2026-09-17
How to Evaluate PV Module Manufacturers: A 7-Step Checklist Before You Sign
A quality compliance manager's practical 7-step checklist for vetting PV module manufacturers — warranty fine print, IEC certificate scope, BOM verification, and pre-shipment EL testing.
-
What this checklist covers
-
Step 1 — Read the warranty as a contract, not a headline
-
Step 2 — Match the IEC certificate scope to the factory and the product
-
Step 3 — Reconcile the datasheet against flash test data
-
Step 4 — Verify the bill of materials (the step everyone skips)
-
Step 5 — Ask for the manufacturing quality system, not just the product certificate
-
Step 6 — Run pre-shipment EL and IV sampling
-
Step 7 — Lock traceability and serialization into the contract
-
The short list of things that go wrong
What this checklist covers
I'm the quality and brand compliance manager at a mid-size solar distributor in Northern Europe. Every module order we place — roughly 35 to 40 containers a year — passes through me before the purchase order goes out, and again before the shipping documents are released. In 2024 I rejected 4 of 11 first-article shipments. In 2023, 6 of 14.
This is the checklist I run. Seven steps. It works for any PV module manufacturer you're evaluating — astronergy solar modules, any OEM or private label program you're setting up, any brand. If you're a distributor, an EPC, or a project developer buying at container scale, it applies to you.
Two hours of work if the vendor sends everything on the first request. They usually don't.
None of it is exotic. All of it is checkable. The problem is that it's boring, and boring is exactly where the problems hide.
Step 1 — Read the warranty as a contract, not a headline
The datasheet says "25-year performance warranty" in a nice font. The actual warranty document — the PDF nobody opens — defines what that means, who honors it, and what voids it.
Three things to find:
Product warranty vs. performance warranty. These are separate. The product warranty (materials and workmanship) is typically 12 to 15 years on standard modules, sometimes 25 on premium lines. The performance warranty is the power-output guarantee. They expire under different conditions, and they're enforced differently.
The degradation curve. Look at the year-one cap and the annual rate after that. Modern monocrystalline modules are usually specified at 1% to 2% in year one, then 0.4% to 0.55% per year. Different cell technologies land in different places on that range — that's not a good-or-bad thing, it's just something the warranty should reflect honestly rather than round off.
The exclusions, and who pays shipping. If a module fails in year 8 and the warranty covers the module but not the labor to pull it off a rooftop and replace it, you've covered maybe 15% of the actual cost. I'd rather pay 3% more per module for a warranty that covers removal and reinstallation than save it and find out later. Ask the question before you sign, not after.
Step 2 — Match the IEC certificate scope to the factory and the product
Everyone sends you the IEC certificates. Almost nobody checks the scope.
What you want on the certificate:
- The legal manufacturer name (not the trading company)
- The specific factory address — a certificate for a headquarters doesn't cover a plant 600 km away
- The exact module model numbers you're buying — if the model on your PO isn't listed, it isn't covered
- The bill of materials the certificate was issued against
IEC 61215-1:2021 and 61215-2:2021 (design qualification — essentially "will this survive hail, thermal cycling, humidity freeze, and mechanical load") and IEC 61730-1:2023 / 61730-2:2023 (safety) are the baseline. Selling into North America, add UL 61730. Into the EU, add CE plus the relevant harmonized standards.
Then there's the extended set — optional, and whether you need them depends on where the modules are going:
- IEC 61701 — salt mist corrosion. Non-negotiable for coastal or offshore-adjacent sites.
- IEC 62716 — ammonia resistance. Agricultural installations, especially livestock barns.
- IEC 60068-2-68 — sand and dust. Desert and arid sites.
- IEC 62804 — PID resistance at system voltage.
- IEC 62782 — dynamic mechanical load. High-wind or hail-prone regions.
Check that the issuing lab is accredited under ISO/IEC 17025, and check the expiration date, not just the issue date. I've seen expired certificates presented as current (not that anyone ever mentioned it).
Step 3 — Reconcile the datasheet against flash test data
The datasheet is a marketing document. The flash test report is evidence.
Nameplate tolerance is usually expressed as 0/+5 W, meaning the module is supposed to output at or above its rated power. Ask for a flash test file from an actual production batch — not a single "golden sample." Every module gets flash-tested at the factory. The question is whether they'll share the file.
Power distribution. If the datasheet says 415 W and the batch averages 418 W, fine. If it averages 412 W against a 0/+5 tolerance, you have a problem. Remember measurement uncertainty is typically around ±3% — a 415 W module could read anywhere from 403 to 427 W depending on the flasher calibration. Which brings us to the next point.
Flasher calibration. A flash simulator that hasn't been calibrated recently is a random number generator with a screen. Ask for the calibration certificate and the date.
Temperature coefficient of Pmax. Usually between −0.29 and −0.35 %/°C for modern modules. A worse coefficient means more power lost on a hot roof, which matters more in some markets than others.
