Solar Inverters Explained: String, Hybrid and Micro (UK 2026)

By Rob Hammond · Updated September 2026 · How we score

PowerGuide is reader-supported. We may earn a commission from links on this page. It never affects our rankings or what we recommend. Learn more

The panels get the attention and the inverter does the work. It converts the array’s DC into 230V AC, decides how hard to push the panels from moment to moment, talks to your battery if you have one, and, because its AC rating is the number the network operator cares about, it quietly decides whether your installation is a form you post or an application you wait for. This is what the three types actually do, how to size one, and where the money goes wrong.

String, hybrid, micro: what each one is

String inverter. One box on the wall, all the panels wired in series into it as one or two “strings”. Cheapest per kW, simplest to fault-find, and the default for an unshaded roof that faces one way. It converts DC to AC and nothing else: no battery port, no storage.

Hybrid inverter. The same job plus a battery port and a BMS communication link, so surplus generation can go straight into storage as DC. Six of the ten inverters in our catalogue are hybrids; three are plain string inverters and one is a microinverter. The Solis S6 Hybrid datasheet lists a battery voltage range of 40–60V, Li-ion or lead-acid, with one battery port and one BMS port over CAN/RS485, and a maximum charge/discharge current of 112A on the 5K model. The Growatt SPH 5000 datasheet lists 42~59V, lithium or lead-acid, and a 10ms UPS transition to backup power.

Microinverter. One small inverter per panel, mounted under it, so every panel does its own DC-to-AC conversion and its own maximum-power tracking. Enphase’s UK datasheet for the IQ8 series gives maximum AC output of 330W (IQ8MC), 366W (IQ8AC) and 384W (IQ8HC), with European weighted efficiencies of 96.7%, 96.6% and 96.8%. The trade is cost and the number of components on the roof against per-panel independence, and the same architecture is what makes plug-in and balcony kits work, where one 800W micro sits behind one or two panels and plugs into a socket.

Which fits:

Sizing: your array is in kWp, your inverter is in kW

These are different units measuring different things, and they are deliberately not equal. Your array’s kWp is its output under standard test conditions: bright, cold, perpendicular light that a UK roof sees for a handful of hours a year. Sizing the inverter to match that peak means buying capacity that is idle almost always.

So installers oversize the DC side, and manufacturers expect it. Solis states the S6-EH1P range “supports PV input up to 160% of the inverter’s rated DC power”. That 160% is the datasheet’s max. usable PV input power of 8kW on the 5K model, while its recommended max PV array size of 10kW is the larger STC figure, because an array rarely delivers its rating. Growatt’s SPH 5000TL BL-UP allows a max recommended PV power of 9,500W against a 4,000W AC output.

When the array does produce more than the inverter can pass, the inverter simply holds output at its ceiling and the surplus is never harvested. That is clipping, and the only question that matters is how much of the year’s generation it costs you.

What clipping costs, on our own model

PowerGuide’s balcony solar calculator contains a clipping curve, and it is the site’s own model, not a measured UK rooftop dataset. We calibrated it on 2 July 2026 against German Balkonkraftwerk simulations, where a 2,000Wp array on an 800W inverter (ratio 2.5) loses about 19% to clipping; our curve returns 20% at that point. Treat it as the right shape with uncertainty of a percentage point or two, and be aware it was fitted to small plug-in systems, not to a pitched roof in Yorkshire.

Worked through, using the array from how many solar panels do I need: ten 440W panels is 4.4kWp, and PVGIS gives a south-facing Midlands roof at 35° pitch 991kWh per kWp, so 4,360kWh a year before clipping.

• On a 3.68kW inverter. Ratio 4,400 ÷ 3,680 = 1.20. The curve returns 0.98%, about 43kWh a year. At the Ofgem cap rate of 26.32p that is £11 if you would have used it yourself, and £5 at Outgoing Octopus’s 12p export rate if you would have exported it.
• On a 3kW inverter. Ratio 1.47. The curve returns 4.6%, about 200kWh, worth £53 self-used or £24 exported.
• On the Solis 5K at its own recommended maximum. A 10kWp array on 5kW is ratio 2.0. The curve returns 12%. The manufacturer’s ceiling is a permission, not advice.
• On the Growatt 5000TL at its 9,500W maximum. Ratio 2.375, about 18%.

The practical reading: a ratio around 1.2 is close to free, 1.5 starts to cost real money, and anything at or beyond 2.0 is only rational when the inverter’s AC rating is capped by something else, which in the UK it often is. See the G98 section below.

MCS does not settle this for you. We read MIS 3002:2025 Issue 1.0, the current Solar PV installation standard, in full, and it contains no clause setting an array-to-inverter ratio. What MCS does require is the paperwork you need to check it yourself: MCS 032:2025 Issue 1.0, the pre-sale information standard, clause 6.2.4, says that “as a minimum, the following technical information shall be communicated in writing to the customer before the point that the contract is awarded”, and item (c) on that list is the “Manufacturer’s datasheet for the proposed inverter”. So you are entitled to the datasheet before you sign. Ask for it, and check the array against the numbers above yourself.

