Single-axis tracker vs fixed tilt: measured gain, US costs and a Szolnok model

The European Commission's Photovoltaic Geographical Information System (PVGIS) gives a single-axis tracker 16.2% more energy than the optimal fixed tilt at Szolnok, Hungary: 1,456 against 1,254 kWh per kilowatt-peak (kWp) a year. Metered field results are lower, about 12.8% to 15%. The gain comes from March to September; in winter the tracker produces less.

What do measured and benchmark data say about the energy gain?

Metered field results run from about 12.8% to 15%. Calculations and reviews give a wider range, 8.5% to 25%. Each study uses its own fixed-tilt baseline, so the rows are indicative rather than strictly comparable.

The Danish fixed array sat at 25°, about 3% below the optimal tilt in the authors' simulation, so 12.8% flatters the tracker slightly. The Spanish model gave 13% against 8.5% from measured irradiance, so models can overstate the gain.

One US benchmark is not a percentage gain. The 2025 data update from Lawrence Berkeley National Laboratory (LBNL) reports that single-axis tracking adds about 5 percentage points to median capacity factor (alternating-current basis) against fixed tilt. The addition is larger in sunnier regions. It also reports single-axis tracking on 99% of new utility-scale capacity in 2024.

Why does the gain depend on latitude, diffuse light and ground cover ratio?

A tracker improves how squarely the module faces the direct beam, and its rows need space to avoid shading each other. Three factors follow.

Latitude and the sun's path

A north-south axis follows the sun's east-west travel but not its height. The tracker gains in the mornings and evenings of long summer days, when the sun's path spans a wide range of azimuth angles. In winter the sun is low and a tilted fixed array faces it better, so the tracker loses. The Spanish study found the fixed array ahead from October to February, as does the Szolnok run below.

The same PVGIS request for five sites shows how the gain changes with latitude.

Slope and azimuth are the PVGIS optima for each site, with the azimuth within 4° of south. The horizontal-axis gain falls from 23.0% at Seville to 16.2% at Szolnok, then levels off near 16% at Hamburg and Stockholm. Tilted-axis trackers behave differently. Huld et al., the study the PVGIS manual cites for its tracking method, found that optimised tilted-axis and vertical-axis trackers keep gaining toward the north. Their yield over an optimal fixed mounting is about 30% in southern Europe, 20% to 25% in central Europe and up to 50% in northern Scandinavia.

Diffuse light

Diffuse light arrives from the whole sky, so aiming the module at the sun adds little. Lave and Kleissl found that the benefit of two-axis tracking in the United States was larger in clear-sky regions and smaller in cloudy ones. The diffuse share in the table is the PVGIS diffuse-to-global ratio weighted by monthly irradiation, 2005 to 2023. At Szolnok it is 40.4% from April to September and 50.2% from October to March. The gain falls from Seville to Szolnok as the diffuse share rises from 30.2% to 42.9% over the year, then levels off.

Ground cover ratio and backtracking

The ground cover ratio (GCR) is module area divided by land area. Close rows shade each other when the sun is low, so controllers backtrack: they tilt the modules flatter than the sun angle to avoid shadows, which gives up some gain. PVsyst's documentation notes that a higher GCR means higher shading losses, and the Danish authors add that wide spacing lets trackers backtrack less.

In US plants the GCR typically ranges from 0.40 to 0.50 for fixed tilt and 0.25 to 0.40 for tracking, according to Bolinger and Bolinger. Nextracker's datasheet gives a typical range of 25% to 45% for its tracker.

What does PVGIS give for Szolnok in Hungary?

For 1 kWp of crystalline-silicon modules near Szolnok (47.18°N, 20.19°E), PVGIS 5.3 gives 1,253.72 kWh a year at the optimal fixed tilt (38°, 2° east of south). A horizontal north-south single-axis tracker gives 1,456.38 kWh, a gain of 16.2%. These are the two requests to the free PVGIS API, run on 10 October 2026:

https://re.jrc.ec.europa.eu/api/v5_3/PVcalc?lat=47.18&lon=20.19&peakpower=1&loss=14&raddatabase=PVGIS-SARAH3&outputformat=json&optimalangles=1
https://re.jrc.ec.europa.eu/api/v5_3/PVcalc?lat=47.18&lon=20.19&peakpower=1&loss=14&raddatabase=PVGIS-SARAH3&outputformat=json&inclined_axis=1&inclinedaxisangle=0

The first request lets PVGIS choose the slope and azimuth. The second describes the tracker as an inclined-axis system with zero inclination, which the PVGIS manual gives as the horizontal-axis case. Radiation data are PVGIS-SARAH3 for 2005 to 2023, with 14% system losses. The methods page lists the data sources used on this site.

The tracker is ahead from March to September (26.0% more energy) and behind from October to February (12.1% less).

On a direct-current (DC) basis the capacity factor is 14.3% for the fixed array and 16.6% for the tracker (annual kWh divided by kWp times 8,760 hours). The performance ratio (PR) barely moves, 0.7875 against 0.7940, because it divides by the in-plane irradiation (1,592 and 1,834 kWh per m²) that the tracker also raises. The fixed-tilt run is the one used in the capacity factor vs performance ratio explainer, which shows why capacity factor is the better yardstick.

