Engineering Reality vs. Installer Claims
The phrase "it is the same panel" is a category error, and it is the single most expensive category error in residential solar procurement. Across the SolarQna dataset — 607 verified panels, 336 inverters, 267 batteries, 359 manufacturers and 220 tracked shops in 22 markets — the observed price dispersion for a nominally identical model number routinely reaches 30–40% between two authorised distributors operating in the same metropolitan area. The installer's claim of identity collapses four distinct identities into one.
Identity 1 — the model number. A model number such as TSM-NEG19RC.20Q identifies a datasheet envelope: 670 Wp nominal, 24.8% aperture efficiency, γ_Pmax of −0.26 %/°C. It does not identify a bill of materials. Manufacturers re-binning a line, switching encapsulant suppliers, or migrating from 182 mm to 210R wafer formats will retain the model number while the physical stack changes.
Identity 2 — the factory and BOM revision. A module produced in a Tier-1 Chinese fab in Q3 2023 and one produced in Q1 2025 under the same model number may differ in cell supplier, EVA versus POE encapsulant, and frame alloy. Both pass IEC 61215. Neither is interchangeable in a warranty claim.
Identity 3 — the power bin. A 670 Wp nameplate is a bin centre. Flash-test data at ±3% tolerance means two physically identical modules can be labelled 645 Wp and 670 Wp. The 670 Wp bin carries a price premium that has nothing to do with manufacturing cost.
Identity 4 — the warranty channel. This is where the 35% spread actually lives. An authorised distributor purchase carries a registered serial number, a valid performance warranty and a local claims path. A grey-market pallet bought cross-border carries the manufacturer's warranty on paper and nothing enforceable in practice.
The dispersion is therefore not a market inefficiency to be arbitraged. It is a price signal for four different assets wearing the same label. The remainder of this study quantifies each term.
Core Physics & Quantitative Equations
Before attributing price spread to commercial factors, we must establish the physical baseline against which any module is evaluated. The thermal derate is the first-order correction.
Cell temperature (NOCT model):
T_cell = T_amb + ((NOCT − 20) / 800) × G
Power at operating temperature:
P_op = P_stc × (G / 1000) × (1 + γ_Pmax × (T_cell − 25))
Worked example. Take a 670 Wp N-type TOPCon module with γ_Pmax = −0.26 %/°C, NOCT = 44 °C, at G = 800 W/m² and T_amb = 35 °C:
T_cell = 35 + ((44 − 20) / 800) × 800 = 35 + 24 = 59 °C
ΔT = 59 − 25 = 34 °C
derate = 1 + (−0.0026 × 34) = 1 − 0.0884 = 0.9116
P_op = 670 × 0.80 × 0.9116 = 488.6 W
The module delivers 488.6 W, not 670 W. Every economic comparison must be built on P_op, not P_stc.
Levelised cost of electricity:
LCOE = ( C_0 + Σ_{t=1}^{N} O&M_t / (1+r)^t ) / ( Σ_{t=1}^{N} E_t / (1+r)^t )
E_t = Y_1 × (1 − d)^(t−1)
where C_0 is installed CAPEX per kWp, Y_1 is year-one specific yield, d is annual degradation and r is the discount rate.
Price-spread decomposition. The shop-to-shop delta for an identical model number decomposes additively:
ΔP_shop = ΔC_cell + ΔC_freight + ΔC_FX + ΔC_inventory + ΔC_channel + ΔC_margin
For a 670 Wp module moving from €0.148/Wp to €0.11/Wp, ΔP_shop = €0.038/Wp = €25.46 per module. Typical attribution in our tracked EU distributor set:
- ΔC_cell (container vintage, deflationary wafer cycle): €0.014–0.018/Wp
- ΔC_freight and pallet consolidation: €0.003–0.005/Wp
- ΔC_FX (EUR/CNY hedge timing): €0.002–0.004/Wp
- ΔC_inventory (carrying cost on aged stock): €0.004–0.006/Wp
- ΔC_channel (warranty registration and claims reserve): €0.003–0.005/Wp
- ΔC_margin (single-unit versus full-pallet): €0.004–0.008/Wp
The channel and inventory terms are the ones a buyer can actually influence. The cell-vintage term is not — it is a function of when the container was priced.
