IIRC the smallest setting is 1 joule, but I don’t know if that works in practice (@Robert.Wall might be able to confirm or expand on that) as I said above, the default setting is 1Wh, the aim is to match the granularity to the import meter’s “energy bucket” to swing to and forth within the metered granularity so as not to trigger a count.
You could go as low as ~£20 for the component parts for a mk2pvdiverter (load switching only) and drive that from the emonpi with some mods, ie new diverting firmware (derived from a MartinR’s “PLL” or Robin Emley’s “emonTx v3.4 as MK2PVdiverter” stretches) and physical access to a GPIO pin within the emonPi case.
Many years ago . . I set up the mk2 in a box by the heater “just to test” and simply ran a telephone wire back to the emonTx in the garage, this took it’s mains from a socket beside the heater and the heater just plugged into the diverter. Many, many moons later, it is still in that “just to test” setup. It has worked so well and the phone wire is so unobtrusive I’ve had no reason to prioritise moving it, although it is still on my do to list.
It is my intention to move the diverter to the garage and run a 13a spur from it to the heater location (I am able to run hidden mains cable, it’s just a bit of a chore to do so) . When I do do this, I intend to do away with the “other” diverter I had installed to the dhw prior to learning about the mk2pvdiverter as the mk2 is so much more efficient, it even mops up after the DHW diverter has taken what it can. I will also run another “diverter spur” outside so that when i eventually hack out hot tub heater circuit, I can have a second “diverter” supply to the hot tub to heat using only surplus PV in the summer months with just a changeover relay/contactor on the diverter output.
This should be possible, some of robin’s diverters have an RF capability, but with any RF/Wifi there is the issue of network speed and connections etc that could lower the effective resolution and/or leave the load on or off if there,s a connection issue etc, much of this can be catered for in software, depends how dirty you want get your hands. A receiver at the diverter end also obviously adds to the cost, but not by a lot.
I think the hardest part of the project would be adapting the emonPi firmware, there are examples you can draw from, but there is not currently any known diverting firmwares for the emonpi, a shame really as I feel there would be significant interest in one if it existed. IMO if you cracked that the rest would be pretty simple, cheap and very efficient.