Most solar calculators are glorified guessing games dressed up with satellite imagery. PVWatts, built and maintained by the National Renewable Energy Laboratory (NREL), is different: it’s the tool utilities, engineers, and state incentive programs actually use to verify production claims. And most homeowners either skip it entirely or punch in one number and trust whatever comes out.

That’s a mistake I’ve watched cost people real money. A reader emailed me last spring, Mark from Phoenix, who’d been quoted a 9.6 kW system with a 14,200 kWh/year production estimate. When he ran PVWatts himself, the output came back at 16,800 kWh for that system size at his address. His installer had either misconfigured the tilt or was padding the system size to hit a higher sale price. Either way, PVWatts caught it in about eight minutes.

Here’s what you actually need to know to use this tool correctly.

Key takeaways
  • PVWatts is NREL's free, publicly available solar production calculator used by utilities and installers, not just homeowners.
  • A south-facing roof at a tilt equal to your latitude is the baseline; every degree of deviation costs you measurable output.
  • The default DC-to-AC derate factor (0.86) is reasonable for new systems but may be too optimistic for systems over 8-10 years old.
  • PVWatts uses 30-year average weather data (TMY3 or TMY4), so a single bad weather year won't tank your estimate.
  • Always run PVWatts before accepting a contractor's production guarantee; discrepancies over 10% deserve an explanation.

What PVWatts Actually Does (and Doesn’t Do)

PVWatts Version 8, the current iteration as of July 2026, models hourly solar irradiance data drawn from NREL’s Typical Meteorological Year datasets (TMY3 and TMY4, depending on location). It runs those hourly figures through a fairly complete performance model that accounts for panel efficiency degradation with temperature, inverter conversion losses, wiring resistance, shading from horizon angles, and more. The output is an estimated annual energy production in kilowatt-hours, along with a monthly breakdown.

What it doesn’t do: shade analysis from actual trees and structures on your specific property. That requires a site-specific tool like Solmetric SunEye or Aurora Solar’s shade import feature. PVWatts lets you input a shading derate, but you have to calculate or estimate that separately. I’ll show you how below.

It also doesn’t model battery storage. If you’re sizing a system for self-consumption with a Tesla Powerwall 3 or a Panasonic EverVolt, PVWatts gives you the generation side only.

The Step-by-Step Walkthrough

Recommended Emporia Vue Gen 2 Home Energy Monitor 16 Sensors

Monitor 16 circuits in real time with solar and net metering support to optimize your solar ROI.

Check Price →

Helpful resource: Jackery Explorer 300 Portable Power Station is a top-rated option for this. (As an Amazon Associate this site earns from qualifying purchases.)

Go to pvwatts.nrel.gov. You’ll see a map and an address bar. Type in your exact street address. This matters: PVWatts pulls the nearest TMY weather station and your exact roof coordinates. I’ve seen estimates differ by 4-6% just from entering a ZIP code versus the actual street address, especially in hilly areas where nearby stations vary.

Step 1: Confirm your weather station. After you enter your address, PVWatts shows you which weather data station it’s using and your annual average solar resource in kWh/m²/day. For context, Phoenix, AZ runs around 5.5-6.0. Seattle hovers closer to 3.7-4.0. Boston lands around 4.3. If your station looks wildly off (like it’s pulling data from 60 miles away in a different climate zone), click “Change weather data source” and select a closer option if one’s available.

Step 2: System info. This is where most people get it wrong.

  • DC System Size (kW): This is your panel array’s total rated output in kilowatts under standard test conditions (STC). If you’re getting quotes for a 10 kW system, enter 10. Don’t confuse this with AC output.
  • Module Type: Choose Standard (monocrystalline or polycrystalline silicon), Premium (high-efficiency mono like Panasonic HIT or REC Alpha), or Thin Film. Most residential installs in 2026 use Standard or Premium. I default to Premium for anything using Maxeon or REC panels.
  • Array Type: Fixed (roof-mounted), 1-axis tracking, 2-axis tracking. Residential is almost always Fixed.
  • System Losses (%): Default is 14.08%. This is the composite of wiring, connections, light-induced degradation, nameplate errors, age, availability, and shading. More on this in a minute.
  • Tilt (degrees): Your roof pitch. A 5/12 pitch is 22.6 degrees. A 6/12 is 26.6. A 7/12 is 30.3. If you don’t know yours, measure it or look at your home inspection report.
  • Azimuth (degrees): Compass direction the array faces. True south is 180°. East is 90°, west is 270°.

Step 3: Adjust system losses honestly. The default 14.08% is reasonably accurate for a new, unshaded system. If you have partial shading from a chimney or dormers and you don’t have microinverters or power optimizers, bump this up. Light shading adds roughly 3-5% loss; moderate shading (nearby tree, covers 15-20% of array for a couple hours) can add 8-12%. I routinely use 18-20% for systems with string inverters in partially shaded installs.

Step 4: Read the output. PVWatts returns annual kWh production plus a month-by-month table. Cross this against your actual 12-month electricity usage (available on your utility bill or account portal). If the solar estimate covers 90%+ of your annual usage, you’re probably right-sized. Under 80%, ask your installer why.

The Numbers That Actually Move the Needle

Azimuth and tilt have more impact than most salespeople admit. NREL’s own modeling data shows that a south-facing array at latitude tilt is roughly the optimum. Deviating 45° east or west (facing southeast or southwest) costs you 5-8% annual production. Deviating 90° (full east or west) costs 15-20%. A flat roof (0° tilt) in a northern climate like Chicago or Minneapolis loses 10-15% compared to the latitude-optimized tilt.