Low-light performance. The datasheet gives STC numbers (1000 W/m², 25 °C). Real roofs spend most of their time below 800 W/m². Ask for relative efficiency at 200 W/m² — it's part of the IEC 61853-1 testing framework, and vendors who have it tend to send it quickly.
Step 4 — Verify the bill of materials (the step everyone skips)
This is the one. If you skip everything else, don't skip this.
Two modules can look identical on a datasheet — same power class, same dimensions, same efficiency — and have completely different bills of materials. And the BOM is what determines how the module behaves in year twelve.
Ask for the formal BOM:
- Cell supplier and cell technology
- Encapsulant: EVA or POE, and which layer
- Front and back glass: thickness, coated or not
- Backsheet type, or glass-glass construction
- Frame alloy and profile thickness
- Junction box: IP rating, diode configuration
- Connectors: original MC4 or a compatible equivalent
- Cable length and cross-section
When I put two vendor files side by side in late 2023 — same nameplate, same price bracket, one using a POE encapsulant and the other a standard EVA — I finally understood why the field failure rates we'd seen didn't track with the spec sheets at all. The spec sheets were identical. The modules weren't.
Two things to insist on: a BOM change notification clause in the contract, and the right to reject a shipment if the BOM changes without notice. Manufacturers change cell suppliers all the time. Sometimes for good reasons. You still need to know.
Step 5 — Ask for the manufacturing quality system, not just the product certificate
A type-approval certificate tells you one module passed a test once. It says nothing about whether the thousandth module off the line will be the same.
For that, look at the quality management system. The baseline is ISO 9001:2015. The PV-specific layer is IEC TS 62941 — the module-manufacturing quality guideline covering incoming material control, in-process inspection, calibration, and BOM change control.
Questions worth asking:
- What percentage of modules get electroluminescence inspection, and at what stage?
- Is every module flash-tested? (It should be.)
- How often is the flash simulator calibrated, and by whom?
- What's the incoming inspection protocol for cells?
- Has the factory been audited by an independent third party in the past 24 months?
If the answer to any of these is a vague "yes, we have a system," ask for the document. It exists or it doesn't.
Step 6 — Run pre-shipment EL and IV sampling
Do this at the factory, before the container leaves. Once it's on the water, your leverage is gone and you're negotiating from a different position.
Pull samples from the actual production run — not from a sample room. Have a third-party lab with ISO/IEC 17025 accreditation run:
Electroluminescence (EL) imaging. EL shows microcracks, broken fingers, bad solder joints, and cell defects that are invisible to the eye and to a handheld meter. Pull them at random out of the pallet. If the vendor only lets you inspect modules they picked, that's an answer in itself.
IV curve at STC, plus low irradiance. The STC number tells you whether the module meets nameplate. The low-light number tells you how it behaves for most of its actual operating life.
Third-party testing usually runs into the low four figures per shipment depending on how many samples you pull. On a full container, that's rounding error. Compared to a field failure requiring a truck roll per module, it's free.
Step 7 — Lock traceability and serialization into the contract
Every module should have a unique serial number, and the serials in your shipment should map to the flash test data.
Why does this matter? Because in year four, when eight modules on one rooftop underperform the other twenty, you need to prove which production batch they came from. Without that mapping, your claim is a conversation. With it, your claim is a document.
Ask for:
- Serial number ranges by pallet and by container
- The flash test file mapped to those serials, delivered within an agreed number of days after shipment
- Confirmation that serials are durable and readable — not printed with ink that fades in six months
If you're buying under an OEM or private label agreement, there's a separate set of terms to nail down: who holds the certificate, who issues the warranty to the end customer, and who is the manufacturer of record. I've seen programs where the answer to all three was "we'll figure it out." That's not a program. That's a liability with a logo on it.
The short list of things that go wrong
Accepting "industry standard" as a spec. Industry standard is not a number. Ask for the number.
Checking the certificate logo but not the scope. A certificate for the parent company doesn't cover a module assembled at a different plant, under a different BOM, with a different model number.
Buying on price alone. The lowest price is usually a different BOM wearing the same nameplate. Not automatically worse — but you should know which trade-off you're accepting.
Relying on the vendor's own lab results. Their lab. Their flasher. Their numbers. Nothing wrong with it, but it isn't independent evidence.
Skipping the BOM conversation because the datasheet looks fine. The datasheet is designed to look fine.
Not documenting the flash file handoff. "We'll send it after shipment" is how you end up with no data and no lever.
I have mixed feelings about how much of this belongs on the buyer's plate. On one hand, a good manufacturer should make most of it unnecessary. On the other, I've seen what happens when you assume that. The batch that made me write this arrived in March 2023 (this was back when we were sourcing through a program we've since wound down). I want to say it was 612 modules, but don't quote me on that. Wrong encapsulant in half of them.
So I run the checklist. Every time. Seven steps, two hours, and a much quieter year.
Written by Farah Qureshi