Empirical Data From Verified SolarQna Datasheets
The following table compares eight verified modules spanning the current 660–805 Wp envelope. Note that the top four entries cluster within 0.2 percentage points of efficiency while differing by up to 25% in product warranty and by a factor of two in degradation rate.
| Module | Cell Type | P_stc (Wp) | η (%) | Wp/m² | γ_Pmax (%/°C) | Product Warranty | Perf. Warranty | Degradation (%/yr) | ALMM |
|---|---|---|---|---|---|---|---|---|---|
| AIKO AIKO-A805-GRH78Dw | Back Contact | 805 | 25.10 | 250.6 | −0.26 | 15 yr | 30 yr | n/a | No |
| Maxeon SPR-BE66-GD615 | Back Contact | 615 | 24.80 | 247.6 | −0.26 | 40 yr | 40 yr | 0.25 | No |
| Trina Solar TSM-NEG19RC.20Q | N-Type TOPCon | 670 | 24.80 | 248.0 | −0.26 | 12 yr | 30 yr | 0.35 | Yes |
| JA Solar JAM66D50-670/GB | Monocrystalline | 670 | 24.80 | 248.0 | −0.26 | 12 yr | 30 yr | 0.35 | Yes |
| JinkoSolar JKM645-670N-66QL6-BDX | N-Type TOPCon | 670 | 24.80 | 248.0 | −0.26 | 12 yr | 30 yr | 0.35 | Yes |
| Sunpro Power SPDG665-BC132R12 | Back Contact | 665 | 24.62 | 246.2 | −0.26 | 25 yr | 30 yr | 0.35 | No |
| LONGi Solar LR7-72HVDF-665M | Monocrystalline | 665 | 24.60 | 246.2 | −0.26 | 15 yr | 30 yr | 0.35 | Yes |
| Vikram Solar ENERSURGE VSMDH.66.660.05 | Back Contact | 660 | 24.43 | 244.3 | −0.26 | 12 yr | 30 yr | 0.35 | Yes |
Three observations follow directly from the data. First, the 670 Wp cluster — Trina, JA Solar and JinkoSolar — is functionally indistinguishable on electrical parameters: identical efficiency, identical power density, identical temperature coefficient, identical degradation. Any price difference between them is commercial, not physical. Second, Maxeon's 40-year product warranty is a genuine outlier and is priced accordingly; it is the only module in the set where the warranty itself is a differentiated engineering asset. Third, the AIKO 805 Wp module at 250.6 Wp/m² and 25.1% efficiency represents the current density ceiling in our dataset, but it carries no ALMM listing, which disqualifies it from Indian subsidised deployment regardless of its technical merit.
The dataset-wide medians provide context: 21.49% efficiency, 490 Wp, 216.8 Wp/m², −0.30 %/°C, 0.40 %/yr degradation, 26.65 kg. The modules above sit in the top decile of every one of those distributions — which is precisely why their price dispersion is so visible and so contested.
System Sizing, Inverter Matching & Field Pitfalls
Price spread is irrelevant if the cheaper module cannot be legally or electrically installed. Consider a 10 kWp residential array using the Trina TSM-NEG19RC.20Q: Voc = 50.95 V, Isc = 16.6 A, 670 Wp, bifacial.
Cold-temperature string limit. Assume β_Voc = −0.25 %/°C and a record low of −10 °C:
Voc(−10 °C) = 50.95 × (1 + (−0.0025 × (−10 − 25)))
= 50.95 × (1 + 0.0875)
= 55.39 V
On a 1,100 V maximum DC input inverter:
N_max = floor(1100 / 55.39) = floor(19.86) = 19 modules
Twenty modules would produce 1,107.8 V — a 0.7% overvoltage that voids the inverter warranty and, under NEC 690.7, constitutes a code violation. The array must therefore be configured as 19 + 19 = 38 modules = 25.46 kWp, or 19 + 16 with a second MPPT.
