Solar, Battery, and EV Charging in UK Homes: The Contention Problem

You have 5 or 6 hours of cheap electricity overnight. Your EV needs 4 of them. How big should your home battery be?

Most battery guides size for daily consumption. Most EV guides assume the charger has the full overnight window. But if you have both (and especially if you have two EVs) those hours are shared, and the battery only gets what is left after the cars are fed.

I run a 16.1kWh battery and a 3.6kW inverter on Octopus Go. My cheap window is 00:30–05:30. If I also need to charge an EV at 7.4kW, the battery and the car are competing for the same fuse capacity and the same hours. Size the battery wrong and it costs you either way: an oversized battery may not reach its target charge before peak pricing begins, and an undersized one runs empty during the expensive evening hours.

This guide works through the real constraints: fuse limits, inverter bottlenecks, and charger capabilities. Then it shows how to size and schedule a battery around what your EVs leave behind.

The Short Answer: Size for Peak Cover, Then Check You Can Replenish It

Size usable capacity to the expensive-period load you want to cover, then check that the cheap window can actually replenish the energy you will use. Three numbers decide it:

1. Spare fuse headroom: how much current is left under the main fuse for the battery inverter at any given moment. The battery can charge alongside an EV when there is headroom (one 32A EV, the 16A inverter and base load sit around 55A on a 60A supply) and throttles back when a bigger load needs the capacity.
2. Cheap hours you can use: how many half-hours of the window the battery actually gets headroom to charge in. Multiply by the inverter’s charge rate for the most grid energy it can take on.
3. Peak cover you want: the usable kWh you need for the 16:00–19:00 peak and the evening. This sets the capacity you want; the first two numbers decide whether you can refill it.

As an illustration with my own hardware: my 3.6kW inverter draws about 3.6kWh of grid energy per hour, roughly 3kWh reaching the battery after conversion losses. Take one deliberately sequential case where the cars leave the battery two clear hours of headroom: it takes on about 7kWh of grid input, roughly 6kWh stored. Sizing beyond what the window plus next-day solar can replenish just buys capacity that sits empty, and next-day solar cannot backfill the evening you have already paid peak rates for. So set the capacity by the peak load you want to cover, then sanity-check that the window headroom and solar can actually replenish it. If they cannot, the window, not the battery, is your binding constraint.

The Demand Stack at Midnight

Here is what actually competes for your supply overnight:

A typical UK single-phase supply has a 60A main fuse, and the full stack wants 84–89A. On 60A you cannot run everything at once: two 7.4kW EVs alone would blow past the fuse, so devices must run sequentially or throttle down. On a confirmed 100A supply the whole stack fits, but with so little headroom that one big load (an electric shower, say) tips you over. That is exactly why dynamic load balancing matters. Either way, the question becomes: in what order, and how much time does each device get?

In this scenario the EV usually takes priority, because you need the car ready in the morning. A car with 30% battery needs 25–30kWh to fill. At 7.4kW, that is 4 hours: most of your cheap window gone. The home battery gets whatever time remains. If you have a 3.6kW inverter, every hour of leftover window draws roughly 3.6kWh of grid energy, which becomes about 3kWh stored after conversion losses. Two hours left? Around 6kWh stored. One hour? About 3kWh. This is why battery sizing for an EV household is fundamentally different from battery sizing for a solar-only home.

The Three Constraints

1. The Fuse Ceiling

The DNO cut-out fuse is the hard limit. The ENA’s cut-out ratings guidance for EV and heat-pump installers says to assume 60A unless confirmed otherwise; they must never open the DNO cut-out to check. Older metalclad cut-outs from the 1940s–60s may contain fuses as low as 30A.

At 60A (13.8kW), the maths is brutal. One 7.4kW EV charger at 32A consumes 53% of your entire supply capacity. Add 1.5kW of base load (7A) and you are at 39A, about 65%. Add a 3.6kW battery inverter (16A) as well and you reach 55A, roughly 92% of a 60A supply. A second 7.4kW EV charger tips you past the fuse entirely.

At 100A (23kW), it is workable but tight. Two 7.4kW EV chargers (64A) plus a 3.6kW inverter (16A) plus base load (7A) equals 87A. Someone boils a kettle (13A) and you hit 100A exactly. An electric shower takes you well beyond.

2. The Time Window

Intelligent Octopus Go guarantees 6 hours at cheap rate (23:30–05:30) plus discretionary extra hours. Standard Octopus Go gives 5 hours (00:30–05:30). If you must run devices sequentially because of fuse limits, the available energy per device shrinks dramatically.

