I review solar equipment before it reaches customers. Inverters, batteries, EV chargers, and everything that connects them. Over the past four years, that's meant reviewing 200+ unique product configurations annually, from single-unit residential kits to 50,000-unit commercial rollouts. And here's a number I don't need to look up: I reject small orders at a higher rate than large ones.
That's not a flex. It's a warning.
I believe the solar industry treats "small" as permission to skip quality steps that keep people safe. That's backwards—and it's dangerous.
It's not that small customers are sloppy. More often than not, they're more careful than procurement teams because it's their own money on the line. The problem is that vendors assume a small order won't be audited. They cut corners on the things that don't show up in a product photo—the earthing conductor, the BMS specification, the terminal torque spec. Those are exactly the things that start fires. I've watched a $22,000 rework happen because someone decided a grounding detail was "optional" on a small batch. That cost came straight out of somebody's margin.
Earthing Requirements Are Where Corners Get Cut First
When I audit a hybrid inverter installation, the first thing I check is the earthing. For a Deye hybrid inverter, the earthing requirements are documented in the installation manual: conductor size, AC and DC side bonding, and the conditions for neutral-earth bonding in backup mode. These specifications exist because people have been hurt.
What I see in small-batch installs, more often than I'd like, is a step down in conductor size. The conversation usually goes like this: "It's only a 5 kW system. Does it really need a 10 mm² earth?"
Yes. It does. The inverter doesn't care about the project size. The fault current cares about the conductor. Period.
I learned that lesson the hard way. I once assumed two vendors' "same specifications" meant identical results—same material, same dimensions, same markings. Didn't verify. Turned out each had a slightly different interpretation of what "same" meant. Now I measure the first unit from every batch. Every time.
In Q1 2024, we received a batch of units where the earthing stud dimension was visibly off against the approved drawing. The vendor argued it was "within industry tolerance." It wasn't. We rejected the batch, and they reworked it at their own cost. That clause is now in every contract I write.
The smaller the order, the more likely a vendor will assume nobody checks. That's exactly why I do. It's not about being strict for the sake of it—the test takes five minutes with a caliper and a spec sheet. There's no excuse for skipping it.
Deye Inverter Features: What I Actually Check at Acceptance
People often ask me which deye inverter features matter most during acceptance testing. They expect something technical—maximum PV input voltage, peak efficiency, that kind of thing. Those matter, but they're not what I look at first.
- Earthing and bonding points—are they specified consistently across the whole product family, including the entry-level models?
- DC and AC terminal ratings—do they match the datasheet on every variant, not just the flagship?
- Monitoring and communication—does the protocol behave the same on a 5-unit order as it does on a 500-unit one?
In my experience, Deye has been fairly consistent on these points compared to other vendors I've audited. But "fairly consistent" is not the same as "verified." I still measure every grounding point, every terminal, every batch. The day I stop checking is the day I start approving failures.
LiFePO4 Battery Fire Hazard: The Chemistry Isn't the Weak Link
Now the topic that generates the most emails: LiFePO4 battery fire hazard.
I'm not a cell chemist, so I can't speak to the electrochemistry at that level. What I can tell you from years of reviewing test reports and audit data is this: LiFePO4 is genuinely less prone to thermal runaway than NMC. But "less prone" is not "immune."
In the battery incidents I've studied, the root cause was rarely the cell chemistry. It was almost always the battery management system being under-specified for the application, or a terminal connection that worked loose under thermal cycling, or a BMS that didn't isolate a fault quickly enough. When I see a small order with an "I'll save money on the BMS" line item, I don't approve it. Simple.
The cells were never the weak link. The management and connection quality were. And that's exactly the kind of thing that shows up in inspection—if you bother to do one. We run the same acceptance tests on a single battery pack as we do on a full container load. The pass/fail line doesn't move.
Solar Generator for Electric Car: Same Load, Same Bar
One of the fastest-growing categories I review is the "solar generator for electric car" setup. The name makes it sound simple: plug in the sun, charge the car. What you're actually buying is a portable battery, an inverter, and a charge controller in one enclosure. And that enclosure has to handle a load profile completely different from powering a campsite.
EV charging draws sustained current for hours. That sustained load tests the inverter's thermal design, the battery's discharge limits, and every terminal connection in the path. I've tested solar generators that handled a 2 kW load for 20 minutes without issue. Pushed to two hours, the terminal temperature crept past the spec. They didn't fail. But they were on the edge—and edges get closer with age.
Is a solar generator a bad idea for charging an EV? I don't think so. But the quality bar has to be the same as any grid-connected charger: proper grounding, proper cable sizing, and a BMS that matches the load. If a vendor wants to lower that bar because the order is small, I'm not interested.
"It's a Small System" Is Not a Defect Allowance
The most common pushback I hear: "It's just a small system. What's the worst that could happen?" Or: "It's a test order. Minor variations are acceptable."
Let me answer with a question. If you've ever wondered what a tire monitoring system light means on your dashboard, the short answer is: it means something is wrong with a tire. Nobody responds by checking the light bulb. You check the tire pressure. Ignoring the warning because "it's probably fine" is how a blowout happens at 120 km/h.
A BMS alarm is the same. An earthing fault indicator is the same. The warning isn't the problem—the thing it's pointing at is. And the size of the system doesn't change what the indicator means.
And regarding test orders: I understand the business logic, but I disagree. A test order becomes a reference order. Customers remember the vendor who checked everything, and they remember the one who didn't. When I was starting out, the vendors who treated my $200 orders seriously are the ones I still use for $20,000 orders today.
Small doesn't mean unimportant. It means potential.
Quality at Any Scale
So let me restate the position clearly: the number of units in an order doesn't change the safety requirements of a single unit.
Per FTC advertising guidance (ftc.gov), we don't make claims like "zero fire risk" or "safest battery ever." A claim is a commitment. A spec is a commitment. Neither should change based on the quantity ordered.
I've lost orders because I wouldn't fudge a test result or accept a "close enough" conductor size. I'm okay with that trade. I've also seen vendors who grumble about small orders become the first call when those customers scale up. The ones who don't? They're usually still grumbling.
A small order done right is how a big order starts. A small order done wrong is how someone's house catches fire.
The choice is simple. I'll take the former, every time.