Kirchhoff decides the layout, not the spare way

Where the inverter connects is usually decided by what is convenient — a spare way in the consumer unit, a short cable run, a tidy job. It should be decided by where the current goes, and that is a question with a definite answer.

Kirchhoff's current law: at any node in a circuit, the current flowing in equals the current flowing out. Nothing accumulates at a junction. It is the first thing anyone learns about circuits and the first thing forgotten on site, because on a load-only installation it never has to be thought about — there is one source, so every current in the building points the same way.

Add a second source inside the installation and it stops being trivial. The busbar in the consumer unit is a node. Whatever arrives at it adds up.

On a conventional installation, one device bounds every current in the board: the distributor's cut-out fuse. A 100 A fuse means no part of that assembly can ever see more than 100 A, whatever anyone plugs in, because there is nowhere else for current to come from.

That reasoning is load-bearing, it is usually unspoken, and it stops being true the moment there is a source downstream of the fuse.

A PV inverter and a battery inverter are both sources. Connect them into the consumer unit and part of that assembly can be carrying the supply current and the inverter current together — a total the cut-out fuse will never protect against, because from its point of view nothing unusual is happening.

BS 7671 has requirements about exactly this — on the rated current of the assembly and of the circuits within it — and they are the ones that decide the layout.

In substance, the wiring regulations require that the relevant design current does not exceed the rated current of the assembly — I_nA — or the rated current of a circuit within that assembly — I_nc — once any applicable diversity has been taken into account. Those ratings are declared by the manufacturer of the assembly; they are properties of the consumer unit, not things a designer gets to choose.

And the sting in it: overload protection is not to rest on the diversity of the downstream circuits alone. “The sum of the breakers is more than the board, but they will never all be on at once” is precisely the argument the regulation exists to rule out.

Paraphrased, not quoted — from the IET's published guidance and BEAMA's material on assembly rated currents, not from the standard itself. Read the requirement in the current edition of BS 7671 before you rely on it — this area has been redrafted by recent amendment, so check which edition applies to your work.

Alongside that, BS 7671 states something people find genuinely surprising: an RCCB or a switch offers no protection against overload at all.It will disconnect on an earth fault, and it will carry its rated current, but it does not watch for too much current the way a breaker does. It has to be protected by an overcurrent protective device, and the rating of that device comes from the RCCB manufacturer's instructions.

The everyday case is the 63 A RCCB in a split-load board, sitting behind a 100 A supply with more than 63 A of breakers on its bank. That is a live argument in the trade without any generation involved.

Add an inverter to that board and you have added a current path the original assessment never included. Whatever position you take on the 63 A question, adding a source is not a neutral change to it.

There are a few ways this gets answered. None of them is a move that makes the problem go away — each one relocates the exposure, and the job is knowing where it went.

1. Option 1

   Connect upstream of the consumer unit

   A separate switchfuse from the tails, or a Henley block. It is the most common answer, and it is routinely described as though it settles the question. It does not.

   It needs a 100 A fuse ahead of all the consumer units in the property — and that potentially leaves the Henley block itself exposed to too much current, because the block is now a node with two sources arriving at it.

   Worse, connecting the PV board upstream — nearer the supply than the main board — still increases the total supply available to the installation. The main board's loads can now draw from the grid and from the generation at the same time, so the main DB is still exposed to a current the cut-out fuse never permitted. The fuse only ever limited the grid's share of it.

2. Option 2

   Use an assembly rated for the sum

   Establish I_nA and I_ncfrom the manufacturer's documentation, and show that the total — supply plus generation — sits inside them. Legitimate, and it depends entirely on paperwork you have to actually obtain rather than assume.

3. Option 3

   Curtail the load

   Limit what can be drawn or delivered so the total cannot reach the assembly's rating. Current wiring-regulations guidance brings load curtailment in as one of the conditions that can be satisfied — which matters for EV chargers and storage, where an intelligent limit is something the equipment can already do.

4. Suggested — and unresolved

   Feed the board at the opposite end from the supply

   Adding the PV board at the opposite end of the main DB to the supply has been suggested as one way to satisfy Kirchhoff's-law effects: the two sources enter the busbar at opposite ends, so no single length of it carries the whole sum.

   The current may distribute, but the total energy passing through the board may lead to overheating — and that is a thermal question about a specific assembly, not something the arithmetic settles. Seek the manufacturer's guidance before making an unauthorised modification.

All of which reduces to one habit. Before you decide where the inverter connects, write down every source that can deliver current into that board at the same time, and add them up. Then compare the total against what the assembly is rated to carry.

Then do it again for every node that total passes through — the busbar, the main switch, the Henley block, the tails. As the chapter above shows, moving the connection point moves which of those is carrying the sum; it does not reduce the sum. A layout is only settled when every node on the path has been checked against what it is rated to carry.

If it all fits, you have an answer and the paperwork to prove it. If it does not, you have found that out at the design stage — which is the entire point — rather than at an inspection, or at a warm consumer unit.

It takes two minutes and it decides the layout. Everything else on the job — cable routes, isolator positions, where the meters go — follows from where the generation lands.

Where the meters go, once the layout is settled →

Field notes, and a starting point for a conversation — not a design authority. The current edition of BS 7671, your DNO's own requirements and the equipment manufacturer's instructions all outrank anything on this page.