MPPTs and shading

An MPPT, or maximum power point tracker, is the circuit that continuously hunts for the voltage at which a string produces most power. One MPPT can hold only one operating point. Put an east-facing string and a west-facing string on the same tracker and it compromises between two arrays that peak five hours apart; the weaker one drags the stronger one down all day.

Every full-size inverter in our catalogue has two. The microinverter is the sole exception, and for the opposite reason: it serves one panel, so one tracker is all it needs. The Solis datasheet lists “MPPT number / Max. input strings number 2/2” with 16A maximum input current per tracker; the Growatt lists “No. of MPP trackers 2”, one PV string per tracker, 13.5A maximum input current each. For a split east–west roof, which our sizing guide puts at about 80% and 78% of south-facing yield respectively, so genuinely worth doing, two MPPTs is the difference between two independent arrays and one confused one.

Two MPPTs do not fix shading. A chimney shadow crossing three panels in a twelve-panel string pulls the whole string’s current down, because series-wired panels share a current path. The fixes are panel-level: optimisers, which keep a string inverter and add a DC device per panel, or microinverters, which remove the string entirely. Enphase’s UK datasheet is explicit that there is “no enforced DC/AC ratio and maximum input power” for module pairing, with per-unit maximum input power of 480W (IQ8MC), 530W (IQ8AC) and 560W (IQ8HC). Panel-level kit is sized per panel, which is exactly why it suits awkward roofs with three panels here and five there.

Battery readiness

There are two routes to storage and the inverter choice is what separates them.

DC-coupled, via a hybrid. One inverter, one conversion step from panel to battery, slightly better round-trip efficiency, one app. This is what the six catalogue hybrids are built for. The cost of it is commitment: the hybrid’s battery voltage window decides what you can buy later. The Solis takes 40–60V, the Growatt 42~59V, the GivEnergy Hybrid 5.0 Gen 3 45–58V and the Sunsynk ECCO 5kW 40–60V, all low-voltage platforms, so a high-voltage battery stack is not a later option on any of them. The Fox ESS H1 5.0 (G2) is the counter-example in our catalogue: its datasheet gives an 80–480V battery voltage range, a high-voltage platform that rules the 48V-class packs out instead.

AC-coupled, retrofitted. Keep the string inverter, add an all-in-one battery with its own inverter built in. More conversion losses, two systems to monitor, but no requirement to have predicted the future. Our are solar batteries worth it guide sets out the economics, including the trap that makes this decision expensive: a modular battery pack is only the cheaper route if you already have a compatible hybrid inverter, because otherwise you are buying an inverter alongside it.

The honest rule: buy a hybrid if a battery is a decision you have already made or nearly made. If storage is a vague someday, buy the cheaper string inverter and AC-couple later. You will have lost a few percent of round-trip efficiency and saved the cost of a battery port you never used.

The 3.68kW line

This is the single most consequential number in inverter selection, and it is not on any datasheet’s performance page. ENA Engineering Recommendation G98 Issue 2, in force from 10 March 2025, covers micro-generators “up to and including 16 A per phase” and converts that for you: 16A “corresponds to 3.68 kilowatts (kW) on a single-phase supply and 11.04 kW on a three-phase supply”. MCS repeats the split in MIS 3002:2025: G98 up to and including 16A per phase, G99 above it. It notes that the threshold is “the total aggregated AC output of all generators”, batteries included.

So: 3.68kW is the largest single-phase inverter you can install on the connect-and-notify route. Manufacturers build to it deliberately. Growatt’s range has an SPH 3600TL BL-UP whose max AC output power is exactly 3,680W at a rated output current of 16A, a model that exists for one regulatory reason. Enphase’s UK datasheet caps microinverters at 11 IQ8MC units per single-phase circuit, which is 3,630W, just under the same line. (That row is headed “Max. units per single/Multi-phase 20 A circuit”, but the note beneath it does the calculation at 16A per phase. The 16A is the number that matters for G98, so don’t be thrown by the header.)

Three of the nine full-size inverters in our catalogue are under it, and each one gets there on its datasheet’s rated output current. The Growatt SPH 3600TL BL-UP is the hybrid: 16A rated output current at a 3,680W maximum AC output. The Solis S6-GR1P3.6K is a string inverter at 16.0A rated grid output current at 220V (15.7A at 230V), with a 16.0A maximum output current, and the SMA Sunny Boy 3.6 is the other, at a 16A maximum output current on 3,680W rated power. (The tenth inverter, the IQ8MC microinverter, is under 16A many times over per unit, at 1.43A, so for panel-level kit the question is how many units go on the circuit, not what one unit does.) The 5kW units are not. The Solis S6 5kW hybrid’s (S6-EH1P5K-L-PLUS) rated grid output current is 22.8A at 220V (21.8A at 230V), and the S6-GR1P5K string unit’s is 22.7A (21.7A at 230V); the Growatt SPH 5000TL BL-UP’s datasheet gives a rated output current of 22A, higher than the 17.4A you get by dividing its 4,000W maximum AC output by 230V, and 22A is the figure that would go on a DNO form. Both mean a G99 application, approved before connection, not a notification afterwards. That is not an argument against them, because a 5kW hybrid on a 6kWp roof with a battery is a perfectly sensible system, but it is a three-month conversation rather than a form, and you should know that before you order. G98 vs G99 walks through which route you land in and what the DNO can say, including export limitation, which is the usual answer.

