Solar and EV charging

The constraint is the panel, and the federal research says so plainly

People arrive at this question thinking about solar output and vehicle range. Neither is the binding constraint. The constraint is how much current the busbar in your electrical panel can carry, and how the code requires that current to be calculated.

Start with what a home charger physically is. A federal technical brief describes a residential Level 2 installation as the price and labour associated with installing "one 40-ampere, 208/240-volt dedicated branch circuit and a circuit terminating in a receptacle, junction box, or EVSE". One circuit. Forty amps. That is the whole thing.

Forty amps at 240 volts is a very large residential load. It is comparable to an electric range or a heat pump, it can run for hours rather than minutes, and unlike a range it is likely to be running at the same time as everything else in the evening.

And the research names the panel as the reason installations get expensive. One study, cited in that brief, found the average high-cost Level 2 installation in an existing home was approximately $2,900, and identified the key factors as "insufficient electrical panel capacity for a dedicated 40-ampere charging circuit", the location of the electric panel relative to the garage, and permit costs.

The distributional finding in that same work is the part worth repeating. The capacity limitation "was found to be more prevalent in less-affluent areas". Older housing has smaller services. So the households for whom the running-cost saving of an electric car matters most are disproportionately the households facing a panel upgrade before they can charge one at home. That is a structural problem, and it is the reason load management is being taken seriously by national laboratories rather than treated as a gadget.

Why the code makes the panel look smaller than it is

The frustrating part of a panel upgrade is that it is often driven by arithmetic rather than by anything actually happening in the house.

The traditional load calculation adds up what could be connected, not what runs together. One industry description of the problem puts it directly: the calculation "assumes almost all of the home's appliances are on at the same time", which "some[times] result[s] in over building the electrical system". Your range, your dryer, your water heater, your air conditioner and your car charger are unlikely ever to draw full current simultaneously, but the calculation that sizes your service does not know that.

So a house whose real peak demand is well within its service can fail a calculation and require an upgrade. That is the situation load management exists to address, and the code now has a route for it.

Here is where we have to be careful about sourcing. The national electrical code is a copyrighted standard behind a paywall, and we could not obtain its text. The clause language below comes from a whitepaper published by a battery manufacturer that sells the product that does this, reproduced in a two-column format alongside its own commentary. We are publishing the mechanism, which is real and well attested, and attributing the wording to where we got it. Nothing below should be treated as read from the code itself.

The mechanism, as that document describes it: the 2023 edition moved energy management system provisions into a new article, and the provision governing the current setpoint permits "a single value equal to the maximum ampere setpoint of the EMS" to be used "for calculating connected load". In other words, a listed energy management system that physically limits current lets that limit be used in the calculation instead of the sum of everything that could be connected.

The equipment has to be field marked, according to the same reproduction, with the maximum current setting, the date of the calculation and setting, and identification of the loads and sources associated with the current limiting feature. That marking is the audit trail, and if you go this route it is what an inspector will look for.

There are things such a system may not shed, including fire pumps, emergency systems, legally required standby systems, critical operations power systems, and circuits supplying emergency lighting. The whitepaper marks those categories as not applicable to residential products, which is fair, but they tell you the design principle: load management is permitted for convenience loads, not for safety systems.

The measurement: what load management actually did, with numbers

Everything above is a mechanism. This is a measurement, and it is the best-sourced thing on this page.

A national laboratory ran a load-management system on a test setup in which the electrical panel was, in the report's words, "effectively downsized to 70 A in software", while the actual panel limit was 150 amps. The stated purpose was to understand how the approach would work in "older homes or apartments with sub-100 A panel capacity, often found in low-income and disadvantaged communities". So the test was deliberately harsh, and deliberately aimed at the households the cost research had identified.

The headline result. The system "successfully throttled the Wallbox EV charger load from 28 A to 6 A" through interface integrations, "reducing the peak energy load from 64 A to 42 A (35% reduction)". The panel was kept within safe operating limits "while ensuring the consumer received a sufficient EV charge".

