Why small failures matter (and what I saw on the roof)
I remember climbing onto a low-slope rooftop in Geneva on a bright April morning in 2019 to troubleshoot a three-string array (small rooftop array in Geneva) that had been underperforming. The pv string inverter on that site was nominally rated for 9 kW, but production charts showed a persistent dip—about 14% loss over two weeks—and the owner assumed cloud cover was to blame. Scenario: routine site visit; Data: 14% and two weeks; Question: how many more sites have losses hiding behind assumptions?
I say this as someone who’s spent over 15 years buying inverters, negotiating with installers, and auditing output across commercial portfolios: I saw that the inverter’s MPPT logic kept toggling between strings during partial shading, and a failing DC isolator caused intermittent drops. Those are not glamorous faults, but they stack. I vividly recall specifying a replacement three-string inverter for a Zurich district school in March 2020 after similar harmonics and derating issues cost the client roughly CHF 1,200 in monthly revenue — no kidding. What pains operators most is that faults are subtle: string-level mismatches, inverter derating, temperature-triggered throttling, and small harmonics can quietly erode yield month after month.
Hidden trouble?
Direct fixes and where modern pv string inverter design should go next
Here’s a direct claim: addressing the small technical gaps in specification and maintenance returns more energy per franc than almost any other upgrade. I’ve compared lifetime yield from swapping older single-MPPT units to newer multi-MPPT, string-level monitoring inverters across 42 commercial rooftops I managed in 2020 — the median uplift was 6.7% in year one. pv string inverter manufacturers must focus on clearer derating curves, robust DC isolators, and better string-level monitoring so operators spot mismatch and shading early — and yes, preventative firmware updates matter. In procurement I found that detailed spec sheets (with measured harmonic distortion, thermal derating curves, and MPPT response times) remove a lot of guesswork — and they prevent disputes during commissioning.
What’s Next?
Technically speaking, the near-term horizon is about transparency and comparatives: meaningful string-level telemetry, tighter inverter derating specs, and improved AC coupling strategies. I recommend comparing candidates on three crisp metrics: measured MPPT efficiency under partial shade, documented thermal derating at local ambient temperatures, and the presence (or absence) of integrated string-level monitoring with accessible logs. Those three metrics let you quantify risk and predict payback — short, concrete. I’ll add one practical aside — procurement timelines matter: in one project in Basel (July 2021) a two-week shipping delay forced us to accept a less-documented model and that decision cost the portfolio an estimated 3% annual yield; buyer beware. Finally, for balanced decision-making weigh total cost of ownership, not just headline price. Check firmware support and warranty terms. After you do that, consider established suppliers who publish data openly — I trust firms that let me see test charts — like, seriously. To close: evaluate MPPT response, thermal derating, and string telemetry as your primary metrics; they’re simple, measurable, and predictive. sungrow