Your Pyranometer Reads Low Every Afternoon: Levelling Errors and How to Fix Them

Key Takeaways
  • A small tilt makes a pyranometer measure radiation on the wrong plane, so readings run high when the sun favours the tilt and low when it does not, with the largest error when the sun is low.
  • An afternoon shortfall that repeats on clear days and fades on cloudy days points to levelling, shading, or soiling patterns rather than a faulty detector.
  • Choose a sensor with a bubble level and adjustable levelling screws, and check its cosine response and mounting design before buying.
  • Level the sensor itself rather than trusting the mounting surface, tighten bolts in a cross pattern and recheck the bubble, keep the cable from pulling on the sensor, and avoid afternoon shading.
  • Verify levelling with a bubble check and a clear-sky curve or side-by-side reference sensor, recheck after storms or mount work, and correct tilt before adjusting calibration.
Table of contents

    Every afternoon at about half past two, the numbers stop making sense. The solar plant is still generating and the sky is clear, but the pyranometer reports less and less solar radiation as the day goes on. By late afternoon the gap is obvious: the plant's energy meter shows a solid day's output, while the irradiance log says the sun was weaker than at noon. The dome looks clean, the cable is intact, and the logger records without gaps, yet the daily total comes in low day after day, making the performance ratio look worse than it really is.

    If the plant is under warranty, that gap becomes a dispute about whether the panels are underperforming or the monitoring can be trusted. The first instinct is to blame the sensor, so a replacement is ordered and swapped in, until the same afternoon pattern returns. The problem was never a faulty detector. It was the way the sensor was sitting on its mount. A pyranometer tilted by even a small amount reads correctly only when the sun is in the direction the tilt favours, and under-reads for the rest of the day. Because the sun moves, a levelling error shows up as a time-of-day pattern, worst in the afternoon.

    Why a tilted pyranometer reads low in the afternoon

    A pyranometer measures global horizontal irradiance, the total solar radiation falling on a flat horizontal surface. Its output depends on the angle at which sunlight strikes its sensing element. When the sun is overhead, light hits a level sensor head-on; when the sun is lower, the same sunlight spreads across a larger area and strikes at a slant, so a level sensor correctly reports less. This is called the cosine response, because the ideal response falls off with the cosine of the angle between the sun's rays and the vertical above the sensor, known as the zenith angle.

    A well-designed pyranometer follows that cosine law closely, with a domed cover shaped so light arriving at a shallow angle is captured in the right proportion. But that design assumes the sensor itself is level. If it is tilted, it measures radiation on a plane tilted by the same amount, in the same direction, as the mounting error. The cosine law still applies, but now it applies to the wrong plane.

    Picture a sensor tilted slightly towards the east. In the morning the tilt works in its favour, so the reading runs a little high. At solar noon a small tilt makes little difference. In the afternoon the same tilt works against the sensor, so the reading runs low. The error grows as the sun gets lower, because the cosine response is steepest at large zenith angles. A tilt of one or two degrees, barely visible to the eye, makes a small error near noon and a much larger one when the sun is low.

    The tilt direction decides which part of the day is affected. A sensor tilted west reads high in the afternoon and low in the morning. The sensor is not broken or drifting; it is faithfully reporting radiation on a tilted plane. If the error favours the morning, the afternoon reads low and the daily total comes in short. That mismatch is often the first clue that the problem is in the measurement, not the panels.

    A tilted sensor also changes how much diffuse radiation it sees. Diffuse radiation is sunlight scattered by sky, clouds, and ground, arriving from all directions. On a clear day the direct-beam error dominates, while on a cloudy day, when diffuse radiation is a larger share, the tilt error is smaller, so the afternoon shortfall can look worse on clear days and almost disappear on overcast days. Soiling, by contrast, usually lowers readings all day and worsens gradually, while a failing detector usually gives a constant offset or noisy signal. A levelling error is distinctive because it is systematic, repeatable, and tied to the sun's position.

    What to look for when choosing a pyranometer

    First, check whether the sensor has a built-in levelling aid. Many field pyranometers include a small circular bubble level in the base, with adjustable levelling screws. That lets an installer set the sensor level on site and a technician check it later without removing the instrument. Without them, levelling depends on the flatness of the mounting surface, which is rarely good enough on a rooftop or pole bracket.

    Second, consider the cosine response: how closely a sensor follows the ideal cosine law, usually stated as an error figure at a large zenith angle. A tighter response is less sensitive to small angular errors, but no sensor is immune to tilt, because tilt changes the plane being measured. Choose one whose levelling hardware makes correct mounting achievable and verifiable.

    Third, think about mounting and maintenance. A sensor on a plate with three levelling screws can be adjusted in place; one bolted directly to a fixed bracket can only be levelled by shimming the bracket or reworking the mount. Also ask how the sensor is secured once level: screws that vibrate loose in wind, or a mount that flexes, will undo a careful job within weeks.

    Finally, know the honest limits. A single pyranometer measures irradiance at one point, on one plane. It cannot give irradiance on a tilted panel surface without a transposition calculation, and it cannot correct for shading or a partly soiled dome. If the site needs plane-of-array irradiance as well, that is a second sensor or a calculated value. Settle the output and logging chain before buying too: an analogue voltage needs a logger input with enough resolution, while a digital output needs the right protocol support.

