The utility-scale solar secondary market exists because new-module lead times sit at 6–12 months for tier-1 product, and project cancellations + repower programs constantly liberate inventory in the multi-MW range. That's a real procurement option. It's also a category where the wrong purchase can cost you 10–30% of nameplate output without ever showing up on the data sheet. Every used module pallet, central inverter, or tracker row needs structured evaluation before you sign the bill of sale.
This guide walks through the evaluation sequence Refound recommends for any used solar equipment in the MW-block range. Use it as a checklist before bidding. Use it again as the framework for an on-site walk-through if the inventory clears the desk review.
Before you bid
The first decision is whether the inventory fundamentally fits your project. Mismatched module specs cost more in re-engineering than the modules themselves are worth. Confirm the following before you spend a minute on test data:
- Voc and Isc compatibility. Module open-circuit voltage at low temperature must not exceed inverter MPPT max + 1500 V system voltage limits. Verify against your project's coldest expected ambient and string sizing math.
- Wattage class within ±5% of design. Modules from a 545 W lot can substitute for a 550 W design with minor performance impact. Modules from a 400 W lot cannot — your tracker rows, combiners, and inverter loading were sized to a different watt-density.
- Frame + connector compatibility. EVO/MC4-EVO2 connectors don't seal reliably with original MC4 connectors. Frame profiles vary between OEMs in ways that affect tracker clamping. Confirm both before purchase.
- Bifaciality factor. If your tracker design assumes 70% bifacial gain at the module rear, you can't substitute monofacial modules. Bifacial module data sheets cite both BSE (bifacial efficiency) and rear-side response under standard test conditions.
- UL listing. UL 1703 (legacy) and UL 61730 (current) modules can both deploy in North America. Some AHJs require specific listings — verify with your jurisdiction before assuming substitutability.
Module evaluation
Get a clear photograph of the module nameplate label plus the original OEM data sheet for the exact model + revision. Cross-check the two yourself. The serial number prefix typically encodes manufacture date and factory line — useful for verifying the lot wasn't a mid-production-revision batch.
What to confirm on the nameplate:
- Manufacturer + model number + revision
- Manufacture date (older modules see more degradation)
- Serial number range across the pallet stack
- Wattage class (Pmax) at STC
- Voc, Isc, Vmp, Imp at STC
- Temperature coefficient of Pmax (typically -0.30 to -0.40 %/°C)
- Fire rating + UL listing reference
- Maximum system voltage (1000 V or 1500 V)
Flash test + IV curves
Flash testing measures actual Pmax against nameplate Pmax under controlled-light-pulse conditions, simulating STC. Used-module inventory should arrive with per-pallet flash data at minimum, ideally with per-module data for revenue-grade procurements. Read flash data with these standards in mind:
- Mean Pmax vs. nameplate. 0–3% degradation indicates near-new condition (fresh OEM excess). 3–7% indicates 1–4 years of operating life. 7–15% suggests longer service or environmental stress. Above 15% requires explanation.
- Standard deviation across the pallet. Tight std dev (under 1% of mean) suggests the modules came from a single project + factory lot. High std dev (3%+) suggests mixed inventory or batch-level issues.
- Outlier rate. Flag any module more than 2σ below mean — those are typically lightning-strike candidates, transport-damaged units, or manufacturing defects worth pricing in.
- IV curve shape. A clean smooth curve is a healthy module. Curve "knees" or step discontinuities indicate cell-string failures, broken bypass diodes, or partial shading at flash-test time. Demand the curve plots, not just the Pmax number.
PID resistance
Potential-induced degradation can knock 10–30% off nameplate output if modules sat for years in a high-voltage string with grounding asymmetry. Flash testing alone won't always catch PID — affected modules can flash at full power then lose 20% over the next 6 months in service.
For modules from operating projects, demand:
- IEC 62804 PID-resistance certification on the original module model
- String-level performance ratio data from the source project, with any underperformer-string root cause analysis
- For modules pulled from known-PID projects: a recovered-from-PID test result (typically 96 hours at -1500 V at 60 °C / 85% RH) showing recovery to within 5% of nameplate
Tier-1 manufacturers (Jinko, Trina, JA Solar, Canadian Solar, LONGi, First Solar, Heliene) publish PID-resistance certification publicly. Tier-2/3 modules vary widely. If the seller can't produce IEC 62804 documentation, treat the modules as PID-risk and price accordingly.
Delamination + EL imaging
Visual inspection catches the largest defects (cracked glass, severe delamination, blackened bypass diodes) but misses the cellular-level issues that drive long-term degradation. Electroluminescence (EL) imaging is the standard tool for modular reuse evaluation:
- EL imaging is a low-current bias test that lights up cells in proportion to their electrical health. Microcracks, finger-line breaks, and shunted cells show up as dark regions invisible to the naked eye.
- For revenue-grade procurements, demand per-module EL imaging on a sample of at least 5% of the pallet stack (or 100 modules, whichever is greater).
- Acceptable: minor finger-line breaks, occasional single-cell dark regions in < 5% of cells. Reject: cell-level cracks crossing busbars, multi-cell dark regions, signs of moisture ingress at the frame edges.
Central inverters
Modern central inverters (1–5 MW per unit, 1500 Vdc class) carry 20–25 year design lives. The major capital cost is the cabinet plus the IGBT power stage. A 1 MW central with 8–10 years of service and a documented IGBT replacement is often a strong value buy at 30–50% of new cost. The key questions are:
- Service hours + IGBT replacement history. Each major IGBT replacement resets the wear clock. Demand the service log going back to commissioning.
