NMOT In Solar: The Faiman Thermal Model, IEC 61853-2, And Why NMOT Replaced NOCT (2026)

I built a 6 kW PV array on my own house in 2024. The roof is asphalt shingle with about 10 cm of standoff under the racking — typical residential install. Looking at the LONGi Hi-MO 6 datasheet, I see two power numbers: 410 W at STC and 308 W at NMOT. Neither one tells me what the panel will deliver on a 30 °C August afternoon when the cells are cooking. The point of this article is to explain why NMOT exists, how it's measured, and how to translate it into a number you can actually trust on your own roof.

This is the deep-dive in our STC/NOCT/NMOT trio. If you want the basics of STC, read STC In Solar Panels. If you want the side-by-side comparison and the temperature math for modern 2026 panels, read STC vs NOCT (NMOT). This article is the part where we slow down on NMOT itself.

What NMOT Actually Measures

NMOT is not a power number. NMOT is a temperature.

The Nominal Module Operating Temperature is the back-of-module temperature the panel reaches when the following four conditions are met simultaneously:

Per IEC 61853-2:2016, a Class A facility outdoor-tests the panel for several days, records back-of-module temperature, ambient temperature, irradiance, and wind speed, then fits the data to the Faiman equation (next section) to extract the U-coefficients. NMOT is then the back-of-module temperature calculated from those coefficients at the four reference conditions above.

For a typical crystalline silicon module in 2026 — whether monocrystalline or polycrystalline — NMOT lands between 41 °C and 47 °C. LONGi's Hi-MO 6 reports 41 ± 2 °C. JinkoSolar Tiger Neo reports 45 ± 2 °C. REC Alpha Pure-R reports 44 ± 2 °C.

The corresponding electrical numbers (Pmax at NMOT, Vmp at NMOT, Imp at NMOT) are the second table on the datasheet — and they are the realistic mid-day output at 800 W/m² irradiance with the panel at its NMOT temperature.

The Faiman Thermal Model — The Equation Underneath

The whole reason NMOT replaced NOCT in 2016 is that the IEC adopted a physically grounded thermal model instead of a single-point measurement. That model is the Faiman model, published by David Faiman in Progress in Photovoltaics in 2008.

The equation is dead simple:

T_module = T_amb + G / (U₀ + U₁ · v_wind)

Where:

• T_module = back-of-module temperature (°C)
• T_amb = ambient air temperature (°C)
• G = plane-of-array irradiance (W/m²)
• v_wind = wind speed at the module (m/s)
• U₀ = constant heat-loss coefficient (W/m²/K) — represents radiative + free convective cooling
• U₁ = wind-driven heat-loss coefficient ((W/m²/K)/(m/s)) — represents forced convective cooling

You can read the equation as: the panel sits above ambient by an amount proportional to incoming irradiance, divided by how good a heat sink the surroundings provide. Low U → poor cooling → hot panel. High U → good cooling → cool panel.

For an open-rack mounted Tier 1 silicon module the typical fit values are:

Plug in NMOT reference conditions (G = 800, T_amb = 20, v = 1):

T_module = 20 + 800 / (25 + 6.8·1) = 20 + 800 / 31.8 = 20 + 25.2 = 45.2 °C

That is right in the middle of the published NMOT range for 2026 panels. The model checks out.

Why Two Coefficients Instead Of One?

The earlier Ross 1980 thermal model, still occasionally used in textbooks, is a one-coefficient version: T_cell = T_amb + k · G. It works fine when wind is constant, but it cannot tell you what happens when wind speed changes — and wind changes constantly on a real roof.

The Faiman model splits the cooling into a static term (U₀) and a wind-dependent term (U₁·v). This lets a single panel be used in different climates by re-applying its U-coefficients to local hourly wind data. PVsyst, SAM, and PVWatts all use a Faiman-style or Sandia-style two-term model internally for exactly this reason.

The Sandia King model uses an exponential rather than linear form T_module = T_amb + G·exp(a + b·v) with manufacturer-fit a and b constants — same idea, different functional form. PVsyst uses a slightly different parameterization with Uc + Uv·v heat-loss coefficients, but it is mathematically the same Faiman shape.

NOCT Versus NMOT — What Actually Changed

The reason this article exists is that the literature is full of "NOCT 45 °C" and "NMOT 45 °C" labels and people assume they mean the same thing. They don't, and the differences are subtle but real.

Two changes matter:

1. Cell temperature → back-of-module temperature. A modern crystalline silicon module has its cells laminated between EVA and glass on top, EVA and a backsheet on the bottom. The back-of-module sensor is taped to the polymer backsheet, just below the cells. The cell itself runs about 2–3 °C hotter than the back-of-module, so for the same physical panel under the same conditions, NMOT (back-of-module) reads ~2–3 °C lower than NOCT (cell). Counter-intuitive at first.

2. Open circuit → MPPT loaded. When a panel is open-circuit, every photon's worth of absorbed energy that doesn't go into reflection becomes heat. When a panel is operating at its maximum power point, roughly 20 % of the absorbed energy is being extracted as electricity and is no longer heating the cells. So loaded operation runs about 3–5 °C cooler than open-circuit. NMOT subtracts that thermal advantage from the measurement.

Net effect: NMOT and NOCT for the same panel under the same conditions are usually within 2–5 °C of each other, but they are not interchangeable. The cleaner way to think about it: NMOT was the IEC's chance to fix the realism gaps in NOCT, and they took it.

The Open-Rack Trap (Why NMOT Is Still Optimistic For Most Roofs)

Here is the part that almost no marketing copy explains, and the part that matters most if you are sizing a residential rooftop install:

NMOT is measured on an open rack, with airflow on both sides of the panel. Real residential rooftops are not open racks. They are flush-mounted or close-mounted to the roof, with very restricted back airflow.