This is also the honest reason to accept a high DC:AC ratio. If G98 is worth more to you than the top 4% of your generation, and for many households it is, then a 4.4kWp array on a 3.68kW inverter at ratio 1.20, losing about 43kWh a year, is a deliberate and cheap trade.

Replacement: when, and what nobody publishes

Panels are warranted for decades. Inverters are not, and the gap is the point.

The Solis datasheet carries no warranty row at all. Resellers advertise five-year and ten-year terms, but a reseller’s listing is not a manufacturer’s commitment and we are not going to cite one as if it were. Ask Solis or your installer for the warranty document in writing before you buy.

There is no official UK statistic for what a replacement inverter costs fitted, and we are not going to invent one. Our solar panel cost guide says the same thing for the same reason: DESNZ publishes installed cost per kW for whole systems, not component replacement prices, and neither Solis nor Growatt publishes an RRP. What we can give you is the scale of the hardware: our catalogue carries the Solis at £966 and the Growatt at £950, both indicative catalogue estimates rather than quotes, and neither includes labour, scaffolding if the array must be isolated, or re-commissioning. Budget for a replacement somewhere in the second decade and treat any payback calculation that ignores it as optimistic, including ours, which is why the solar payback calculator is a hardware model, not a lifetime one.

One note on our own data. Reading these datasheets for this guide turned up three errors in our own catalogue, and we corrected them on 21 September 2026. We had the Solis at 98.4% maximum efficiency where the S6-EH1P(3-8)K-L-PLUS datasheet says 96.2% (EU efficiency 96.1%); we had the Growatt at 97.6%, which is the 6000TL’s column, where the 5000TL’s is 97.50%; and we rated the Growatt at 5kW where the current datasheet gives the SPH 5000TL BL-UP a 4,000W maximum AC output. The catalogue now matches the datasheets. Where anything here still disagrees with a manufacturer’s own document, believe the document and its date, not us.

When the answer is no

Do not buy a hybrid “just in case”. A battery port you never connect is money spent on a constraint: the voltage window narrows what you can buy later, and the hybrid’s efficiency on the PV path is no better than a good string inverter’s.

Do not buy microinverters for an unshaded south roof. You are paying a premium per panel for independence that a single string would have given you for free.

Do not buy a bigger inverter to “future-proof”. Going from 3.68kW to 5kW to leave room for a battery you might add moves you from G98 to G99 immediately. You take the slower route today for a benefit that may never arrive, and a G100 export limiter may be imposed anyway.

Do not let an oversized array excuse an undersized inverter without doing the arithmetic. At ratio 1.2 the loss is about 1%; at 1.5 it is 5% and worth £53 a year on the worked example above. The line between “clever” and “cheap” is somewhere in between, and it moves with your export rate.

Do not accept a quote that does not name the inverter model. MCS 032:2025 clause 6.2.4(c) requires an MCS contractor to give you the manufacturer’s datasheet for the proposed inverter in writing before the contract is awarded. Every number in this guide came off one; yours should too. Start at the solar hub if you are still deciding whether to install at all, and the inverter listings when you are comparing specific units.

All prices, rates and figures were checked on 21 September 2026. The price cap is reset quarterly and manufacturers revise datasheets without notice, so re-check both before committing.

What’s changed (changelog)

• 21 September 2026 (revised same day). Catalogue expanded from two to ten inverters, each added against a manufacturer datasheet read on the day: Growatt SPH 3600TL BL-UP, Solis S6-GR1P3.6K, Solis S6-GR1P5K, SMA Sunny Boy 3.6 (SB3.6-1AV-41), GivEnergy Hybrid Inverter 5.0 Gen 3, Fox ESS H1 5.0 (G2), Sunsynk ECCO 5kW (SUN-5K-SG05LP1-EU) and the Enphase IQ8MC microinverter. That added the first string inverters, the first microinverter and the first G98-size units to the category, so the counts and the G98 paragraph in this guide were rewritten to match.
• 21 September 2026. First published, against the Solis S6-EH1P(3-8)K-L-PLUS and Growatt SPH 3000-6000TL BL-UP manufacturer datasheets, the Enphase IQ8 UK datasheet, MCS MIS 3002:2025 Issue 1.0, MCS 032:2025 Issue 1.0 and EREC G98 Issue 2, with clipping losses from PowerGuide’s own balcony clipping curve. Writing it corrected three specification errors in our own inverter listings: the Solis efficiency, and the Growatt’s efficiency and AC output rating.