- Cooling system type + leak history. Liquid-cooled designs (most modern utility centrals) have heat exchangers that degrade with cycling. Air-cooled designs are simpler but less efficient. Confirm any history of coolant leaks or thermal shutdowns.
- Firmware version + grid-support capability. Post-2022 interconnections in most US ISO regions require UL 1741-SB grid-support functions (frequency-watt, volt-watt, smart-curtailment). Older inverters may not have firmware-upgradeable controllers — confirm before purchase.
- Available spares. Some discontinued inverter lines (older Solectria, older AE-Series) have constrained spare parts inventory. A used inverter without a parts pipeline is a higher-risk purchase than the lower price suggests.
String inverters
Three-phase string inverters (25–350 kW) are increasingly the architecture of choice for new utility builds. The secondary market is growing as early adopters refresh from 1000 V class to 1500 V class systems. Evaluation focuses on:
- Hours-run plus cycle-count from the original SCADA log
- DC + AC connector condition (the most common failure mode)
- Internal-fan health if the unit is fan-cooled
- Firmware update history and current revision
- Original installer + commissioning report (ideally with the first 30 days of operating data)
String inverters are physically smaller and easier to swap, so the freight and rigging risk profile is much lower than central. Most string inverters in the secondary market come from project upgrades or commissioning errors rather than failures.
Trackers + racking
Single-axis trackers are the dominant utility-scale racking architecture. Tracker rows can outlive the modules they support — a 25-year-old tracker line from a repower project, with refreshed bearings and motor units, can deploy on a new build. Key evaluation points:
- OEM model number + design wind class. Trackers are sized to specific wind speed + soils combinations from the original project. Site-relocation usually requires a structural review.
- Torque tube length + section count. Consistent across the lot? Cut sections from the original project lengths add labor at install.
- Bearing condition. Sealed bearings have limited service lives. Demand a sample disassembly + inspection on at least one bearing pillar per row.
- Motor unit + controller compatibility. Nextracker NX Horizon controllers and Array Technologies DuraTrack controllers each have specific firmware revisions and self-powered modes. Mixing generations across a project doesn't always work cleanly.
- Pile + foundation hardware. Driven piles are usually project-specific length and section. Helical anchors are more reusable but have soil-condition constraints.
Freight realities
MW-block module orders are freight-heavy. Plan accordingly:
- Roughly 1.6 MW of standard 550 W modules ≈ 18 standard truckloads (≈ 90 pallets)
- Pallet weight: 700–900 kg for crystalline silicon, 600–800 kg for thin-film
- Truckload max for highway transport: ~22 t (subject to per-state weight limits)
- Forklift requirements at delivery: 5,000+ lb capacity for full-pallet handling, 8,000+ lb for double-stacked
- Weather + storage planning: modules in original packaging tolerate 12+ months of outdoor storage on pallets; without the OEM crate, exposed modules can suffer micro-cracking from thermal cycling within months
For deliveries inside 500 mi of the supply origin, FOB origin with the buyer's freight broker is typically cheapest. Longer distances often favor seller-arranged dedicated freight where the seller handles loading + insurance.
Warranty transfer
Tier-1 module OEMs typically allow warranty transfer to a second buyer if the original purchase + serial-number range is documented and the modules haven't been deployed-and-removed. The warranty transfer details vary by OEM:
- Jinko, Trina, JA Solar, Canadian Solar: Transferable with original purchase invoice + transfer notification. Typically 25 years performance, 12 years product.
- First Solar (CdTe): Transferable but with stricter documentation. Original purchase invoice + commissioning records required.
- Tier-2/3 manufacturers: Vary widely. Some require manufacturer notification within 30 days of transfer. Some void on resale entirely.
Inverters are typically warranted 5–10 years from original purchase, with optional extended warranties to 25 years. Most inverter warranties don't transfer cleanly without the OEM's specific approval.
Common red flags
- Vague flash-test data. A seller who can't produce per-pallet (or per-module) flash data is selling untested inventory at full-test pricing.
- "Removed during operation." Modules pulled from a working project for "performance reasons" are PID candidates until proven otherwise.
- Frame discoloration or jacket cracking. Indicates UV degradation; common on modules stored uncovered for years.
- Mixed serial-number lots in a single pallet. Suggests inventory consolidation from multiple sources; harder to assess as a coherent lot.
- Inverter cabinet condensation marks. Watermarks inside the cabinet indicate gasket failure + moisture ingress; expect downstream component issues.
- Trackers without OEM service logs. Unmaintained trackers from old projects often have bearing seizures + motor unit failures that weren't visible at decommissioning.
Pre-bid checklist
- OEM data sheets matched to nameplate revision
- Per-pallet (or per-module) flash test data with IV curves
- EL imaging on a representative sample
- IEC 62804 PID-resistance certification
- String-level operating-history data (for in-service modules)
- Original purchase invoice + warranty transfer paperwork
- Pallet weights, dimensions, and freight quote
- Photos of the laydown yard / storage conditions
- For inverters: SCADA log export, IGBT replacement history, firmware version
- For trackers: OEM model + revision, design wind class, sample bearing inspection
If the seller can produce all of the above and the documentation is consistent, you have a defensible procurement. If any line item is missing, the price discount should reflect the residual risk.