The control thresholds are worth knowing because they explain the behaviour. Load management actions occurred when 49 amps was reached, which is 70 percent of the simulated capacity, in order to keep total panel load under 56 amps, which is 80 percent. It acts early, at 70 percent, to avoid ever reaching the 80 percent line.

And the worked example is the part that makes it concrete. Twice, at 6:39 in the evening and again at 6:56, a dryer turned on while the car was charging at its maximum of 28 amps. With the dryer running, the maximum load was approaching 49 amps, so the system "throttled the EV charging down to 16 A to keep the panel within safe capacity". When the dryer finished, it "sensed the available load in the panel and increased the EV charger amperage to 28 A after approximately one minute".

A second run using winter usage patterns for a colder climate saw the system detect that panel load had approached 80 percent of capacity and throttle the charger down from 21 amps to 6.

The laboratory's own summary of applicability is the sentence to hold onto: the system "effectively maintained the panel within its capacity, even for 100 A panels, which are common in the US".

Where solar fits, and the honest limit on what we can tell you

Everything above has been about the panel. Solar changes the picture in ways that are real and in ways that are frequently overstated, and we want to be clear about which is which.

What is structurally true is that an electric car is a large load you can schedule. That is unusual. Most household consumption happens when the household decides, but a car parked overnight can be charged at any hour before morning. A schedulable load is exactly what pairs well with generation that arrives on its own timetable and with tariffs that price hours differently.

Both battery manufacturers we read ship time-of-use scheduling as a product feature. One describes a time-of-use mode in which the homeowner sets on-peak and off-peak periods to match their rate, the system draws on solar and stored energy during peak periods, and charges to full during off-peak. The load-management report describes rate optimisation in the same terms.

What we cannot give you is a number. We did not obtain any primary source quantifying what charging an electric car under a time-of-use rate saves. That figure depends on your tariff's peak and off-peak prices, the shape of the differential, your annual mileage, your vehicle's efficiency and how much of the charging you can genuinely move. Anyone offering you a single figure has either modelled your specific tariff or has made it up. We are not going to publish one.

And there is a timing mismatch worth stating plainly, which is our reasoning rather than a citation. Solar generates during the day. Most cars are away during the day and charge at night. Unless you work from home, have a second vehicle, or can charge at your workplace, the electricity going into your car overnight is not the electricity your roof made that afternoon unless a battery moved it there, or unless your tariff credits exports at a rate that makes the round trip worthwhile. That is not an argument against pairing them. It is an argument for being precise about which benefit you are actually buying.

The backup question nobody asks until the power goes out

If you are adding a battery at the same time, one question deserves to be settled deliberately rather than discovered: does the car charger sit on the backed-up side?

A charger is by far the largest discretionary load in most houses. Forty amps at 240 volts is roughly the entire continuous output of a typical home battery. Putting it on the protected panel means that, during an outage, charging the car and running the house are competing for the same stored energy, and the car will win by simply drawing what it is allowed to.

Manufacturers treat these large 240-volt loads as a distinct category. One system's controllable circuits are described specifically as the large 240-volt loads, which is the category a car charger, a range and a well pump sit in. That categorisation exists because these loads need managing rather than merely connecting.

There is a case for backing it up, and it is a real one. In a long outage, a car with a large battery is a meaningful energy reserve and being able to top it up from a roof that is still generating has obvious value. Some vehicles can supply power back to a house, which changes the calculation entirely, though whether yours can and what equipment that requires are questions for your vehicle manufacturer rather than for us.

The case against is that an unmanaged charger on a protected panel can empty a battery overnight while you sleep. If it goes on the protected side, it needs to be a circuit you can control, and you need to know how to turn it off.

This is a decision made at quote time and then physically wired, which is why we cover which circuits get protected as its own subject. Raise the charger explicitly, because it is the single load most likely to be assumed rather than decided.