What the tool does
A string that is too long can exceed the inverter's maximum voltage on a cold morning; a string that is too short drops below the MPPT range on a hot afternoon. The tool works out both limits from the datasheet coefficients and the site's temperatures, checks the currents per MPPT input, and then estimates how much energy the array gives over a year.
- Strings: open-circuit voltage at the lowest ambient temperature, MPP voltage at the hottest cell temperature and at the coldest, the allowed number of modules per string, margin to the maximum voltage, operating and short-circuit current per MPPT, start-up voltage and DC/AC ratio.
- Design temperatures: record and mean yearly low and high from 24 years of daily NASA POWER data for the site, with the basis written into the report.
- Energy: hourly simulation with solar position, sky diffuse on the tilted plane, reflection losses, module temperature, temperature coefficient, system losses and the inverter efficiency curve with clipping. Monthly table, specific yield, performance ratio, capacity factor and a loss chain from sunlight to AC.
- Variants: up to four designs on the same site and equipment, side by side.
How to use it
- Enter the site, or open a weather file, which sets the site too.
- Get the design temperatures and choose the basis your code or utility asks for: record low or mean yearly low.
- Paste the electrical data of the module datasheet and of the inverter datasheet, press Read values and check every field. Values can be typed as well.
- On Strings, pick a length from the allowed range and set the number of strings and MPPTs; each check shows the value, the limit and the result.
- Load weather: one year of hourly NASA POWER data in the browser, or a typical year from PVGIS or an EPW file. On Energy and Variants read the results, and on Losses adjust the losses to your site.
- Export the design report, the workbook or the hourly CSV.
Which weather data
NASA POWER can be loaded directly for any place and year. A typical meteorological year (TMY) from PVGIS represents the long-term climate better than a single year; download it on the PVGIS website and open the file here. The same roof in Seoul gives about 1,428 kWh/m² of horizontal irradiance in NASA POWER 2023 and about 1,525 kWh/m² in the PVGIS TMY, so the energy differs by roughly the same share. The report names the data used, so two estimates can be compared fairly.
How it was checked
String voltages reproduce published hand calculations: 23 modules of Voc 22.60 V with -0.075 V/°C at -7 °C give 575 V (Mike Holt's NEC example), and a 600 V inverter with a corrected Voc of 44.6 V allows 13 modules. The energy model was compared with pvlib-python running the same equations on the same PVGIS typical years: the annual energy agrees within 0.01 % (Seoul 1,475.05 against 1,475.03 kWh/kWp). Against PVGIS's own results the performance ratio, which removes the difference between a typical year and PVGIS's multi-year irradiance, agrees within 1.5 points:
| Site | PVGIS performance ratio | This tool |
|---|---|---|
| Seoul | 80.0 % | 79.9 % |
| Tokyo | 79.2 % | 79.4 % |
| Berlin | 80.1 % | 79.7 % |
| Delhi | 75.1 % | 75.3 % |
| Denver | 79.2 % | 77.9 % |
Conditions: 1 kWp crystalline silicon, PVGIS optimum tilt and azimuth for each site, 14 % system losses, no horizon, PVGIS 5.3, results read on 2026-10-19.
Limits
Shading from the horizon or nearby objects is not modelled except as a percentage loss, and neither are bifacial gain, trackers or spectral effects. String limits use the coefficients you enter; national codes may ask for a different basis or a correction table, and the authority and the utility decide. This is a design aid, not the stamped drawings or the utility review.
Sources
- Models ported from pvlib-python (BSD-3-Clause); inverter and DC model as in the PVWatts Version 5 manual.
- Weather: NASA POWER (no restrictions on use, please cite), PVGIS (European Commission JRC).
- String voltage example: Mike Holt, NEC 690.7 sample.
Last checked 2026-10-19