Buying ten solar cells and buying a thousand are not the same transaction with a different number on it. At ten, you inspect every cell yourself and any bad one is a minor annoyance. At a thousand, you cannot inspect them all, the bad ones hide inside the good ones, and the failure does not show up until the array is already built.
This is what actually changes as the quantity climbs, and what to specify before money moves.
The thing that breaks first: bin spread
At small quantities, cell-to-cell variation barely registers. At volume it becomes the dominant factor in whether your array hits its design number.
Cells wired in series all carry the same current. A cell with lower current capability limits every other cell in that string. So the number that matters is not the average of your lot — it is how tightly the lot clusters, and where the bottom of the distribution sits.
| Lot character | What happens in a 40-cell string |
|---|---|
| Tight bin, narrow spread | String output lands close to the design target |
| Wide spread, same average | String pulls toward the weakest cells — the average is misleading |
| Mixed bins, undocumented | Output is unpredictable per string; two identical panels test differently |
That last row is the one that costs real money, because it is not a uniform shortfall you can design around. It is variance between finished units, which means your product's spec sheet becomes a guess.
What to specify: ask for the bin distribution of the actual lot, not the nominal wattage. If you are stringing in series, say so explicitly and ask for matched bins across the shipment — including across the reorder, if you expect one.
Tolerance stacking: small numbers that compound
Physical tolerance behaves the same way. A ±0.3 mm cut tolerance is irrelevant on one cell. Lay thirty cut cells edge to edge in a fixture sized for the nominal dimension and the accumulated error can exceed the space you left.
This bites hardest on cut cells, which is exactly what most weight- or shape-constrained builds use. Two questions to ask before ordering:
- What is the dimensional tolerance per cell, and is it stated per cut edge or across the whole cell?
- Is the cut laser or mechanical? Laser cutting holds tighter tolerances and leaves less edge damage; mechanical scribing is cheaper and shows it at volume. The custom cut service guide covers the geometric limits that the rear busbar layout imposes.
Design defence: size your fixture against worst-case accumulated tolerance, not nominal. It costs nothing at design time and saves a re-tool later.
Breakage is a line item, not an accident
Loose cells are brittle, thin, and shipped in stacks. A percentage will arrive broken no matter who ships them. Treating that as a surprise is the mistake — it should be a number in your plan.
Ask three things before the lot ships: how are cells separated within the stack; what is the outer packaging spec for the carton; and what is the agreed handling of breakage on arrival. Suppliers who ship cells regularly answer all three immediately. Suppliers who do not, cannot.
The related risk is worse because it is invisible: micro-cracks. A cell can survive transit visually intact and still carry cracks that propagate during lamination or thermal cycling. Those failures appear weeks or months later, inside a finished product, which is the most expensive place to find them.
What to specify: visual and micro-crack inspection before packing, and an agreed over-ship percentage so a few broken cells do not stall your build.
The volume tiers, and what actually changes at each
| Quantity | What dominates | What to insist on |
|---|---|---|
| Under 100 | Speed. Your own bench is the quality system | Nothing beyond fast shipping — just test what arrives |
| 100–1,000 | Consistency within the lot | Bin data, stated grade, breakage handling |
| 1,000–5,000 | Consistency and repeatability | All of the above, plus written measurement conditions and an off-spec process |
| 5,000+ recurring | Allocation across reorders | Bin matching across shipments, not just within one |
That last row is the one buyers discover late. Getting a good lot once is not hard. Getting the same lot characteristics three months later, when the fab has moved to a different production run, is the actual supply problem — and it is why recurring buyers care more about who holds allocation than about unit price.
Why the cheapest lot is often not the cheapest lot
Compare landed cost per usable watt, not price per piece. The arithmetic that matters:
- Start with unit price × quantity
- Subtract cells lost to breakage and micro-cracks
- Subtract the performance gap if bins are loose — a string that lands under target is a partial loss of the whole string
- Add the cost of schedule slip if you have to wait for replacements
A lot that is 15% cheaper per piece and arrives with a wide bin spread routinely ends up more expensive per delivered watt. This is not a sales argument — it is why experienced buyers ask for bin data before they ask for a price.
Sample first, always
Regardless of your target volume, the sequence that avoids nearly all of the above is boring and effective:
Retail price, no minimum. Test against the label on your own bench. This is where you validate the supplier's claims cheaply.
Non-standard cuts run 7–14 days, with the tooling fee refunded against your production order.
Quoted per specification — format, grade, bin tightness and quantity all move the number.
The general logic of sampling before volume, and what to actually measure when the sample lands, is covered in testing samples before a bulk order. For cells specifically, the grades guide tells you what the grade designation should mean physically, and this hands-on build guide walks through testing and wiring loose cells so you know what a healthy reading looks like.
What to send when you ask for a bulk quote
Cell format or finished dimensions · target voltage and power at the array level · quantity per batch and expected reorder interval · whether cells will be wired in series (and therefore need matched bins) · destination market, if RoHS or REACH declarations are required.
The item most often left out is the reorder interval — and it is the one that determines whether allocation needs to be held. Saying "1,000 pieces now, likely again in three months" produces a materially different answer than "1,000 pieces."
If you want to compare formats before specifying, 125 mm vs 166 mm and custom cuts covers where each makes sense, and IBC vs standard monocrystalline covers when the back-contact premium is worth paying and when it is not. Stock formats are in the cells collection.
