/product_info.php/products_id/20516/product/WattcoreWC3%20SERIES%20WC3-100-MA100%20Split-Core%20Current%20Transformer%20(100:100mA)

To measure a full 100 Amp load you will need a R (Burden Total) = 1.1 V / 0.1 A = 11.0 Ohms.  Use a 22 Ohm in parallel with the 22 Ohm Burden Resistor on the emonTx V3 circuit board = 11.0 Ohm and you won’t have to remove that resistor.

Try a Calibration Coefficient of Turns Ratio (1,000) / Burden Resistor (11.0 Ohms) = 90.9

If you are going to measure with a CT every individual breaker off the service panel, I do question why you also need to also measure the 2 Main feeders, because the Total of the individual circuits will be the same as what the feeders measure.

Also, you will only need a CT on one of the hot legs from the Solar Inverter. All 240 V Grid Tie Inverters use the Neutral wire only for sensing (doesn’t carry current), so the 2 hot legs will be balanced and only need to measure one.

If your Solar Inverter max output is less than 50 Amps X 240 V = 12,000 Watts --- this CT should work:

Wattcore WC1 Split-core 50A:0.1A  (500 Turn Ratio)  $25  Has Zener Diode clamping  at 6.5V RMS or 18.4V Pk-Pk

​R (Burden Total) = 1.1 V / 0.1 A = 11 Ohms  Again, use a 22 Ohm in parallel with the 22 Ohm on circuit board.

Calibration Coefficient = Turns Ratio (500) / Burden Resistor (11 Ohms) = 45.5

Hi SD Solar - 

Thanks for the info. I wanted to measure the mains just as a backup to the other circuits and the cost to monitor two additional legs is not that much.

Since I only need to monitor one leg on the inverter, I assume that this is the same on my other 240V circuits such as my AC, electric range, etc. Those circuit should be balanced. I am going to pull my electrical cover this afternoon and do some testing to verify that that is the case. 

Then I am assuming that when I dump the info to emoncms, I just multiply what I get on the current side * 2 and that would be accurate. Am I thinking correctly?

I am new to all of this so I assume that the Calibration coefficient is something that I will need to know when I program the units themselves to read the CTs or is this something that needs to be done my emoncms once I get the data to it?

Thanks for the info!

It's a given that the currents in each leg - and the same goes for a proper 3-phase supply - will not normally be balanced. So when you made a point of saying that the feeds were not balanced, that meant to me that the voltages were not balanced. 

As they are, you need to revisit the "Use in N. America" page because almost everything is explained there.

Hi rjsears, I have a rather similar setup (only 16 CTs, no solar), but same US split phase. I did everything with the "normal" SCT-013-000 split-core CTs. They (just) went round my 100A supply lines (one each - mine are nowhere close to balanced either). For the 240V circuits, I took the wires out of both sides of the breaker through the same CT - but in opposite directions so the currents add (there's probably a good principle here, but I just used trial & error to get them adding properly). Not sure this is completely kosher, but numbers I get are reasonable. Remember to flip the DIP switch on the EmonTxs for US voltage. Also need to set it in the EmonCMS configuration.

I did order the programmer, but haven't needed to use it yet! I just asked Glyn & Tristan at the store to ship the EmonTxs with different IDs, and they were great, no problems. But there is a new monitoring sketch "in the works", when that's stable I suspect I will be reprogramming.

Thanks, Sandy

Thanks Robert - 

Actually, only my main feed is not balanced. All of my other 240v circuits are evenly balanced. Sandy mentioned that he took both of the leads of his 240V circuits through the same CT except in opposite directions. This seems like a great concept to same some $$ on some CTs and seems to make sense. 

What do you think of this idea? I noticed it in your For use in the US document...

Sandy - 

Thanks for the great info, I never thought of running both of the wires through the same CT only with one going the opposite direction. I wonder why more people are not doing it (or maybe they are). I asked Robert (same thread) for him input.

When you received your units did you have to add/change/remove resistors with the SCT-013-ooo CTs you were using?

Thanks for the info!

I am assuming that his is the configuration that you are referring to:

Hi Rich,

A "load audit" will give you an idea of which CT to use.

Add up a rough total of all the large energy consumers you have.
Make an estimate of the max number of them you'd ever have powered on at the same time.
Use a CT with a rating that meets or exceeds your "load total, plus a bit for headroom." (15 to 20%)

More info in this thread, and here. <-- Info on 200 Amp service

Wally, Thanks for helping out!

Regards,

Bill

I never thought of running both of the wires through the same CT only with one going the opposite direction. I wonder why more people are not doing it

Unless the wires are smaller than say, AWG10, it's nigh on impossible to get two of them through the CT in opposite directions. Definitely not going to happen with 2/0 or 4/0 feeders! It might be done with AWG 8 or
AWG 6 in the case of branch circuits for larger 240V appliances, but that would require some slack in the wiring.

The Electrician doing the wiring is supposed to leave just enough slack in the wiring to enable connections to circuit breakers, bus bars, etc, but not more. Depending on the person who did the wiring, there may, or may not, be enough slack in the wiring to enable putting two wires through one CT, in opposite directions. If the job was done IAW the NEC, there won't be much, if any, slack in the wiring.

