-
Step 1: Lock Your Spec Sheet Before You Talk to Anyone
-
Step 2: Verify Factory Capacity, Not Just Company Size
-
Step 3: Request Same-Batch EL Images Before Shipment
-
Step 4: Ask About the Warranty Process, Not the Warranty Period
-
Step 5: Calculate Total Cost, Not Price Per Watt
-
Step 6: Negotiate the Rejection Clause Before the Trial Order
-
Step 7: Specify PID and Hot-Spot Testing in the Contract
-
Three Mistakes I See Most Often
-
When This Checklist Doesn't Apply
This checklist is for procurement managers, EPC contractors, and distributors placing wholesale PV module orders—typically 1 MW and up. It's not a primer on what solar panels are. It assumes you already know the difference between mono PERC and TOPCon, and you're now standing at the point where a supplier has sent you a quote and you need to figure out whether to move forward.
I've worked on the quality side of module sourcing for over four years. My role involves reviewing supplier deliverables before they clear our procurement gate—roughly 40 to 60 container-level orders a year, plus smaller trial batches. In 2023, I rejected somewhere between 15% and 20% of first-article shipments. Not because the modules were broken, but because what arrived didn't match what was quoted. That gap is the entire reason this checklist exists.
Here are seven steps. They're ordered by sequence, not by importance. Every one of them has a failure story behind it.
Step 1: Lock Your Spec Sheet Before You Talk to Anyone
Most quality problems in wholesale PV sourcing start before the first email.
The typical mistake: a buyer collects three or four quotes, then compares them on price per watt. But the quotes aren't quoting the same thing. One supplier quotes 550W TOPCon with a 144-cell layout. Another quotes 545W mono PERC. A third quotes 550W but with a different junction box and a 1,100mm cable length instead of 1,300mm. The buyer picks the cheapest per-watt number and discovers the mismatch two months later when the mounting hardware doesn't align.
Before you send a single RFQ, write down what you actually need:
- Module wattage and tolerance range (e.g., 0/+5W, not ±5W)
- Cell technology (TOPCon, HJT, PERC—be explicit, don't let the supplier interpret)
- Dimensions and frame thickness
- Junction box IP rating and cable length
- Connector type (MC4 compatible or proprietary?)
- Certifications required for your destination market (IEC 61215, IEC 61730, UL 61730 if you're shipping to North America)
What most people don't realize is that suppliers will quote to the loosest interpretation of your spec if you leave room for it. They're not being dishonest—they're optimizing for their own inventory. If your spec says "550W ±5W" they can ship you 545W modules and be technically compliant. If your spec says "550W, 0/+5W" they can't.
Step 2: Verify Factory Capacity, Not Just Company Size
I want to say this is obvious, but I've seen it missed at companies that should know better.
When a supplier says "we have 5 GW annual capacity," that number refers to the company's total production across all lines and all product types. It does not mean they can deliver your specific module configuration at the volume you need within your timeline.
Ask specifically:
- Which production line will run my order?
- What is the monthly output of that specific line for my module type?
- How many other orders are currently queued on that line?
- Can I see the line's last three months of output data?
If they can't answer these, you're talking to a trading company, not a factory—or you're talking to a factory that's brokering your order out to a third-party line. That's not automatically disqualifying, but it changes your risk profile significantly. A brokered order means you have less visibility into the actual manufacturing conditions.
Here's something vendors won't always volunteer in the first conversation: many "OEM" arrangements in the PV space involve one factory producing unbranded modules that three or four different companies sell under their own labels. The modules might be fine. But you need to know whether your supplier controls the line or is just reselling.
Step 3: Request Same-Batch EL Images Before Shipment
This is the step most buyers skip, and it's the one that catches the most problems.
Electroluminescence (EL) imaging shows micro-cracks, cell mismatch, and soldering defects that aren't visible under normal light. Any serious factory runs EL testing on finished modules. The question is whether they'll share the results with you before the container leaves.
Ask for:
- EL images from the actual production batch your order is drawn from (not a sample batch from six months ago)
- Randomly selected units—at minimum 10 modules out of every 1,000
- Images taken after lamination, not before
- Consistent imaging parameters (same exposure, same camera angle) so you can compare across the batch
In 2022, we received a batch of 2,400 modules where the EL images showed a recurring soldering anomaly—a small dark spot near the busbar on about 8% of units. Not catastrophic. But it indicated a calibration issue on the stringer that the factory hadn't caught. We flagged it, they re-tested, and the affected modules were pulled before shipping. Without the EL review, those modules would have gone into a project in Uzbekistan with a hot-spot risk down the line.
If a supplier resists sharing EL images or says "that's internal quality data," treat that as a signal. It doesn't mean the modules are bad. It means you won't know either way until it's too late to change anything.
Step 4: Ask About the Warranty Process, Not the Warranty Period
Every supplier offers a 25-year performance warranty and a 12-to-15-year product warranty. Those numbers are standard and nearly identical across the industry. They tell you almost nothing about what happens when something goes wrong.
The questions that matter:
- If a module fails in year three at a site in Indonesia, who pays for the removal, shipping, and replacement?
