Engineering Reality vs. Installer Claims
The phrase "you can mix any panels as long as the voltages are close" circulates widely on installer forums and is repeated by sales teams during retrofit quotations. It is technically true that no code prohibits heterogeneous modules on a single inverter, but the claim conflates legality with electrical neutrality. A PV array is not a bag of independent generators — it is a coupled circuit whose operating point is set by the weakest element in each series branch and the lowest-voltage branch in each parallel group.
Three distinct coupling mechanisms govern mixed arrays. First, series current bottlenecking: every module in a series string carries the same current, so the string current collapses to the minimum Imp across all modules. Second, parallel Vmp clamping: when two strings of different Vmp are paralleled at a single MPPT input, the combined I-V curve peaks near the lower Vmp, forcing higher-voltage strings to operate off their maximum power point. Third, thermal divergence: modules with different γ_Pmax coefficients drift apart as cell temperature rises, so a mismatch that is tolerable at 25 °C STC can widen by several percentage points at 65 °C operating cell temperature.
The SolarQna verified corpus — 607 modules, 336 inverters, 267 batteries, 359 manufacturers across 22 markets — shows exactly how wide the electrical spread has become. Median module efficiency is 21.49%, median power is 490 Wp, and median temperature coefficient is -0.30%/°C. But the tails are extreme: the AIKO-A805-GRH78Dw reaches 25.1% efficiency at 805 Wp, while the Peimar SG330P sits at 17.0% and 330 Wp. Placing those two on the same string is not a rounding error; it is a structural redesign of the array's I-V curve.
This monograph answers the mixing question with equations, verified datasheet values, and a lifetime cost-per-kWh framework. The conclusion is nuanced: mixing is acceptable — sometimes optimal — when each sub-array has its own MPPT channel or its own module-level power electronics. It is economically destructive when dissimilar modules share a single series string or a single parallel bus.
Core Physics & Quantitative Equations
Series String Current Limiting
For a series string of n modules, Kirchhoff's current law forces a single common current I_string. The maximum power point of the string occurs where the combined I-V curve peaks, which for modules of unequal Imp is bounded by:
I_string,MPP ≤ min(Imp,1, Imp,2, …, Imp,n)
P_string,MPP = I_string,MPP × Σ Vmp,i
The mismatch loss relative to the sum of individual MPPs is:
L_series = 1 − [ I_min × Σ Vmp,i ] / [ Σ (Vmp,i × Imp,i) ]
Parallel String Vmp Clamping
When two strings share one MPPT input, the inverter tracks a single voltage. The combined current at voltage V is I_total(V) = I_A(V) + I_B(V), and the MPP of that sum is typically pulled toward the lower-Vmp string:
V_bus,MPP ≈ argmax_V [ I_A(V) + I_B(V) ]
P_parallel = V_bus,MPP × [ I_A(V_bus,MPP) + I_B(V_bus,MPP) ]
Temperature Divergence
Each module's power at cell temperature T_cell is:
P(T) = P_STC × [ 1 + γ_Pmax × (T_cell − 25) ]
Worked Example: Heliene 410 + Peimar 330 in Series
Take the Heliene Solar 108HC-410 M10 SL (Voc 37.77 V, Isc 13.54 A, γ_Pmax -0.34%/°C) and the Peimar SG330P (Voc 45.1 V, Isc 9.78 A, γ_Pmax -0.43%/°C). Approximating Vmp ≈ 0.83 × Voc and Imp ≈ 0.95 × Isc:
- Heliene: Vmp ≈ 31.35 V, Imp ≈ 12.86 A, Pmp ≈ 403 W
- Peimar: Vmp ≈ 37.43 V, Imp ≈ 9.29 A, Pmp ≈ 348 W
Series string current is capped at 9.29 A. Heliene contributes 9.29 × 31.35 = 291.3 W instead of 403 W — a 27.7% loss on that module. Peimar contributes its full 348 W. Total string output = 639.3 W versus 751 W of isolated MPP sum, a 14.9% array-level loss from a single mismatched module.
Worked Example: Same Pair in Parallel
Paralleled at one MPPT, the bus settles near 31.35 V. Heliene delivers 403 W. Peimar, forced off its 37.43 V MPP, delivers roughly 8.5 A × 31.35 V ≈ 266 W — an 82 W loss. Total = 669 W versus 751 W, a 10.9% loss. Parallel mixing is less punitive than series mixing here, but still material.
