String sizing and roof hips: why quotes skip your weakest panel
I pulled a quote apart for a mate in Ocean Grove a few months back, a modest hip roof with panels running across three faces. Nineteen panels, one string. The installer had sized it on total wattage and called it a day. Nobody had asked what happens when the north-west corner gets afternoon shade from the neighbour's Norfolk pine for forty minutes in winter. That's the gap I want to talk about today: string sizing on a roof with more than one plane, and why the weakest panel on a string drags the whole string down with it, no matter how good the other eighteen are.
This is different from the usual orientation or shading conversations. You can have a roof with zero shading and still lose meaningful output because of how panels are wired together, not where they're pointed.
What a string actually is and why it matters
Panels wired in series form a string, and a string behaves like the weakest link in a chain. Current through a series string is set by the lowest-performing panel in that string, not the average. If one panel on a hip-roof corner is partially shaded, dirtier, slightly warmer because it sits over a sarked section without a gap, or just a different batch with marginally different output, the whole string's current drops to match it.
On a single flat roof with uniform pitch and no obstructions, this rarely bites you because every panel sees roughly the same conditions. The problem shows up on hip roofs, roofs with dormers, roofs split across more than one orientation, or roofs with a chimney, vent pipe or skylight that shades one corner for part of the day. These are exactly the roofs where installers are most tempted to keep things simple and run one long string across everything, because separate strings mean more homerun cable and sometimes a second MPPT input they'd rather not use.
Where this quietly costs you kWh
I've seen quotes where a 6.6kW array nameplate produces noticeably less than a same-sized array on a simpler roof, and the client assumed it was just a "bad roof for solar." Sometimes that's true. Often it's actually a wiring decision made at quoting stage, not a physical limitation of the site.
The giveaway is usually in the single-line diagram, if the quote even includes one. If every panel from three different roof faces feeds into one string on one MPPT, that's a flag. If the installer has instead grouped panels by roof face, with each face on its own string or its own MPPT input, that's a sign someone actually looked at your roof geometry rather than just counting squares.
Most hybrid and string inverters on the market now, including something like the Symo GEN24 8.0 Plus, ship with two or more independent MPPT trackers specifically so installers can separate roof faces electrically. The hybrid inverter category broadly does the same. The hardware to do this properly is standard. Whether it gets used properly on your job is a design choice, not a hardware limitation.
The chimney, the vent pipe and the one panel nobody mentions
I've walked enough roofs to know the pattern. There's always one panel position that's slightly compromised, tucked next to a plumbing vent, sitting under the shadow line of a chimney for an hour around midday, or positioned where a TV antenna bracket throws a sliver of shade across one cell string inside the panel itself. On its own that panel might lose five or ten percent of its output on a sunny day. Wired into a string with eighteen other panels that have no such issue, it can drag the measured output of the entire string down by a similar percentage, not just its own output.
This is where panel-level optimisers or microinverters earn their keep on awkward roofs, because they let each panel operate at its own best point regardless of what its string-mates are doing. They cost more per watt and add another component that can fail, so I wouldn't default to them on a clean, unshaded, single-orientation roof. But on a hip roof, a roof with three faces in use, or anything with a fixed obstruction like a chimney that isn't going anywhere, the maths often favours them. It's the same logic we cover when looking at east/west split-orientation roofs, just one level further down in the wiring.
How to actually check a quote for this
Ask for the single-line electrical diagram before you sign, not after. A legitimate design should show which panels are grouped into which strings, and which strings feed which MPPT input. If the quote doesn't include one, ask for it. Any accredited designer should be able to produce this in a few minutes because the design software they use, whether that's Aurora, PVsyst-based tools or a manufacturer's own platform, generates it as part of the layout process.
Then walk your own roof, or look at satellite imagery with the time of day in mind, and mark where the fixed obstructions sit: chimney, evaporative cooler, solar hot water tank, aerial, box gutter. Cross-reference those positions against the string diagram. If a compromised panel is grouped with panels on a totally different, unobstructed face, that's worth raising with the installer directly. A reasonable installer will either re-group the string or explain why it won't matter much on your specific layout, for instance if the "obstruction" only shades for twenty minutes near sunset in June.
I'll be upfront about my own view here: I think a lot of installers quote strings based on cable run convenience rather than roof geometry, because it's faster to design and faster to install. It's not usually dishonest, it's just not optimal, and most homeowners never find out because the system still "works," it just underperforms by a margin that's hard to notice without monitoring data to compare against.