Annual production loss by azimuth deviation (Chicago, IL, 10kW system)
Due South (baseline)0 kWh/year
SE or SW (45° off)1,050 kWh/year
East or West (90° off)2,100 kWh/year
North-facing (180° off)3,400 kWh/year
Source: NREL PVWatts modeling, 2026

The real surprise is how much temperature matters. High-efficiency panels like the REC Alpha 430W have a lower temperature coefficient (-0.26%/°C) than generic poly panels (-0.40%/°C). In Phoenix summers where rooftop temps regularly exceed 65°C, that difference adds up to 200-400 kWh annually on a mid-sized system, worth $40-80 at typical Arizona rates. PVWatts handles this automatically based on module type selection, which is why choosing “Premium” for high-efficiency panels isn’t just flattery.

Comparing Key Input Scenarios

Here’s a side-by-side look at how different installation choices affect estimated annual output for the same 10 kW system in Atlanta, GA (solar resource: approximately 5.1 kWh/m²/day):

ScenarioTiltAzimuthSystem LossesEst. Annual Output
Optimal south-facing33° (latitude)180° (true south)14%~14,200 kWh
SW-facing, slight shade22°225°18%~12,600 kWh
East-facing, flat pitch15°90°14%~11,400 kWh
South-facing, heavy shade33°180°25%~11,100 kWh
West-facing, no shade22°270°14%~12,300 kWh

That spread between best-case and worst-case is 3,100 kWh per year. At $0.14/kWh (close to the U.S. average), that’s $434 annually, and over a 25-year system life, it’s $10,000+ in foregone savings before accounting for rate increases. Contractors who don’t discuss this spread aren’t doing their job.

The Derate Factor: Where Honest Math Gets Uncomfortable

The DC-to-AC derate factor is PVWatts’ shorthand for all the real-world losses between what your panels produce and what your meter sees. NREL provides a component-by-component breakdown tool in the “Advanced” losses section if you want to get specific.

The default 14.08% assumes new equipment with no degradation. But panels degrade roughly 0.5% per year (the Solar Energy Industries Association (SEIA) cites this as the industry-accepted baseline for tier-1 manufacturers). By year 10, your system is running at about 95% of original capacity, which translates to roughly 0.5 percentage points of additional annual loss. For a 10-year payback period calculation, ignoring degradation overstates your lifetime production by about 3%.

I made this mistake myself on my own home system sizing back when I first transitioned out of electrical contracting. Ran PVWatts with default losses, calculated a tidy 7.2-year payback, bought the system. The payback is closer to 8.1 years after accounting for actual degradation and two years of slightly below-TMY weather. Not catastrophic, but I’d have done it anyway knowing the real number.

The lesson: use default losses for comparing systems or auditing a contractor’s quote, but for your own financial modeling, add 1-2% to system losses to stay conservative.

Validating a Contractor Quote

Here’s the workflow I recommend before signing anything:

Run PVWatts at your address with the exact system size, panel type, tilt, and azimuth your contractor specified. Compare their annual production estimate to PVWatts’ output. A well-configured quote should be within 5-8% of PVWatts. Under-estimates suggest the contractor is sandbagging (setting low expectations so they look good), which is actually a minor red flag for system sizing integrity. Over-estimates above 10% deserve a direct question: “What are you inputting for system losses, and why is it different from NREL’s default?”

Phoenix homeowner scenario: 9.6 kW, 20° tilt, 185° azimuth, Premium module type, 14% losses → PVWatts returns 16,847 kWh/year. Contractor quoted 14,200 kWh/year → gap of 18.6%. After confronting the contractor, they admitted they’d used a 22% loss factor citing “local dust conditions.” Adjusting to 19% (reasonable for dusty Phoenix without automated cleaning) brings the estimate to about 15,200 kWh. Still a 7% gap the contractor never explained satisfactorily.

Boston scenario: 7 kW system, 30° tilt, 195° azimuth (slightly SW-facing), Standard modules, 16% losses (modest shading from dormer) → PVWatts returns 7,640 kWh/year. Monthly breakdown shows December and January each below 300 kWh, which the installer hadn’t disclosed when projecting year-round bill offsets.

If you want to go deeper on the installation side of things, NREL’s own PVWatts documentation has the full technical methodology, and it’s actually readable. A good solar DIY guide can walk you through translating those outputs into a full system design.

Sources

  • NREL PVWatts Calculator: Official tool with TMY3/TMY4 weather data documentation and methodology notes
  • NREL System Advisor Model (SAM): More detailed performance modeling for those who want to go beyond PVWatts’ inputs
  • Solar Energy Industries Association (SEIA), 2026 Solar Market Insight: Industry benchmark data including panel degradation rates and installation trends
  • NREL, “Best Practices for Operation and Maintenance of Photovoltaic and Energy Storage Systems” (3rd ed.): Component-level loss factor data underlying PVWatts’ derate methodology
  • NREL, Typical Meteorological Year (TMY) Data documentation: Explains how the 30-year weather averages used in PVWatts are constructed and their regional limitations

Photo: Trinh Trần via Pexels

Disclosure: As an Amazon Associate, we earn a small commission from qualifying purchases at no extra cost to you. We only recommend products that genuinely support the topics covered in this article.