Treat 16.2% as an ideal-tracking figure. The manual does not mention row shading, backtracking or rotation limits. PVGIS also models a tilted-axis tracker (1,622 kWh per kWp, +29.4%) and two-axis tracking (1,662 kWh, +32.6%). This article uses the horizontal axis, the most widely used single-axis type according to the Spanish study.

How does a tracker change the daily production shape?

It flattens the midday peak and adds morning and evening output. The table shows mean output for June to August from the PVGIS hourly series, 2005 to 2023. PVGIS time stamps are in Coordinated Universal Time (UTC), so the first column adds 2 hours to give Central European Summer Time (CEST).

The fixed array peaks at 592 W per kWp at 12:10. The tracker's mean peak is lower, 563 W per kWp at 10:10, and its output at 18:10 is about four times the fixed array's. In December to February the picture reverses: the monthly PVGIS values put the tracker's output 15.9% lower.

This shape can raise the value of each MWh. The Danish authors report a consistent midday dip in Nord Pool spot prices. A fixed array earns most of its energy in those hours. The authors' value factor (generation-weighted price divided by average spot price) was about 4% higher for the tracker over the 11-month test, and 1% to 2% higher in most years.

Whether Hungarian prices reward the shoulder hours the same way depends on their shape, which this article does not estimate. For the midday side, see how midday solar changed Hungary's power imports.

What do trackers cost compared with fixed tilt?

Sources disagree on the direction of the cost gap. All figures are US data in US dollars (USD) per watt of direct-current capacity (Wdc), with operations and maintenance (O&M) per kWdc a year.

NREL's Q1 2021 benchmark is a bottom-up model of a 100 MWdc plant. LBNL's 2025 data update shows the reverse for projects built in 2024, but warns that its comparison does not control for site, size or labour. Its 2018 report cautions that such averages do not show that tracking is likely to be cheaper for a given project.

Levelised cost of electricity (LCOE) is annualised cost divided by annual energy. A tracker therefore lowers LCOE only if its annualised cost per kWp rises by less than its yield gain. In NREL's model installed cost is 7% higher and annual O&M is 10% higher, each below the 16.2% ideal gain at Szolnok. Both feed into annualised cost alongside financing and land, so the net effect needs a full calculation. The LCOE calculator lets readers test their own inputs.

Moving parts, wind and hail

Trackers add drives, controllers and sensors that fixed racking lacks. The International Energy Agency's Photovoltaic Power Systems programme (IEA PVPS Task 13) warns of a significant risk that a tracker will not respond appropriately to rare extreme weather, and wants responses standardised. Makers answer hail by stowing rows steeply: Nextracker's Hail Pro option goes up to 75° and ARRAY's stow in development reaches 77° (manufacturer statements). A stowed row is not tracking; the hail testing explainer covers the evidence.

Land, row spacing and terrain

For US plants built to 2019, Bolinger and Bolinger found median power densities of 0.87 MWdc per hectare for fixed tilt and 0.59 for tracking. Energy densities were 1.10 and 0.97 GWh per year per hectare. Per hectare, tracking plants therefore produced less energy despite the higher yield per kWp.

On terrain, Nextracker's datasheet lists slopes up to 15% north-south and east-west. LBNL notes that fixed tilt is still chosen on brownfields, landfills and particularly challenging ground.

Do trackers pair well with bifacial modules?

Mostly yes, because the rear-side gain is larger on a tracker. The Danish test (11 months, August 2019 to June 2020) measured a bifacial gain over grass of 10.5% on trackers and 7.2% on fixed tilt. IEA PVPS Task 13 reports gains on tracker fields over natural ground (albedo 0.2 to 0.3) that are generally below 10%. PVGIS makes no bifacial calculation, so the Szolnok figures are front-side only. The explainer on bifacial gain and albedo covers the rear-side term.

Lumen Park Mezőtúr, SolServices' 66.3 MWp DC park built by Greencells, pairs bifacial modules with a sun-tracking system (project page).

What does this mean for Hungarian owners and analysts?

• Owners and developers. Compare options on annualised cost per kWp against the yield gain, not on capex alone. A real plant with close rows will see less than the ideal 16.2% at Szolnok.
• Analysts. Compare plants by specific yield (kWh per kWp) or capacity factor, since performance ratio hides the tracker's gain.
• Traders and offtakers. A tracker shifts energy from midday to the shoulders and from winter to summer. Whether that earns more depends on the price shape in each year.
• Grid planners. In the June to August mean, the tracker's peak is about 5% lower per kWp (563 against 592 W) and output is spread over more hours.

Limits of this explanation

• PVGIS figures are for one point and a 2005 to 2023 average. They assume ideal tracking and exclude bifacial gain, clipping and curtailment.
• The hourly series gives annual totals about 1.2% to 1.3% below the monthly tool, and a gain of 16.3% against 16.2%.
• Cost and land data come from the United States, and NREL's model dates from 2021. Hungarian prices, land costs and weather exposure differ.
• The metered results come from three studies with different baselines and lengths (11 months to 3 years), and the India and Florida figures are taken from a secondary summary.
• Stow-angle figures are manufacturer claims.