When a solar inverter exports, it pushes the voltage at your terminals up. Push it past the over-voltage stage and the inverter disconnects — usually at midday, exactly when the energy is worth most. This page works from the inverter's rated power, the cable and the voltage already sitting at the connection point, and returns the rise, the headroom left before a trip, and the largest inverter that cable can carry. Runs entirely in your browser — nothing is uploaded.
System
Nameplate AC output, not the array DC rating.
Measured on a sunny day, not assumed nominal.
110 % of nominal is common. Check your grid code.
Cable
Inverter to the transformer or metering point.
Result
| Size | R one way | Rise | Rise | Max inverter |
|---|
Sizes shown are a window around your selection, not a full catalogue.
A voltage drop calculator starts from a load current and tells you what the load loses. This one starts from an inverter's rated AC power and tells you what the network gains — and whether that ends in a disconnection. It answers three questions an installer actually gets asked:
The third one is the useful direction. It is also the one nobody publishes, because it depends on the local baseline voltage, which varies from street to street.
ρ₂₀ is 17.241 Ω·mm²/km for copper and 28.264 Ω·mm²/km for aluminium; α is 0.00393 /°C and 0.00403 /°C respectively. L is one-way length, A is conductor area, n is parallel runs per phase. ΔV_allow is the tighter of your percentage limit and the physical gap between the baseline voltage and the trip threshold.
Current is calculated at nominal voltage, not at the elevated voltage. A grid-following inverter is a constant-power source, so at a higher terminal voltage it draws a slightly smaller current. Calculating at nominal gives the larger current and therefore the larger rise — the conservative direction. The difference is small, typically under 2 % of the rise, but it is always in your favour to round this way.
Power factor drops out of the maximum-inverter formula. It appears once in the current and once in the apparent power, so P_max depends only on voltage, impedance and allowed rise. That is why you cannot fix a rise problem by commanding the inverter to a lower power factor — you can shave the resistive component a little through the volt-var curve, but the arithmetic does not hand you a free fix.
Design guides talk about rise as a percentage of nominal, which quietly assumes the network sits at nominal before you connect. It does not. On a sunny day, every other system on the same low-voltage feeder is exporting at the same time, so the voltage at your connection point is already pushed up — commonly 3 to 8 % above nominal on a rural feeder with a lot of rooftop solar.
That baseline is not yours, but it is charged against your headroom. On a 230 V system sitting at 248 V with a 253 V trip threshold, you have 5 volts of room left, not 23. A rise of 2 % looks compliant in the design and disconnects the system in service. This is the single most common reason a system passes the paperwork and fails on the roof.
A 10 kW three-phase inverter on a 400 V supply, 90 m of 6 mm² copper back to the transformer, conductor at 70 °C, connection point measured at 405 V on a sunny afternoon, trip threshold 440 V.
6 mm² carries 14.4 A without difficulty — typical buried rating is well over 40 A, so ampacity is not remotely the constraint. At a 2 % limit the design just passes, at 1.93 %. Now change one thing: the connection point measures 428 V instead of 405 V, which is unremarkable on a feeder with a dozen other systems on it. Terminals now sit at 435.7 V, the headroom is 4.3 V, and any further drift in the network trips the inverter. Same cable, same inverter, opposite outcome.
Going to 25 mm² drops R to 0.074 Ω and the rise to 1.85 V, or 0.46 %. That is the difference between a system that runs all afternoon and one that cycles on and off between 11:00 and 15:00.
The Max inverter column in the results table is the figure to quote when someone asks whether they can add panels. It answers: at this conductor, this length and this available rise, what is the largest AC rating that still stays inside the limit? Note that it is a ceiling, not a target — leave margin for the network drifting upward, which it does every year as more systems connect to the same feeder.
If the ceiling is far below the inverter you want to install, the fix is not a bigger inverter and not a lower power factor. It is conductor area, distance, or a network-side change. Those are the only three levers, in ascending order of cost and descending order of how much control you have over them.
Work through these in order — the first two cost nothing and resolve a surprising share of cases:
Caps of around 1–2 % at the point of connection are common, and many networks allow more where the measured pre-connection voltage is low. The figures that actually govern your job come from your connection agreement and the national grid code — for example the G98/G99 settings in the UK, VDE-AR-N 4105 in Germany, or the equivalent standard in your country. Those documents also set the over-voltage stages and their timing, which is why the trip threshold field defaults to 110 % rather than to a fixed number of volts.
This tool gives you the physics and the arithmetic. It does not know your contract, your transformer, or what your neighbours installed last month.