Using the EmonPi for larger loads

jim.o · 2020-12-09T08:21:24.888Z

Hi

I’ve got half-hourly consumption data for a building indicating average 3-phase power up to 55kWh per half hour. That’s an average of 150A per phase for the half-hour, but I’m not sure about peak demand. I want to monitor power in real-time.

It’s likely that the 200A CT for an EmonPi is not going to be sufficient. Is that right? Is there an alternative - can I split the current, source a new CT and/or source a power or electronic device to convert the sensing current to an appropriate range? Has anyone done this?

Thanks!

Jim

Robert.Wall · 2020-12-09T13:59:32.708Z

Welcome, Jim, to the OEM forum.

It’s common for users in the USA to use a larger c.t., partly because of their reduced voltage but mainly because of aluminium “service entrance wires” (incomers to you and me). If you look on the “Using the emonTx in N.America” page in Learn, Learn→Electricity Monitoring→AC Power Theory→Use in North America→Use in North America there’s a list of (mainly American) c.t’s that you can use.

But - if you have a 3-phase supply, the emonPi won’t work - unless you have a guaranteed balanced load and only need to measure one phase. If there’s a significant imbalance, you need the emonTx with the 3-phase PLL sketch loaded - then an emonBase (or any server that’ll run Python and emonCMS - or another database) to accept and store the data. Note the ‘front end’ power sensing departments of the emonPi and emonTx are identical, except for the number of channels.

You can use almost any c.t. with the emonTx/emonPi, if you’re prepared to change the burden that’s on the p.c.b. inside, and provided that it has a sufficient VA rating to give about 1.1 V across its burden at your maximum current.

You obviously are in a better position than I am to know what your peak load might be, but if the average half-hour current is about 150 A per phase, you may well be pushing a 200 A c.t. beyond its rated maximum. I think I’d be looking at a 300 A or a 400 A c.t. But I stress that I don’t know how your demand changes from one instant to the next.

This is obviously not a “standard” problem, so feel free to come back with any further questions.
(My background is drives and controls, particularly d.c., but I’m retired now.)

jim.o · 2020-12-11T12:14:31.855Z

Thanks Robert.
I would be prepared to assume a balanced supply if it simpler to use an emonPi over emonTx.
You’re right - only I can assess the peak load!
I was hoping to be able to monitor in excess of 200A without “opening up” the device. Does changing the burden require soldering? Sorry - I can’t find the documentation by simply searching for “burden”!

Is an interposing transformer a solution as per your reply in:

Current Transformer 300A/5A interfacing with Arduino Hardware

I take it you are using the emonTx Shield? And your Arduino has an analogue input range of 0 - 5 V? There are two possible approaches: You need a burden resistor that will give you approx 1.6 V rms at the maximum current that you want to measure. Presumably that will be 300 A. Therefore, the value of that resistance will be 0.32 Ω. The power rating will be 8 W. You will almost certainly need to approach a manufacturer of industrial instruments to have a meter shunt made to that specification…

I think the responder then thought of something akin to another solution I was thinking of - but not sure what it was called!

Robert.Wall · 2020-12-11T13:38:01.872Z

jim.o:

if it simpler to use an emonPi over emonTx

Yes it will be - you have one box only. Apart from that, there’s not a lot in it. If you really will benefit from knowing the loads on each phase, then an emonTx and an emonBase - there are several ways of doing it - is what you need. It’s for you to weigh up what you need to know vs the complexity - perceived or real.

jim.o:

Does changing the burden require soldering?

Yes. (I wouldn’t trust one on a header, or something like that.)

But if you choose a c.t. with a 50 mA secondary - or more precisely, 50 mA at your maximum current, then no change to the burden is necessary. An interposing c.t. will add to the errors and cost of course, so it’s less than ideal, because if the primary c.t. is a 5 A one, then you need a 5 A:50 mA c.t to get into our analogue front end - or a larger one with a multi-turn primary winding.

Or, if the burden you need is less than 22 Ω, you could have a parallel resistor external to the emonPi/emonTx.

jim.o:

I think the responder then thought of something akin to another solution I was thinking of

4-20 mA is going to struggle with getting phase information.
(This is how the emonTx/emonPi front end works: the instantaneous voltage and current samples, taken at about 50 per mains cycle, are multiplied together and averaged to get the real power.)