These figures are grid input, before conversion losses, and this is one deliberately sequential example rather than the only possible schedule. A load-balanced charger can overlap an EV with the battery on a 60A supply until another household load forces it to throttle.

That 7.4 kWh for EV #2 is roughly 25–30 miles of range. If it needs more, you are either paying peak rate or hoping for discretionary cheap slots. Intelligent Go’s guaranteed-window rate is around 7 to 8p/kWh, depending on your region and tariff version (Octopus advertises 8p, and some DNO areas run lower after the April 2026 cut). Cheap, but it does not solve the contention. The window is the bottleneck, not the rate.

3. The Minimum Charge Rate Floor

EVs will not accept less than 6A AC (~1.4kW) per IEC 61851-1. This creates a hard floor: your panels need to produce at least 1.4kW of surplus before the EV will take any of it. You cannot trickle-charge an EV at 3A to share the supply with other devices. It is 6A minimum or nothing.

Load balancing can only modulate between 6A and 32A. If the charger’s CT clamp detects that total household demand would exceed the fuse limit even at 6A, charging suspends entirely. There is no graceful degradation below 1.4kW.

How EV Chargers Handle Contention

The solution to the fuse constraint is a CT clamp on the main supply cable that monitors total current in real time and modulates the EV charge rate to stay within the fuse limit. This is dynamic load balancing, and it is the feature that most separates solar-aware chargers from basic ones.

Zappi v2 (myenergi): £700–850

Among the strongest solar diversion implementations: a CT clamp with three modes (Fast, ECO, ECO+). Dynamic load balancing prevents fuse overload. The Minimum Green Level parameter lets you configure how much surplus is needed before charging starts: at 100%, it needs the full 1.4kW; at 50%, it charges with 700W surplus and imports the rest. PEN fault detection is built in.

Limitation for multi-device setups: No native sequencing across devices. Two Zappis coordinate via the myenergi hub, but battery inverter coordination requires external automation like Home Assistant.

Ohme Home Pro: £435–500

Deepest native Octopus API integration: receives and fills the cheapest charging slots automatically. Solar Boost via included CT clamp adjusts charge rate to prioritise surplus.

Critical limitation: Cannot distinguish between solar generation and home battery discharge. If your battery is discharging, Ohme sees reduced grid import and may throttle charging thinking solar is covering it, or worse, accelerate charging thinking there is surplus while the battery is actually being drained.

Other Chargers Through the Contention Lens

Tesla Wall Connector (Gen 3, £450–550): No native solar surplus awareness. Dynamic Power Management prevents overload but does not actively match generation. Smart scheduling is managed vehicle-side via Tesla’s direct Octopus integration. Requires external PEN fault protection.

Wallbox Pulsar Plus (roughly £550–900 installed): Solar integration needs an optional power meter accessory. No direct UK tariff API integration: manual time-based scheduling only, which is adequate for fixed-window Go but inadequate for Agile or Flux. Standard UK model lacks built-in PEN fault protection.

Hypervolt: Has house load balancing to prevent fuse overload but no solar surplus diversion; there is no CT clamp for generation monitoring. Adequate as a “smart dumb charger” that prevents overload but does not optimise.

The Solar Diversion Trap

Solar diversion sounds perfect: free energy from your roof straight into your EV. The reality is messier, and the efficiency penalty at partial load is the most underappreciated constraint in residential solar-EV integration.

The Efficiency Problem at Low Power

The Zappi in ECO mode modulates down to match available surplus, often settling near the 6A minimum. At this point, the EV’s onboard charger operates deep in its inefficiency zone. A 2016 DTU efficiency study (Kjeldsen, Thingvad, Martinenas and Sorensen) tested a Renault Zoe, a Nissan Leaf, and a Peugeot iOn, and measured total charging losses of 15–40% at minimum current. Charging at full power wastes far less; the penalty comes from running the onboard charger well below its rated load.

At 6A, you might be losing 20–25% of the “free” solar energy to heat in the onboard charger. That 1.4kW of solar surplus becomes roughly 1.05–1.12kW actually reaching the battery. Still better than buying grid electricity at 28p/kWh, but substantially less free than the headline suggests.

Seasonal Mismatch

On a clear summer day a 4kWp south-facing array clears the 1.4kW threshold for most of the middle of the day; on a dull December day it may barely clear it at all. Roughly two-thirds of a UK system’s annual generation falls between April and September (PVGIS). The solar diversion opportunity is strongly seasonal: dominant in summer, near-irrelevant in winter.