The thermal penalty for close-roof mounting is well documented in the TamizhMani et al. NREL study, in PVsyst's documentation, and in Sandia's PVPMC modeling guide. The numbers vary by mounting style:

That extra 5–15 °C costs another 1.5–4.5 % via the temperature coefficient β. So the NMOT power on the datasheet is already optimistic by a few percent for any rooftop install, even before you account for soiling, mismatch, wiring, or inverter losses.

This is exactly why PVsyst lets you set Uc and Uv directly per system and why PVWatts has a "mounting type" dropdown (open rack vs roof mount vs heat sink). They are all applying the open-rack-vs-roof penalty on top of the NMOT-implied U-values.

Practical takeaway: if your install is a typical pitched-shingle residential rooftop with rail standoffs, add 3–5 °C to NMOT cell temperature when sanity-checking expected output. If it is a flush BIPV install, add closer to 10 °C. If it is a ground mount or commercial tilted rack with full back airflow, NMOT is roughly correct.

Worked Example — Translating NMOT To A Real Roof

Take the LONGi Hi-MO 6 LR5-54HTH 410 W. From the datasheet:

The datasheet's 308 W at NMOT is for an open-rack install at 41 °C back-of-module on a 20 °C, 1 m/s, 800 W/m² day. Now let's translate to my actual install in southern Slovenia in mid-July: ambient 30 °C, irradiance 950 W/m², light wind ~2 m/s, residential pitched roof with rails.

Step 1 — Compute the open-rack module temperature for the actual conditions.

T_module_openrack = 30 + 950 / (25 + 6.8·2)
                  = 30 + 950 / 38.6
                  = 30 + 24.6
                  = 54.6 °C (back-of-module)

Cell temperature is ~2–3 °C above that, so call it 57 °C cell.

Step 2 — Add the close-roof mounting penalty.

A standoff rack on a pitched residential roof typically adds about +4 °C cell temperature versus open-rack. So actual cell temperature ≈ 61 °C.

Step 3 — Apply the temperature coefficient to the STC rating.

P = 410 × (1 + (−0.0029) × (61 − 25))
  = 410 × (1 − 0.1044)
  = 410 × 0.8956
  = 367 W

Step 4 — Scale for irradiance (from STC's 1000 W/m² down to actual 950 W/m²):

P_actual = 367 × (950 / 1000) = 349 W

So the realistic instantaneous DC output of that panel on my roof in mid-July at noon, with ~30 °C ambient and full sun, is about 349 W — not the 410 W STC nameplate, and not the 308 W datasheet NMOT either.

The NMOT number was a closer starting point than STC, but it was still off in two directions (lower irradiance than my actual; lower temperature penalty than my actual). The honest model is the full Faiman + temperature-coefficient + irradiance-scaling chain, which is exactly what PVWatts and PVsyst do internally for every hour of the year.

How To Use NMOT Without Fooling Yourself

The cleanest mental model: STC is the nameplate. NMOT is the typical mid-day reference point. The temperature coefficient β is the physics that relates them and lets you generalize to any condition. For the fundamentals of how photovoltaic cells convert sunlight in the first place, see how do solar panels work.

Common NMOT Misreadings

1. "NMOT is the panel's real output." Closer than STC, but still a reference point. The actual instantaneous output depends on your irradiance, your wind, your mounting, and your temperature — none of which equal NMOT reference conditions in any climate I've seen.
2. "NOCT and NMOT are the same thing." They are within a few degrees but use different measurement targets (cell vs back-of-module) and different load states (open-circuit vs MPPT). Modern datasheets all use NMOT methodology even when they print the legacy "NOCT" label.
3. "My roof gets the NMOT temperature so my output should match NMOT Pmax." Roof mounting adds 3–15 °C above the open-rack reference, so even at the same ambient and irradiance, your panel runs hotter than NMOT and outputs less than NMOT Pmax.
4. "NMOT is included on every datasheet, so I don't need to use installer software." NMOT is a single point. Energy yield over a year requires a simulation that walks through 8,760 hours of irradiance + temperature + wind + losses. Use PVWatts or our solar panel calculator.
5. "Higher NMOT means worse panel." Not directly. A high NMOT (say 47 °C) on a panel with very low β (HJT, IBC at −0.24 %/°C) can still beat a panel with low NMOT (42 °C) and worse β (PERC at −0.36 %/°C). What matters is the combined product — and that is what real-world simulation captures.

Bottom Line

NMOT replaced NOCT in 2016 because the IEC wanted a physically grounded thermal model — the Faiman two-coefficient linear fit — instead of a single-point thermal balance. The new methodology is more accurate, more transferable across climates, and more compatible with the simulators (PVsyst, SAM, PVWatts) that everyone in the industry actually uses.

But NMOT is still a reference point, not a forecast. Open-rack measurement understates roof-mount temperatures by 3–15 °C depending on backsheet airflow. Use NMOT as a much better baseline than STC, then apply your mounting penalty, your irradiance, and your climate before you compare it to whatever your inverter actually logged at noon.

Keep Reading

If you found this useful, these guides go deeper into related topics:

• STC In Solar Panels — The Standard Test Conditions Foundation
• STC vs NOCT (NMOT) — Temperature Math And Modern Datasheet Comparison
• Open Circuit Voltage Of A Solar Cell — Formula And Temperature Behavior
• How To Calculate Solar Panel Efficiency
• Solar Panel Output Voltage Explained
• Standard Solar Panel Sizes And Wattages
• Average Peak Sun Hours By State
• Solar Panel Calculator — Full Energy Estimate