Hot-temperature MPPT floor. At Vmp ≈ 42.5 V at STC and a −0.30 %/°C Vmp coefficient, at 65 °C:
Vmp(65 °C) = 42.5 × (1 − 0.0030 × 40) = 42.5 × 0.88 = 37.4 V
String Vmp = 19 × 37.4 = 710.6 V
Comfortably inside a 200–1,000 V MPPT window.
| Configuration | Modules | Voc at −10 °C (V) | Vmp at 65 °C (V) | Verdict |
|---|---|---|---|---|
| 1 × 16 | 16 | 886.2 | 598.4 | Safe, under-utilised |
| 1 × 19 | 19 | 1,052.4 | 710.6 | Optimal |
| 1 × 20 | 20 | 1,107.8 | 748.0 | Overvoltage — rejected |
| 2 × 19 | 38 | 1,052.4 each | 710.6 each | Two MPPT inputs required |
Mixing bins. A field pitfall that price-shopping actively encourages: combining a discounted 665 Wp pallet with an existing 670 Wp string. Isc differs (16.78 A vs 16.6 A), producing roughly 1.1% current mismatch and a 0.4–0.6% string-level yield loss — small, but it compounds with the 0.35 %/yr degradation differential over 25 years.
Bifacial gain. All eight modules in the comparison table are bifacial. Rear-side gain of 5–15% is achievable, but only with ≥1.5 m ground clearance, high-albedo surface and non-uniformity below 10%. On a flush-mounted pitched roof, bifacial gain collapses to 2–4%, and the bifacial premium embedded in the price becomes dead capital.
Lifetime Economics & Cost-per-kWh Trade-Offs
The central economic question is whether a 35% module price spread matters at system level. Using the LCOE framework with C_0 = €1,050/kWp at €0.11/Wp module pricing, a fixed balance-of-system cost of €940/kWp, O&M of €15/kWp/yr escalating at 2%, r = 6%, Y_1 = 1,550 kWh/kWp and d = 0.35 %/yr:
Discounted energy sum (25 yr) = 1,550 × (12.783 − 0.0035 × 115.95)
= 1,550 × 12.377 = 19,185 kWh/kWp
Discounted O&M = €231.7/kWp
LCOE(€0.11/Wp) = (1,050 + 231.7) / 19,185 = 0.06681 €/kWh = 6.68 ¢/kWh
LCOE(€0.148/Wp) = (1,088 + 231.7) / 19,185 = 0.06879 €/kWh = 6.88 ¢/kWh
The 35% module spread produces a 0.20 ¢/kWh LCOE delta — 3.0% relative. The sensitivity table below extends this across module price and specific yield.
| Module price (€/Wp) | CAPEX (€/kWp) | LCOE @ 1,400 kWh/kWp | LCOE @ 1,550 kWh/kWp | LCOE @ 1,700 kWh/kWp |
|---|---|---|---|---|
| 0.090 | 1,030 | 7.28 ¢ | 6.58 ¢ | 6.00 ¢ |
| 0.110 | 1,050 | 7.40 ¢ | 6.68 ¢ | 6.09 ¢ |
| 0.130 | 1,070 | 7.51 ¢ | 6.79 ¢ | 6.19 ¢ |
| 0.148 | 1,088 | 7.62 ¢ | 6.88 ¢ | 6.27 ¢ |
| 0.160 | 1,100 | 7.69 ¢ | 6.94 ¢ | 6.33 ¢ |
The table exposes the real trade-off. Moving from €0.16/Wp to €0.09/Wp — a 44% module discount — improves LCOE by only 0.41 ¢/kWh at 1,550 kWh/kWp. Moving