Electromagnetic harvester feasibility question

[webandrew]

I have a device idea, but my limit is at putting a wee arduino device here and there  . I have an idea in my head that's just beyond my technical knowledge. So I'm wondering if its feasible or am I just embarasing myself. Here's the gist:

A 3.3V (I think 50-170mA depending on usage) ESP-01 running off electromagnetic power harvesting from a cable delivering power to a water heater. So a 230V and quite a few Amps. Presumably more than a tea kettle.

The heater turns on, the ESP turns on. ESP could transmit "I'm on" to some IoT device when attached to a cable. Rest is up to imagination!

Question is, is a harverster to power the ESP in real time possible?

So... I have a feeling its a non-feasible idea. How bad is it? Thanks in advance.

Context for the idea if anyone cares: Where I live rentals are common (So no tinkering with the installation) and everyone forgets to turn off the heater. No plugs available. And maybe there's a neat solution with a battery, but in my head a standalone device that you clip onto a power cable and forget about seems elegant.

[daqq]

Hmm, I don't think that you can you it with a typical non modified cable with the current running both ways. The power requirements are just too large for any kind of capacitive coupling.

You can actually "harvest" some pretty large amounts of energy using current transformers (magnetic coupling), but you need to be able to hook it around a single line (L or N, not both at the same time).

[webandrew]

Ah, knew there would be a trick to it! Too bad. Thank you for your anwser! Back to tinkering with a battery and deep sleep option.... Thanks again!

[daqq]

Well, maybe you could harvest energy all the time from the ambient source, charge a battery or a super cap, and then, when active you discharge the battery. The average charging power can be very small, during the active period, you can draw large amounts of power.

See: http://www.linear.com/product/LTC3331 and its friends. Just add a small solar cell.

Mind you, it might get costly.

[Twistedsnail]

This isn't an answer to your questions, but here is the first article by jeelabs on similar idea:
https://jeelabs.org/2015/02/18/micro-power-snitch/

It's a really excellent 8-part write-up of a device that steals power from an AC line until it has enough juice to send a wireless packet.
Worth the read.
[edit] It's a success by the way, the MPS triggers on any appliance drawing ? 500W (on 230 VAC, or 250W for 115 VAC)

[webandrew]

Thank you both for those ideas. I'll definitely look into both suggestions, and I'm sure I'll learn a ton just by reading up on these. Hopefully I'll manage to put something interesting together at the end of this!

Combining battery with trickle charge from ambient might be the way to go!

Thank you again for your time!

[Ian.M]

If you install your device at the consumer unit rather than at the water heater, you can draw power from a current transformer on a split core clipped round one of the meter tails.   You should be able to sense the load circuit current with a hall sensor in contact with the flat cable to the heater, sensing the loop field between the L and N wires - it will probably work best if you cut a gap in a thin large diameter ferrite toroid large enough for the sensor + thickness of the cable.