Bill - thanks for the insight.

In my panel, it would appear that I might have enough slack to run the 240V house circuits throught a single CT, the lowest right hand 2-pole breaker is my inverter feed to the  panel and it looks like I have enough there, but CTs are cheap for sure and all I would save is some math and an extra EmonTx!

Considering that I wrote that page, with Bill's help....

Looks like you've got plenty of slack (and room too!) in your load center. ;-)

With that much room, you won't have any problem using two CTs.

My load center is a lot tighter than yours, and I've managed to put four Wattcore WC-1s and
two Magnelab SCT-0750s in it!

Robert - 

Will all the great advice you give and the knowledge you have I suspected you wrote that page! Thank you for your patience in helping us newbies get going and understanding all of this electrical stuff! 

Thanks Bill - 

I might try with a single CT per 240V circuit (except the mains) as it will cut down on the number of EmonTXs I will need. I ordered enough of everything to use a single CT on each hot leg so we will see how it all plays out when everything arrives in a couple of weeks!

Answering the questions elsewhere about CT cable length and multiple transmitters in one enclosure...

You should be OK up to 10 m, and probably more, for the CT leads, if you do it properly: use good quality twin twisted screened 'microphone' cable, using the twisted pair for the CT secondary, and earth the screen only at the CT end. And don't run a heavy power cable alongside that carries interference, such as big spikes from switching fluorescent lamps or motors.

5 transmitters inside one box might be pushing your luck, do you feel up to arranging a serial connection and using one emonTx as the master, which will command the others to send their data when requested, rather than when they feel like it (which is what happens with the radio, and can result in collisions)? Martin Roberts did that with 3 emonTx V2's in his "Full-fat 3-phase" design. You should be able to follow that for the serial software. A serial connection from one emonTx to a Pi is documented, but I don't know that anyone else has done exactly what I'm envisaging, however I think it should be possible.

If those 240 V loads don't have a neutral connection, or they do have but it only carries an insignificant current, then of course you only need one CT, and as your voltage is balanced, all you do is double the number you get. It's all on the "Use in N.America" page.

Hi Robert, rjsears, Yes, the concept of passing both legs through the same CT in opposite directions came right from that North America usage post (Robert, sorry, I didn't remember seeing you & Bill as the authors - if I had, I wouldn't have had any worries about doing it!). I've used the same concept to combine several low usage 120V circuits into 1 CT. (The house has rather a lot of branch circuits - and even more light switches!)

I didn't have too much trouble getting both legs through the one CT. I think the biggest I did it for was the oven - 50A on 6 ga. As Robert says, not a hope with the supply lines.

As I remember (from another of Robert's posts) the SCT-013-000 has no internal resistor - so works perfectly with EmonTx that has the burden resistors on the board. There are other variations - with something other than -000 in the part number - that do have internal burdens.

I have 4 EmonTx's (nailed to the wall, in a line right above the breaker box) all sharing a power supply (so all turned on at exactly the same time) and I don't seem to have many conflicts with the transmissions overlapping (looking at EmonCMS, the ages on the feeds rarely go much over 30 seconds - maybe miss 1 every 5 mins). I've also got 3 EmonTH's in the mix, but they transmit less frequently.

The only non-standard thing I've done is in EmonCMS, I added a 'stacked' option to multigraph so I could see how usage breaks down between the circuits. (See https://openenergymonitor.org/emon/node/10919).

Robert - 

I will look at the "Full-fat 3-phase" design and take a look at the serial setup. Sounds like it might be a better way to go since I will have 5 of the EmonTX3 and the Pi. I might get it all set up and see if I am having issues and then move to the serial route.

My thought on the CT length was to move my equipment box out of the outdoors (where it would sit in the sun all day) to the attic, but I did some temp reading there yesterday as well, and it was 45C, not a good place for the units.

So I talked to my better half last night and she is ok with me coring through our back wall which is directly opposite our power panel, drop a nice 2" conduit in through the wall and have all of the hardware sitting inside the house in a nice cabinet in our living room. That way, all of the hardware is inside in cooler temps and my run from the equipment to the power panel is less than 2M. 

I am also going to run my reference power through the same conduit using 16/3 extension cord wire. As I recall, in my setup I need two reference voltages - one for the main panel and one for the Solar array (to measure import or export of power).

I have not read anything about there being an issue with the CT leads being close to power lines, what do you think about running the CT leads inside the same conduit as my reference power?

Robert -

Another questions comes to mind rereading your posts above. You said:

As I hinted above, to measure real power, both voltage and current need to be measured in the same unit. 

Does this mean that I need a reference voltage adaptor for each of the EmonTX3 (all five of them) in order

to get real power from each unit?

Thanks

Rich,

If you want your installation to meet NEC guidelines, your low voltage wiring needs to be in a conduit separate from the 120/240 voltage wiring.