- What's the typical turnaround time from claim submission to replacement delivery?
- Is the warranty backed by the manufacturer or by the seller? (If you're buying through a distributor, the distinction matters.)
- Has the company ever denied a warranty claim? What were the circumstances?
To be fair, most suppliers will give you the standard answer: "We honor all valid claims." The value is in the follow-up questions. Ask for a redacted example of a resolved claim—timeline, logistics, resolution. If they can produce one, that's a good sign. If they can't, it likely means few claims have ever been filed through them, which could mean either excellent quality or a customer base that gave up on the claims process.
Step 5: Calculate Total Cost, Not Price Per Watt
Price per watt is the number everyone compares. It's also the number that lies most often.
Total installed cost includes:
- Module price per watt (including any run-off from wattage tolerance)
- Freight cost (varies significantly by port of origin and destination)
- Import duties and tariffs (these can swing 15-30% depending on the trade lane)
- Inland logistics from port to project site
- Any re-testing or certification you need to do on arrival
- Potential replacement cost if the first batch has a quality issue (budget 2-5% of order value as a contingency)
I've seen a $0.012/W difference in module price disappear entirely once freight and duties were factored in. The cheaper supplier was shipping from a different port with less favorable routing, and the landed cost ended up $0.003/W higher than the "more expensive" option.
Also worth checking: whether the quoted wattage is nameplate or flash-test minimum. Some suppliers quote nameplate wattage (the nominal rating) while others quote the flash-test minimum—the lowest wattage a module in that bin will actually produce. A 550W nameplate module might flash-test at 545W minimum. Over a 10 MW order, that 5W gap adds up.
Step 6: Negotiate the Rejection Clause Before the Trial Order
This is the step most buyers skip, and it's the one that costs the most when skipped.
Everyone does a trial order—a few modules, maybe a pallet, to check quality. That's fine. But the trial order itself isn't the point. The point is what happens after the trial order arrives and something is wrong.
Before you place even a sample order, have this conversation:
"If the delivered modules fail to meet the agreed specifications—on wattage, on dimensions, on EL results, on flash-test data—what is the process? Who pays for return shipping? What's the timeline for replacement? What's the threshold for a full-batch rejection versus a partial credit?"
If the supplier can't answer these questions clearly, you're not ready to place an order. Not because they're going to screw you over, but because ambiguity in the rejection process always favors the party who has the goods.
Get the rejection terms in writing before the PO. I know that sounds bureaucratic. It saves arguments later.
Step 7: Specify PID and Hot-Spot Testing in the Contract
Potential-induced degradation (PID) and hot-spot testing aren't always standard in factory QC reports, especially for suppliers targeting price-sensitive markets. But both are common failure modes in the field, and both are testable before shipment.
- PID testing: Ask for results at 85°C/85% RH, 96 hours, 1,000V system bias. IEC 62804 is the reference standard. Look for less than 5% power degradation.
- Hot-spot testing: IEC 61215 section 10.11. Ask for the maximum temperature recorded on the hottest cell.
Neither test adds significant cost if requested upfront. Both become expensive problems if they surface after installation. A PID-affected module in a utility-scale array can lose 10-20% output over two to three years—and by then, the warranty claim process is much harder to navigate.
Three Mistakes I See Most Often
Comparing quotes without normalizing specs. If two quotes have different wattages, different tolerances, or different certifications, the price comparison is meaningless. Build a comparison spreadsheet. It takes 30 minutes and can save you a six-figure mistake.
Skipping the pre-shipment inspection. Third-party inspection (SGS, Bureau Veritas, TÜV) costs typically $1,000-3,000 depending on scope and location. On a $500,000 order, that's 0.2-0.6%. On a rejected order, it's the cheapest insurance you'll ever buy.
Treating the first order as the relationship. The first order tells you about the factory's capabilities. It doesn't tell you about their consistency. Consistency shows up in orders two through five. Build your vendor evaluation over at least three orders before committing to volume contracts.
When This Checklist Doesn't Apply
I want to be clear about where this approach doesn't fit.
If you're buying fewer than 50 modules for a single-site commercial installation, this level of supplier verification is probably overkill. You're better served working with a regional distributor who already stocks modules and can deliver in days, not weeks. The cost-per-watt will be higher, but the total cost—including your time—will likely be lower.
If your project has a hard deadline under four weeks and you haven't identified a supplier yet, this checklist won't help you. You'll need to buy from available stock, not from a factory production run. That's a different sourcing strategy entirely.
And if you're sourcing modules as part of a larger integrated project—say, a utility-scale development where the module supplier is also involved in financing or EPC coordination—the evaluation criteria shift. At that point, you're not just buying modules. You're evaluating a partner, and the checklist above becomes one input among many. For reference, developers like ACWA Power operate at that integrated level—sourcing modules across multiple global projects while also managing power generation and desalination infrastructure. That's a different scale of procurement, and the supplier evaluation process reflects it.
The bottom line: this checklist is for buyers who have time to do it right and volume that justifies the effort. For everyone else, find a supplier you trust, start small, and document everything. The checklist can wait until order number two.