WFH vs Commuter: The Availability Dividend

The fundamental variable is whether the EV is parked at home during peak solar hours (roughly 10am–3pm). As an illustrative estimate for a typical 4kWp south-facing system and an average-mileage driver: someone working from home might divert 50–70% of annual solar surplus to the EV (1,500–2,500 kWh/year), while a commuter is limited to weekends and the odd home day, perhaps 15–30% of surplus (400–900 kWh/year), with overnight tariff charging covering the rest.

On those assumptions the direction is clear but the gap is modest: the home worker with effective solar diversion comes out perhaps £60–100/year ahead of a commuter on Intelligent Go. That is an estimate, not a measurement, and it sits alongside the export-versus-divert maths in the next section, which often makes diversion the worse deal for either driver anyway. The commuter’s optimal strategy: overnight cheap charging for baseline needs, weekend solar diversion when possible, and battery-to-home discharge during evening peaks.

The Contrarian Tariff Position: Export Your Solar, Charge Overnight

This is counterintuitive for a solar site, but the maths supports it at current tariffs.

At Intelligent Go’s roughly 7 to 8p/kWh overnight rate and Outgoing Octopus export at 12p/kWh:

• Divert 1kWh of solar into the EV: no cash changes hands, but you give up the 12p that kWh would have earned as export
• Export that 1kWh instead: earns 12p
• Buy 1kWh from the grid overnight for the EV: costs ~7 to 8p

Exporting the solar and buying the EV’s energy overnight leaves you roughly 4 to 5p better off per kWh than diverting. That comparison is on metered input energy. The low-power charging losses covered above push it further in the same direction: a kWh diverted at 6A loses more in the onboard charger than a kWh delivered at full rate overnight.

This only reverses in a few situations: you are not on a cheap overnight tariff (diverting then avoids 28p import while forgoing 12p export, a net benefit of roughly 16p per kWh); your export rate is below your overnight import rate; the cheap overnight window cannot supply all the EV energy you need; or you value resilience and self-sufficiency over pure cost optimisation.

For the triple-setup household on Intelligent Go, the rational hierarchy looks like this:

1. Overnight (the guaranteed 23:30–05:30 window): Charge both EVs and the battery at the off-peak rate, sequenced within fuse limits. Any extra cheap half-hours Octopus dispatches outside that window only apply while the car itself is charging, so the battery cannot bank on slots of its own beyond the guaranteed window.
2. Morning: Battery discharges to cover household load, avoiding 28p import
3. Daytime solar hours: Solar powers home directly (highest value at 28p avoided import). All excess exports at 12p. Do not divert to EV unless export rate drops below overnight import rate
4. Peak (16:00–19:00): Battery powers the home exclusively (or exports on Flux for premium rate)
5. Evening: Battery covers remaining capacity, EVs plugged in but charging deferred to overnight

When Home Assistant Becomes Hard to Avoid

A simpler setup copes without it: one EV, a load-balanced charger, and fixed schedules cover most households. The two-EVs-plus-battery case on a variable tariff is different. Manual scheduling breaks down because there are more variables than any single charger app can optimise: two EV SoC levels, battery SoC, the solar forecast, the current tariff rate, fuse headroom, and discretionary slot availability.

Home Assistant with BottlecapDave’s Octopus Energy integration (half-hourly rate sensors, intelligent dispatching) and CJNE’s myenergi integration (Zappi mode switching, CT clamp data) enables automations like:

• When Intelligent Go dispatching activates AND EV #1 SoC is below 80% AND grid current is below 50A: activate Zappi in FAST mode
• When EV #1 reaches target SoC OR grid current exceeds threshold: switch to EV #2 or battery charging
• When solar surplus exceeds 1.4kW AND no cheap rate is active AND export rate is below avoided import cost: switch Zappi to ECO+ mode
• When peak period starts (16:00): ensure battery has minimum reserve for peak discharge

This is powerful but it is not plug-and-play. It requires technical comfort with YAML automations or Node-RED flows. For technically capable homeowners it is achievable. For a consumer-grade solution, we are probably 1–2 years away. Our Solar Assistant automation guide covers a similar approach for inverter control.

DC-Coupled Charging: The Efficiency Future

The conventional AC pathway from solar panel to EV battery involves multiple conversions: solar DC to inverter AC (96–97%), through the EVSE (~99%), through the EV’s onboard charger back to DC (85–93% at rated power, potentially as low as 68–78% at 6A solar diversion). Total efficiency at partial load can be surprisingly poor.

DC-coupled charging bypasses the onboard charger entirely: solar DC passes through a DC-DC converter, a stage that is typically 97–99% efficient, direct to the EV battery. Because the EV’s onboard AC charger is out of the loop, DC coupling largely sidesteps the partial-load losses that hurt AC solar diversion.