You might want to verify this, but I'm thinking the 16/3 extension cable wire you want to use doesn't meet code requirements. I'm not sure whether the requirement to use solid wire for low voltage circuits is the same as it is for high voltage wiring, but I'll find out a bit later today and pass along the answer.

Thought I'd mention it in case you weren't aware of those requirements. Keeps the inspectors happy, and isn't a bad idea looking at it from the insurance company's point of view.

Bill - 

Thank you for pointing that out, it makes sense that per code they would need to be separate. Same on the extension cable wire. I was hoping to bypass an intermediate outlet since the conduit is so short, but you are right, regardless it all needs to be to code.

Thanks!

OK, so giving this more thought, I am going to have upwards of 20 CTs in my panel, all coming through a conduit into the house (less than 4' away) and then into the Emon equipment. If I run all of the CTs from the panel directly to the Emon equipment it might look a bit weird so I was thinking of taking all of the CTs and terminating them inside a dual gang outlet box with a plastic cover.

I would install 3.5MM Female-to-female bulkhead connectors (as opposed to cutting off the 3.5mm jack on the CT and soldering to a 3.5mm female outlet) in the cover of the outlet and on the inside plug in the CTs then close the cover. From there I just get standard 3.5mm male-to-male cables and run them from the box to the emon equipment. 

I think this would look better and if I ever lose a CT, I would not have to desolder the CT, just disconnect the CT from the electrical circuit, open the box, unplug the bad CT and replace it with a new one.

My question is would I have any problem with my readings going from male plug on CT to female jack to female jack to male plug to emon equipment?

Again, if your voltages on the two legs are balanced, and if your emonTx's and Pi all share the same 5 V d.c. supply, they can all share the same a.c. voltage sample too. If you look at the front end circuitry, you'll see that it's OK to tie all the ac voltage inputs together if they all share the same ground, and that in turn means they can all share the same d.c. supply. Just make sure you've got enough current. With the links on the emonTx's removed so that the a.c. voltage isn't used to power it, the current drawn is about 0.1 mA. (You will probably need some -1 multipliers to get all the signs right in the power and VA departments.)

Ha - you posted while I was typing!

I'd avoid as many connectors as possible. It's easy to unsolder the plugs from the CTs, so I'd be inclined to do that and use terminal blocks (get ones small enough because those CT cables are quite fine) in the outside box, and re-use the plugs inside. It'll save having to have enough spare space in the conduit to get the plugs through the wall, apart from anything else, and it will facilitate earthing the cable screens (and those are not the plug body!).

[I think you want 5 terminals per 2 CTs: CT1+, CT1-, Scn, CT2+, CT2-, etc 10 times.]

Rich,

Turns out the 16/3 wire you mentioned is OK for the job. As long as all the LV stuff is kept separate from the HV stuff, i.e. in its own conduit, you're good to go.

Robert - 

I decided to use the following CTs:

2 x Wattcore Part #WC1-100-RV001 Split-Core Current Transformer (100:1000mV) (My Mains)
16 x Wattcore Part #WC1-025-RV001 Split-Core Current Transformer (25:1000mV) (120V 20A Circuits)
2 x Wattcore Part #WC1-050-RV001 Split-Core Current Transformer (50:1000mV) (50A A/C Circuit)

From my understanding (after reading about 200 of your posts), these should work, I just need to remove the internal burden resistors completely from my Pi and TXs. These have the added benefit of not having a connector on them, so I have to do all the soldering. I recall a post telling which wire goes to tip, I'll dig that up when they arrive and I need to start putting ends on them.

As far as powering the units, I am planning on purchasing a single, powered USB hub and plug all 6 of the units into this powered hub. I am going to add a dedicated breaker for the VT (I have the space in my panel) and I purchased the US AC adaptor (actually three of them) from the OEM store. Since each Pi/TX only used 0.1mA for voltage monitoring, I could use a single power supply and a splitter for powering all of the units, correct?

I  could also power my USB hub from that same breaker/outlet and both the USB hub and the AC adaptor would be sharing the same hot, neutral and ground. Would that be the best idea?

Am I thinking about this clearly?

Now referring to THIS document, I have a Type 2 connected solar array. I need to have an AC input in order to monitor the flow of the current direction from the solar array. I assume (maybe incorrectly) that the AC adaptor needs to be connected to the hot line of the solar array itself in order to detect this flow, or is this incorrect and the TX connected to the solar array can use the same reference voltage as the rest of the system? I just do not understand how an AC adaptor can tell the flow.

Thanks

Thanks Bill - 

I decided to just install normal outlets for my reference voltages as opposed to running the extension cord type wiring.

The ac adapter provides the reference so that the relative phase of the current can be measured. Read the first page in Resources > Building Blocks about AC Power Theory. It doesn't matter where you measure the voltage (within reason). If you get the CT wires crossed, you just flip it on the main (primary) conductor. But it would be nice to have them all the same.

" I could use a single power supply and a splitter for powering all of the units, correct?" That's EXACTLY what I wrote above: "if your emonTx's and Pi all share the same 5 V d.c. supply, they can all share the same a.c. voltage sample too."

Thanks Robert - 

I just want to make sure I understand correctly is all.

Thanks