February 2023: The Quote I Was Proud Of
A client in the southwest of the UK wanted a hybrid system that would "keep the lights on and the heat pump running" through grid outages. Three-bed semi, 4.2 kWp roof array already fitted, EV in the driveway.
I quoted a 3kW hybrid solar inverter plus 4× Pylontech US5000. Total storage: 19.2 kWh nominal. Output: 3 kW. I ran the numbers in my head, felt good about them, sent the quote. Signed inside a week.
That quote was £9,400 in hardware. It cost me £2,140 to fix.
Where the Logic Broke Down
The client's brief had two parts and I treated them as one.
Storage is about how long you can run. Output is about how much you can run at once. I sized the first properly. I sized the second with a single sentence of reasoning: "3 kW is plenty for a house."
It isn't—not when the house has a heat pump pulling 2.8 kW and an EV charger pulling 7.4 kW. Not simultaneously; the client was careful. But the heat pump on its own, plus lighting, plus a kettle, hit 4.1 kW peak draw on a cold morning. I verified that after commissioning, not before.
A 3 kW hybrid inverter caps at 3 kW continuous. When the draw exceeds that, the inverter throttles. The battery bank—19.2 kWh of Pylontech US5000 capacity in kWh terms—sits there looking healthy on the app while the meter spins. From the client's view, the battery "wasn't working." From mine, it was doing exactly what I'd specified. The specification was wrong.
That's the first thing I'd tell anyone speccing pv and battery storage: Pylontech US5000 capacity kWh figures tell you duration, not delivery. The datasheet lists 4.8 kWh nominal per module—about 4.56 kWh usable at 95% depth of discharge, and I usually design to 80% if the client wants a comfortable cycle-life margin. Four modules is roughly 15.4 kWh usable at that design point, not 19.2.
Then the Comms Failure
Commissioning, second week of March. The inverter booted, found the battery over CAN, and refused the handshake. Firmware mismatch—the specific 3 kW model I'd picked supported Pylontech US5000 on paper, but only on a later firmware revision than the unit that shipped. The installer on site (not me, thankfully) put the battery into voltage-based fallback mode.
Voltage-based SOC estimation on LFP is imprecise through the flat middle of the discharge curve. The client's app showed 100% for three days, then dropped to 18% in a single evening. They called me at 11 pm.
I should add: this wasn't the inverter manufacturer's fault. Their compatibility document listed the Pylontech series with a footnote about firmware. I skipped the footnote.
The Fix, and What It Actually Cost
Two options. Downgrade the battery to a Pylontech US3000C battery module the existing inverter definitely supported (3.55 kWh per module), or swap the inverter for one on Pylontech's approved list with the right firmware.
I swapped the inverter. £900 for the replacement, £240 for re-commissioning the DC side with a different isolator arrangement, plus about £1,000 of unbillable labour across three site visits.
Total: £2,140. Three weeks' delay. And a client who, reasonably, asked whether they should go with a bigger installer.
What I Got Wrong, Precisely
Not the battery. The US5000 is a solid module—95% DoD, 6,000 cycles to 80% state of health under standard test conditions, and it does what it says on the datasheet. The mistake was mine: I treated "hybrid inverter" as a product category rather than a compatibility question.
It's tempting to think any hybrid inverter will talk to any 48V LFP battery with a CAN port. But Pylontech's BMS protocol has a revision history, and inverters vary in which revisions they support. The compatibility list isn't a suggestion—it's the specification.
The second mistake was sizing. I compared battery capacity to the house's total daily consumption and felt clever. What I should have done is measure peak load. Not annual kWh—peak kW. A 3 kW inverter on a house with a heat pump and EV charger is a 3 kW inverter on a house with a heat pump and EV charger, no matter how many US5000s you stack behind it.
People think more battery equals more backup capacity. Actually, the inverter's continuous output sets what you can run; the battery sets how long. Get them out of balance and the bigger battery becomes dead weight.
What I Changed
I built a pre-commissioning checklist. Seventeen items; the three that matter most:
- Verify the exact inverter firmware against Pylontech's latest compatibility document—not the version on the product page, the version on the support portal.
- Pull the client's peak demand from smart-meter data or a clamp logger for at least two weeks before finalising inverter sizing.
- Order inverter and battery together from a supplier who confirms compatibility in writing. If they won't, order elsewhere.
We've caught eleven near-misses with that checklist in the eighteen months since. Two of them were firmware issues on the US3000C battery that would have looked exactly like my March 2023 failure.
I still kick myself for the 11 pm call. If I'd spent twenty minutes on the compatibility document in February, the client would have had their system in week one and I'd have kept £2,140.
To be fair to myself: the pressure was real. The client had a deposit deadline, the sales team wanted the quote out, and I was juggling four projects. But that's not a defence. It's a description of every project.
What It Taught Me About Quality
The client kept the system. They also referred two neighbours, which I didn't expect. What changed their mind wasn't the fix—it was that I showed up, explained exactly what went wrong, and didn't invoice for the labour.
If I'd tried to save £200 by shopping for a cheaper inverter without checking the list, there wouldn't have been a system to fix. Oversizing the battery and undersizing the inverter was a £2,140 mistake. Cutting the compatibility check entirely would have cost me a client and a referral chain.
That's the thing about what is a hybrid solar system in practice: the customer never sees the compatibility spreadsheet or the peak-load audit. They see whether the battery delivers what they were promised. The spec sheet is invisible; the outcome isn't.
The Short Version
A 3 kW hybrid inverter with 4× Pylontech US5000 (19.2 kWh nominal) is not "a well-sized system." It's a system that can deliver 3 kW for about five hours, then 3 kW again after a recharge—assuming the BMS handshake works, which is a separate question.
Size the inverter to peak load, not to battery capacity. Read the compatibility list twice. And measure before you quote, not after.