The Sigenergy SigenStor EVDC is the first commercially available DC-coupled solar-to-EV system in the UK. It integrates solar inverter, DC EV charger, battery PCS, battery pack, and energy management into one unit with CCS2 bidirectional charging at 12.5 or 25kW. It is available from UK distributors but represents a whole-system replacement, not an add-on. After a price cut in May 2026, UK pricing is roughly £1,900–2,900 ex VAT for the EVDC module, depending on model and cable length.

SolarEdge announced a bidirectional DC EV charger at Intersolar 2023 but has repeatedly delayed it. Given SolarEdge’s financial difficulties, I would not count on it until it actually ships.

Why Can’t I Charge a Battery at 7.4kW?

It depends on the battery. Not all batteries are created equal, and the answer changes depending on your hardware.

A large rack battery (say a 32kWh unit rated at 57V and 300A) can accept nearly 17kW of DC power. That battery is not the bottleneck. But a single Fogstar 5.12kWh rack unit uses 100Ah cells (EVE LF100LA, 16S1P) with a max charge rate around 100A at 51.2V, roughly 5.1kW. Still more than most inverters can deliver, but not the massive headroom of the larger unit.

So the battery can be a constraint on smaller single units. But in most UK residential setups, the inverter is the tighter bottleneck.

The real bottleneck chain, from narrowest to widest:

1. The inverter’s AC-to-DC conversion capacity. A Sunsynk 3.6kW inverter can only convert 3.6kW from AC to DC, regardless of what the battery can accept. Some 3.6kW hybrid inverters max out nearer 3.3kW of battery charge power, so check the spec sheet rather than the model name. A 5kW inverter lifts this to roughly 4.6–5kW, but then you need a G99 application to your DNO.
2. The circuit fuse and cable gauge, a matched pair. A typical inverter circuit runs on a 32A MCB with 6mm² cable (7.4kW max) or a 20A MCB with 4mm² cable (4.6kW). The fuse protects the cable; the cable dictates the fuse. Together they set the hard ceiling on how much AC the inverter can draw, regardless of what the inverter or battery can handle.
3. The main fuse. At 60A, the battery’s share of the overnight window competes with every other device on the supply. At 100A there is more headroom, but the circuit fuse and inverter still cap the actual charge rate per device.

Always check your battery’s spec sheet for maximum charge current and voltage. Then check your inverter’s maximum battery charge power. The lower of the two is your actual charge rate. In most UK residential setups, the inverter is the constraint.

Solutions: run two batteries in parallel to split current (viable with systems like Fogstar rack units), use a higher-power hybrid inverter (5kW+, but G99 required), or accept the 3–4 hour charge window and schedule accordingly.

The practical impact for contention: if your inverter maxes out at 3.6kW, it needs roughly 3 to 4 hours to take in 10–15kWh of grid energy at 3.6kW, about 83% of which reaches the battery. On a 6-hour cheap window shared with two EVs, this is workable but tight.

What I Would Build Today

If I were designing a triple setup from scratch in 2026 for a UK home on single-phase supply:

• Request a 100A fuse upgrade from your DNO before anything else. UKPN treats this as a free service, and many DNOs will upgrade a 60A cut-out to 100A on request, often at no charge. Others vary: some charge a small fee, and SSEN will only replace a damaged fuse, not upsize on request. Contact your DNO directly and ask; the worst they can say is no.
• Zappi v2 for the primary EV: solar diversion for summer, load balancing year-round.
• Ohme Home Pro for the secondary EV: best Intelligent Go integration for overnight slot-filling. The Ohme/solar confusion does not matter for overnight charging.
• 3.6kW hybrid inverter with 10–16kWh battery: stays under G98, charges in around 3–4.5 hours overnight.
• Intelligent Octopus Go as the base tariff: six guaranteed cheap hours, plus discretionary bonus slots while the car is smart-charging.
• Export all daytime solar surplus at 12p/kWh rather than diverting it to the EV at a ~4 to 5p/kWh opportunity cost.
• Home Assistant with Octopus and myenergi integrations for sequencing and optimisation.

Total hardware cost excluding solar panels: my ballpark is roughly £4,500–6,000 for both chargers, the inverter, and the battery. I will not put a single number on the payback, because it depends on your annual mileage, your tariff, how much energy the battery cycles, and how much peak import it displaces. Work those through for your own household before you spend.

Managing all of this properly is not simple. But for the households willing to invest the setup time, the returns are real and the system gets smarter as tariff APIs and automation tools improve. Start with the solar foundation, add one piece at a time, and automate as you go.

Nikola Nedoklanov

UK-based solar DIY enthusiast with 5+ years hands-on experience.

About the author

Published: July 12, 2026

Last updated: July 24, 2026