Empirical Data From Verified SolarQna Datasheets
The table below compares eight modules from the verified corpus, spanning Mono PERC, N-Type TOPCon, HJT, and Back Contact architectures. The electrical spread — from 9.78 A Isc to 17.95 A Isc — is the root cause of every mismatch scenario discussed in this monograph.
| Module | Cell Type | Wp | η (%) | Voc (V) | Isc (A) | γ_Pmax (%/°C) | Degradation (%/yr) | Warranty (product/perf) |
|---|---|---|---|---|---|---|---|---|
| Peimar SG330P | Mono PERC | 330 | 17.0 | 45.10 | 9.78 | -0.43 | n/a | 20 / 30 |
| Heliene 108HC-410 M10 SL | Mono PERC | 410 | 21.0 | 37.77 | 13.54 | -0.34 | 0.50 | 25 / 25 |
| Solar-Fabrik S4-DG460-96TMax | N-Type TOPCon | 460 | 23.02 | 36.46 | 15.80 | -0.29 | n/a | 30 / 30 |
| JinkoSolar JKM470N-48QL6-DB | Monocrystalline | 470 | 23.52 | 36.19 | 15.80 | -0.26 | 0.35 | 25 / 30 |
| DMEGC DM485G12RT-G48HBB | N-Type TOPCon | 485 | 24.30 | 36.91 | 15.98 | -0.29 | 0.35 | 25 / 30 |
| Maxeon SPR-BE54-HD485 | Back Contact | 485 | 23.80 | 40.45 | 15.18 | -0.26 | 0.25 | 40 / 40 |
| MY Solar MS740N-HJTGB | HJT | 740 | 23.82 | 51.35 | 17.95 | -0.24 | n/a | 30 / 30 |
| AIKO-A805-GRH78Dw | Back Contact | 805 | 25.10 | 59.42 | 17.13 | -0.26 | n/a | 15 / 30 |
Three observations follow immediately. First, the Isc spread of 9.78 A to 17.95 A means a single Peimar in a string of MY Solar HJT modules would throttle the entire string to roughly 54% of its native current. Second, Voc spread from 36.19 V to 59.42 V means parallel mixing of Jinko and AIKO strings would clamp the AIKO string far below its MPP voltage. Third, the degradation spread — 0.25%/yr for Maxeon versus 0.50%/yr for Heliene — means a mixed array's annual yield curve diverges progressively, not just at commissioning.
The corpus-wide distribution reinforces this: 36.2% of verified modules are Mono PERC, 34.4% are N-Type TOPCon, 9.1% are HJT, and 4.4% are Back Contact. A retrofit in 2025 will almost always pair a legacy PERC module with a modern TOPCon or HJT replacement — precisely the worst-case mismatch pairing.
System Sizing, Inverter Matching & Field Pitfalls
Worked Sizing Scenario
Consider an 8 kW residential retrofit. The existing array is 12 × JinkoSolar JKM470N-48QL6-DB (5.64 kWp) on a single MPPT. The owner wants to add 6 × DMEGC DM485G12RT-G48HBB (2.91 kWp) using a new hybrid inverter with two MPPT channels.
Option A — Single MPPT, all 18 modules in two parallel strings. String 1 = 12 Jinko (Voc string ≈ 434 V), String 2 = 6 DMEGC (Voc string ≈ 221 V). Paralleling these is catastrophic: the DMEGC string's Vmp is roughly half the Jinko string's, so the combined MPP collapses toward the DMEGC voltage and the Jinko string operates far off-MPP. Estimated loss: 18–25%.
Option B — Two MPPT channels, one per sub-array. Jinko string on MPPT-1, DMEGC string on MPPT-2. Each tracker finds its own MPP. Mismatch loss ≈ 0%. This is the correct engineering answer.
Option C — Single MPPT with DC optimizers on every module. Optimizers decouple each module's MPP, so the string voltage is synthesized. Mismatch loss ≈ 1–2%, but capex rises by roughly 8–15% and the optimizer count (18 units) introduces 18 new electronic failure points.
Edge Cases and Field Pitfalls
- Rapid shutdown (NEC 690.12). Mixing modules from different manufacturers can void the UL listing of a rapid-shutdown device if the module's maximum series fuse rating differs. Always verify the module's max series fuse against the string's OCPD.
- Reverse current. In parallel strings of unequal Vmp, the lower-voltage string can be driven into reverse current at dawn/dusk if blocking diodes are absent. IEC 62548 requires blocking diodes or string fuses sized to the highest Isc in the parallel group.
- Bifacial gain asymmetry. The Solar-Fabrik S4-DG460-96TMax and MY Solar MS740N-HJTGB are bifacial; mixing them with monofacial modules on the same MPPT means the tracker optimizes for a blended rear-irradiance response that neither module was designed for.
- Voltage window violations. A string of 12 × AIKO-A805 (Voc 59.42 V) yields 713 V at STC — dangerously close to the 1000 V or 1500 V system limit once cold-temperature Voc correction (typically +10–13% at -10 °C) is applied. Mixing high-Voc modules into an existing string can push the array past the inverter's maximum input voltage.