Monitoring is how you'd actually catch it after the fact
If you've already got panels installed and you suspect a string issue, panel-level or string-level monitoring is the only real way to confirm it. Whole-system monitoring that just shows total daily kWh won't tell you anything, because a 15% shortfall on one string of three gets diluted across total output and looks like normal day-to-day variation. You want monitoring that reports per-string or per-MPPT data at minimum, ideally per-panel if you've got optimisers.
Compare output between strings on a clear, cloudless day when every panel should theoretically be performing close to its rated capacity scaled for conditions. If one string consistently underperforms by more than what you'd expect from orientation differences alone, that's your evidence. From there it's either a re-wiring job, cleaning, or confirming there's a physical obstruction nobody flagged at design stage.
Where this intersects with warranty claims
This matters for warranty conversations too. If you're chasing a performance warranty claim under terms like the ones we've gone through in solar warranty terms that actually matter, the manufacturer will want evidence the shortfall is a panel defect, not a design or shading issue upstream of the panel itself. A string wired across three roof faces with one compromised panel muddies that evidence considerably, because the manufacturer can reasonably argue the system design, not their product, is the cause. Clean, well-documented string design isn't just about day-one performance, it protects your position if you ever need to make a claim under the product versus performance warranty distinction years down the track.
The Clean Energy Council's installer guidelines in Australia and the relevant National Electrical Code requirements for PV source circuits in the US both leave string configuration largely up to designer judgement within safe voltage and current limits. There's no regulation that forces an installer to split strings by roof face. It's a quality decision, not a compliance one, which is exactly why it varies so much between quotes.
What to ask for before you sign
Request the string diagram, ask specifically whether strings are split by roof orientation or grouped by convenience, and ask what happens electrically if one panel underperforms, whether due to shading, soiling or a manufacturing variance within normal tolerance. A good installer will have a ready answer. One who hasn't thought about it will usually just repeat the total kW figure back to you, which tells you everything about how the quote was put together.
For sizing decisions more broadly, our piece on how many kW your roof can actually fit covers the geometry side of this; string design is the wiring decision that sits on top of that geometry, and it's worth getting both right before panels go on the roof, not after.
— Sam Whitfield, Solar & Batteries, IceWarm
Common questions
- Does string sizing matter on a simple single-face roof?
- Less so. If every panel sits on one plane with the same tilt, orientation and no shading, panels in a string tend to perform similarly and the weakest-link effect is minimal. It becomes a real issue once a roof has multiple faces, fixed obstructions like a chimney or vent, or partial shading at certain times of day.
- Are microinverters or optimisers always worth it on a complex roof?
- Not always, but they're worth pricing for comparison. On hip roofs or roofs with permanent shading from a chimney or similar, per-panel optimisation can recover output that a poorly strung system would otherwise lose. On a clean, single-orientation roof the extra cost usually isn't justified.
- Can I ask my installer to split strings by roof face even if it costs more?
- Yes, and it's a reasonable request. It may mean an extra MPPT input or additional cable runs, which can add modest cost, but on a multi-face roof it's often a better use of money than oversizing the system to compensate for string losses.
- How do I check if my existing system already has this problem?
- You'll need per-string or per-MPPT monitoring data, not just total daily output. Compare string performance on a clear day; a string that consistently underperforms relative to its expected orientation output may indicate a weak-link panel dragging the group down.
Sam spent six years installing residential solar before moving to writing full-time, and covers panels, inverters and home batteries.
Former licensed solar installer (C-46, California).
More from Sam Whitfield
- Roof age vs panel warranty: the 25-year mismatch nobody flagsYour panels are warrantied for 25 years but your roof might not last that long. Here's how to check before you sign, not after the array's up.
- Why your solar quote's "system size" number ignores your switchboardA 10kW solar quote means nothing if your switchboard and main fuse can't handle it. Here's how to check before you sign anything.
- Microinverters vs string inverters: the sizing question nobody asks firstMicroinverters and string inverters get sold on shading and monitoring, but the real decision hinges on how your roof will be sized and expanded over time.
- Shading from next door's tree: why one quote accounts for it and one doesn'tTwo solar quotes for the same roof can disagree by 20% on output because only one actually modelled the gum tree next door. Here's what to ask for.
- How many solar quotes list your roof's actual tilt angle (most don't)Most solar quotes assume a generic tilt for your production estimate. Here's why your roof's real pitch matters and how to check the number yourself.
- How many kW of panels your roof can actually fit (not the brochure number)Roof size and system size aren't the same maths your installer's software assumes. Here's how to work out what will actually go up there.