I’m looking to expand my residential electrical monitoring to include some complete branch circuits, specifically 2 240V circuits dedicated to HVAC systems. I plan to do this with a Shelly EM and have been reading some docs/forums about it, but a couple questions remain.
This is a US power system, so it’s split-phase with 2 120V lines 180 degrees out of phase with each other (L1 and L2) along with a common neutral (N). The 240V branch circuits to be monitored are 2-wire circuits that just tie into L1 and L2 (no neutral connection), so the current on each wire should be equal. As a result, I plan to monitor only one leg of each. Here’s how I think that shakes out (blue are components to be added):
As you can see though, I’m not sure how to handle the N connection on the Shelly EM. It could go to the common neutral bar in the panel (120V power to Shelly) or it could go to L2 (240V power to Shelly). They Shelly specs say it can handle either voltage but the correct wiring may depend on whether-or-not those L/N terminals also act as voltage and phase references for the power calculations the Shelly makes. That’s where I’m seeing lots of conflicting info:
This discussion implies that the Shelly should always use 120V independent of the circuits being monitored, and that it will detect voltage directly via the CT clamps. However, it also notes that Shelly will still detect incorrect power values which have to be adjusted for in HA.
This discussion (and reference from Shelly support) seems to imply that the voltage and phase details are in fact determined by the L terminal, presumably relative to the N terminal, on the Shelly. I’ve seen a couple other similar discussions, but not conclusive.
So I was hoping someone may know if those L/N terminals on the Shelly are doing more than just supplying power and if they are also involved in phase/voltage references that have impacts on the power calculations. That would inform correct wiring and also what data adjustments may be required on the HA-side (if any).
L provides voltage and phase for measurement using both clamps.
Ok, thanks, then this implies that L is indeed used for more than just device operational power. For phase I don’t see any issue since I’ll ensure that the Shelly L connection is tied to L1 in the panel (also the same phase direction as the wires CT clamps are connected to). However, for a voltage calculation the Shelly must be looking at the difference between L and N. So, if connecting N to the neutral bar this will be 120V, which is not correct for monitoring a 240V line. I assume it would work but power calculations might then be off by a factor of 2. Would connecting the Shelly N to L2 solve for this, or is that just introducing other computation problems?
I’m not an expert, so remember that, this is my opinion from the docs…
That has CT ports that measure the load current and with that I would think phase… Could be wrong.
I would expect L and N to be the power supply for the thing’s internals only. Likely that is an internal power supply and the specs state it works with international voltages, meaning it would have to be isolated and separate. And according to the specs Shelly EM can use:
Power supply AC 110-240 V, 50/60Hz
That means it should be able to power L & N to L1 and L2
or L & N to L1(or L2) and the white neutral wire which likely comes in from the vault.
Didn’t think of that. So in that case I would treat it like EU and L & N goes to L1 & L2 to measure all the voltage and any device internals that power to ground are treated as noise, I guess.
But to be safe, ask the shelly people directly. (as maxym said)
Thanks all. My Shelly EM is still on-order so I’ll need to report back more when I have it and can test more. I believe I ordered a gen 1 Shelly EM (the only one I think is available in the US) which unfortunately does not seem to have a lot of documentation about how the L and N terminals must relate to circuits being monitored. However, I do see that there’s also a gen 3 Shelly EM that shows some additional wiring details printed directly on it:
This seems to suggest device L and N should be matched to the circuits being monitored. For a US 240V with no neutral connection I think N would have to be L2.
I’ll also see if there is a way to get more info from Shelly directly on this.