Lifetime Economics & Cost-per-kWh Trade-Offs
The economic case for mixing hinges on whether the mismatch loss exceeds the cost of decoupling. Below is a sensitivity table for a 10 kW array over 25 years at 1,600 kWh/kWp/yr specific yield and $0.12/kWh retail tariff.
| Scenario | Mismatch Loss | Annual Yield (kWh) | 25-yr Revenue () | Decoupling Capex () | Net 25-yr ($) |
|---|---|---|---|---|---|
| Homogeneous array, single MPPT | 0% | 16,000 | 48,000 | 0 | 48,000 |
| Mixed, single MPPT, series | 15% | 13,600 | 40,800 | 0 | 40,800 |
| Mixed, single MPPT, parallel | 11% | 14,240 | 42,720 | 0 | 42,720 |
| Mixed, dual MPPT | 0.5% | 15,920 | 47,760 | 400 | 47,360 |
| Mixed, DC optimizers | 1.5% | 15,760 | 47,280 | 1,200 | 46,080 |
| Mixed, microinverters | 1.5% | 15,760 | 47,280 | 1,800 | 45,480 |
The dual-MPPT option dominates: it recovers 96% of the mismatch loss for roughly 0.8% of lifetime revenue. DC optimizers and microinverters recover slightly less due to their own conversion losses and add meaningful capex. The economically irrational choice is single-MPPT series mixing — it forfeits 7,200 of lifetime revenue to save 400 of hardware.
Degradation divergence compounds this. A Maxeon module at 0.25%/yr and a Heliene at 0.50%/yr, both starting at 485 Wp, diverge to 456 Wp and 428 Wp respectively by year 25 — a 6% spread that further widens string mismatch over time.
Global Standards & Regional Compliance (India, USA, Europe)
India — MNRE ALMM, DCR, and PM Surya Ghar
The Approved List of Models and Manufacturers (ALMM) governs which modules may be deployed in subsidized Indian installations. Mixing an ALMM-listed module with a non-listed replacement voids eligibility for PM Surya Ghar: Muft Bijli Yojana subsidies and net-metering approvals. Domestic Content Requirement (DCR) rules further require that a minimum share of cells and modules be manufactured in India. Of the eight modules benchmarked above, only the JinkoSolar JKM470N-48QL6-DB carries ALMM approval in the SolarQna corpus — a critical constraint for Indian retrofits. Expanding an older ALMM-listed array with a non-listed module is technically feasible but forfeits subsidy and warranty continuity.
United States — NEC 2023 Article 690, UL 61730, ITC
NEC 690.8 requires conductor and OCPD sizing based on 125% of the maximum current, which for mixed strings must be computed from the highest Isc in the parallel group. NEC 690.12 rapid-shutdown requirements apply per-array, and mixing modules with different shutdown-compatible electronics can invalidate the listing. UL 61730 (the US adoption of IEC 61730) governs module safety certification; a mixed array must have every module UL-listed. For the federal Investment Tax Credit (ITC), all equipment must be new and placed in service in the same tax year — mixing salvaged modules with new ones can disqualify the entire system from the 30% credit.
Europe — IEC 61215/61730, VDE, CEI, VAT
IEC 61215 (design qualification) and IEC 61730 (safety) are the baseline. IEC 62548 specifically addresses array design with mixed modules, requiring string fuses or blocking diodes and mandating that parallel strings have matched Vmp within 5% unless independently tracked. Germany's VDE-AR-N 4105 and Italy's CEI 0-21 grid codes impose inverter-level requirements that are indifferent to module mixing but strict on rapid shutdown and anti-islanding. VAT treatment varies: most EU member states apply reduced VAT (5–10%) to residential PV, but mixed arrays using second-hand modules may be reclassified as refurbishment, triggering standard VAT (19–22%).
Methodology, Dataset Coverage & Limitations
This study draws on the SolarQna verified dataset: 607 modules, 336 inverters, 267 batteries, 359 manufacturers, 220 tracked shops, and 22 markets. Module-level parameters (Voc, Isc, γ_Pmax, degradation rate, warranty terms) are sourced from manufacturer datasheets and cross-validated against at least two independent distributor listings where available. The eight modules cited here were selected to span the full architectural spectrum — Mono PERC, N-Type TOPCon, HJT, and Back Contact — and to bracket the electrical extremes of the corpus.
Three limitations apply. First, the mismatch equations assume a single-diode model with negligible shunt resistance variation; in practice, modules with different cell architectures exhibit subtly different I-V curve shapes that can shift the combined MPP by 1–3% beyond the linear estimate. Second, the economic sensitivity table uses a fixed $0.12/kWh tariff and 1,600 kWh/kWp/yr yield; real projects in high-insolation markets (Rajasthan, Arizona, Andalusia) will see proportionally larger absolute losses from mismatch. Third, degradation rates are datasheet warranty floors, not field-measured values; independent field studies frequently show 0.1–0.3 percentage-point higher annual fade, which would amplify the divergence between mixed modules over 25 years.
The engineering verdict is unambiguous: mixing is permissible and often necessary during retrofits, but it must be paired with independent MPPT tracking or module-level power electronics. Series mixing on a shared MPPT is the only configuration that is both electrically and economically indefensible.