Archived Forum

Probably a combo of looping the AC line through the CT multiple times to scale up the current detected and replacing the burden resistors to scale down the range of the CT channels to increase the sensitivity of genuine readings without amplifying the digital noise.

On the emonTx the only difference between the "hi-sensitivity" channel and the other 3 (or the emonPi') is the burden resistor value, which isn't too difficult to change.

I think you could accurately measure low power ranges on either the emonPi or emonTx the choice is yours, the deciding factors would be 

how many channels do you want, 2 or 4?

How are you connecting to the network? wifi, mobile, tethering etc

Do you want a local emoncms instance or just remote?

How power efficient are you wanting this to be?

The emonPi can quite heavy on energy consumption if you just want to send data to emoncms.org, whilst the emontx is wasting some power using RF to send data to a device located in the same place, but you do get 4 channels. Plus the emonPi doesn't buffer data for delivery to emoncms.org so if the network is not available the data destined for emoncms.org is dropped. Therefore you will have to keep a local server if the network is intermittent.

Perhaps a Pi zero or A+ with a "serial-direct" connected emonTx is the most energy efficient option as the Pi B's ethernet and usb ports add considerably to the power consumption but add no value if the Pi is just being used to forward data via wifi.

If you are concerned about making the burden resistor changes, perhaps the shop would do it prior to dispatch for a small fee. Worth asking, as they are very helpful.

Paul

The best solution for small currents is to use a more appropriate current transformer. It's easy to say that, somewhat harder to find one, as the majority are made to give a 0.333 V output, which seriously degrades the performance of the emonTx as that is looking for 1.1 V rms at maximum current.

Robin Emley uses one of these two inside his Mk2PVRouter, but even those are designed for a load significantly bigger than yours.
The 20 A version here: http://www.yhdc.com/en/product/425/ or this 20A:10mA one http://eppep.com/?action-viewnews-itemid-140364 from http://www.ebay.co.uk/itm/281436762683?_trksid=p2059210.m2749.l2649&ssPa...

Both will need the burden resistor changing, but both will be a better starting point than the shop CT, especially if you use the YHDC 10 A version and a multi-turn primary winding (that's the wire on which the CT sits). 500 W is 2 A in round numbers, so 5 turns and a 20 A CT will effectively make it a 4 A CT, or 4 turns and a 10 A CT will give you a 2.5 A CT, and that's looking reasonable. Your normal load is then about 3 - 12% at worst of full load (~ 1 kW), and accuracy should be reasonable (but not outstanding!) at your lowest 'normal' power level.

If you're worried about power drawn by the monitoring system, consider adding a display to the emonTx. I believe it's been done, but I can't recollect when or by whom.

Here's one example of an LC display added to the emonTx:

http://openenergymonitor.org/emon/node/2550

I agree about the "closed ring" CT's I prefer to use them when ever I can, usually only the main incomers (and any "absolutly only qualified person circuits") get the "clip-on"  CT's. I hadn't mentioned them as they aren't "ready to fit", don't have the zener protection and require breaking the mains circuit to install, (the question re how difficult was it to replace the burdens didn't suggest a fluency in electronics/electrics, but that may very well not be the case)

The question was originally about whether the emonPi was suitable or if the emonTx "high sensitivity" input was better, I was just pointing out the CT sensitivity was not really a deciding factor between them as either could do the job equally well if the circuits/ct's were adapted to suit. and to that I added some other considerations that were possibly more important when choosing the device.

I don't know that power consumption is a priority but IMO adding a display actually adds to the power consumption if the data is wanted to be logged rather than just displayed in realtime. but this is drifting somewhat from the OP's CT sensitivity questions.

Paul

The emonTx and the emonPi input circuits are to all intents identical, so exactly the same considerations apply to both. As the enquirer lives in a "trailer", then presumably the rules for a fixed installation don't apply, though other rules for mobile installations might. My guess is the main incoming supply will be via a 16 A connector, thus it will be very easy to isolate and break into the circuit to install a ring-core CT. That's why I thought one of those should be considered.

We can take it further and look at the details when Petr comes back to us.

Thank you very much for all the thoughts I really appreciate it a lot. :)

- Yes Robert is right electricity supply is via standard 16A extension cord and i actually broke it already to put cable inside through a smaller hole and then I reinstalled the plug from inside. So to install ring-core CT is not a problem.

- one channel is enough for now and later when I add the photovoltaic system I would need two, i do not need more.

- also the power efficiency is not critical now, but later when I move out of the grid and use just solar it will be more important

- it would sit right beside the router connected to the internet, so idea was to connect it via ethernet cable, wifi is of course an option too.

-  logging is the important part, local is enough, remote would be great. Now I use one of those plug-in energy meters so I know the actual real time data, but it is complicated to evaluate them in the longer run, that's why i am looking for a solution.

- I do have some basic electricity knowledge, but it is more practical experiences with operating couple of home made photovoltaic systems and I am not very good at the theory behind it. That is why I do have troubles understand what do I need to do :)

I am now little bit confused about when does the burden resistor comes in and when I need to change it. I thought that together with software calibration it is the difference between normal and high sensitivity CT input on emonTx.  But if they are the same I do not get it. Or is it part of the CT? so if I get one of the Ct's (preferably yhdc 10a) Robert Wall is suggesting I will just adjust the software numbers and it is good to go?

Can emontx itself log the data locally? 

Thanks for help again and sorry for silly questions 

Petr

The easy one first: No, the emonTx cannot log the data inside itself. You need to connect it either serially or by radio to something else. And everything below applies to the emonTx too (except you need a programmer to reload the sketch).

Every current transformer needs a burden resistor. A current transformer outputs a current that's proportional to the current you're trying to measure. You generate a voltage by passing the current through a burden resistor. Some CTs have their burden resistor inside the case. We believe a YHDC SCT-013-010 is a SCT-013-000 with an internal burden resistor chosen to give you 1 V @ 10 A.  The emonPi requires a maximum of about 1.1 V across its internal burden resistor. We chose the burden resistor to give that voltage (approximately) at 100 A with the YHDC SCT-013-000.

Your most versatile and most accurate solution is to find a small ring-core CT (because these are inherently more accurate than split-core ones) like the ones I mentioned in the 3rd post here, remove the burden resistor in the emonPi and replace it with a higher valued one which you calculate will give 1.1 V at the maximum current that you expect. Then you'll need to edit the sketch to change the calibration, compile it and load it into the emonPi. The instructions for that are in the emonPi Wiki. If you do not want to measure power closer than whole watts, you could do the calibration inside your emonPi - but it only sends the power in integers to emoncms, so precision is limited.

If you use the SCT-013-010, that version has a burden resistor inside, and you will need to remove the burden resistor that is fitted inside your emonPi. If you don't, it will read to 110 A, which is not what you want. And you will still need to change the sketch.