    Installation and levelling pitfalls

    First, do not trust the mounting surface. A rooftop rail, pole bracket, or concrete plinth is rarely perfectly level, and a sensor bolted directly to it inherits that slope. A bracket that looks level by eye can easily be a degree or two out, enough to create the afternoon pattern. Always level the sensor itself, using its own bubble level and adjustment screws, or a small spirit level on the sensor body.

    Second, do not assume the sensor stays level while you tighten the bolts. Each bolt can tilt the plate slightly, and by the time the last one is torqued down the bubble has moved. Level the sensor, tighten gradually in a cross pattern, then check the bubble again. If it has moved, loosen, re-level, and tighten again.

    Third, watch the cable. A stiff cable pulled tight, or routed so it tugs on the sensor body, can tilt the sensor or slowly work the levelling screws loose. Leave a service loop so the cable hangs freely without pulling, and secure it to the pole or bracket below the sensor. Check that it does not rest on the dome or shade it.

    Fourth, check the site for shading. A pyranometer needs a clear view of the sky, and anything shading the dome for part of the day creates a time-of-day pattern that looks like a levelling error. A nearby panel row, vent pipe, or parapet wall can shade the sensor in the afternoon even if the morning is clear. Check the horizon, especially to the west, and move the sensor if shading is unavoidable there.

    Fifth, remember that a mount can move after installation. Roofs expand and contract with temperature, poles flex in wind, and ground-mounted stations can settle after rain. Treat levelling as a condition to be checked, not a one-time event.

    Verification and calibration

    The simplest verification is to check the bubble level. On a calm day, note whether the bubble sits in the centre of its circle; a bubble clearly off-centre should be corrected on the spot if access allows. Without a bubble level, place a small spirit level on the base in two directions, north-south and east-west, and record the date and direction of any tilt.

    A bubble check confirms the sensor is level at that moment, but it does not prove the readings are correct. If a second pyranometer is available, mount it next to the suspect sensor, level it carefully, and log both side by side for several clear days. The two should track each other closely all day. If the suspect sensor reads low in the afternoon relative to the reference while matching it near noon, the pattern confirms a tilt. If both read low, the problem is more likely shading affecting both equally.

    With no second sensor, use the shape of the clear-sky curve. On a cloud-free day, irradiance should be smooth and roughly symmetrical around solar noon. A curve that rises normally, peaks near noon, then falls faster than it rose hints at a tilt or afternoon shading. Plot the data: the shape is often more informative than any single value. Compare the daily total with the plant's generation over many days too. Output also depends on panel temperature and soiling, so the two never match exactly, but a performance ratio consistently lower in the afternoon, repeating on clear days and not on cloudy days, points to the measurement.

    Calibration is separate from levelling. Levelling corrects the geometry; calibration corrects sensitivity, the factor converting electrical output into irradiance. A sensor out of calibration reads high or low by a roughly constant percentage all day, not just in the afternoon. If a comparison shows a constant offset rather than a time-of-day pattern, check against a reference instrument or return the sensor for recalibration. Do not correct a levelling error by adjusting a calibration factor. Check the bubble at every maintenance visit, plus after any work on the mast or bracket and after any severe storm, and keep a log of each check.

    Troubleshooting table

    Symptom Likely cause What to check
    Low every afternoon, normal near noon Tilted away from the afternoon sun Bubble level; re-level, then compare curves
    High in the morning, low in the afternoon Tilted towards the east Bubble level in both axes; reference sensor check
    Low all day, gradually worse over weeks Soiled or obscured dome Inspect and clean the dome; log the date
    Sharp drop at the same time daily, then recovery Shading from a nearby structure Horizon to the west; shadow on the dome
    Low by a roughly constant percentage all day Calibration drift or logger scaling error Reference comparison; logger scaling; recalibration
    Noisy or jumping readings, especially in wind or rain Loose cable, poor connection, or water ingress Cable strain; connectors for moisture
    Correct at first, then slowly asymmetric over months Mount settling or screws working loose Bubble level; bracket and fixings; re-level

    Checklist

    • Check the bubble level on the sensor itself, in both directions, and record it.
    • Confirm a clear sky view, especially west, with nothing shading the dome.
    • Level with the adjustment screws, tighten bolts in a cross pattern, re-check after.
    • Leave a service loop in the cable so it does not pull on the sensor.
    • Plot the daily curve on a clear day; look for asymmetry around solar noon.
    • Compare with a reference sensor side by side for several clear days if possible.
    • Clean the dome, note the date, and rule out soiling.
    • Correct the tilt first; do not adjust calibration to fix levelling.
    • Check the bubble at every visit and after any storm or mount work.
    • Log each level check, cleaning, and adjustment.

    A pyranometer that reads low every afternoon is usually telling the truth about the wrong plane. Before ordering a replacement, check the bubble level, look at the shape of the daily curve, and rule out shading and soiling. In many cases the fix is a screwdriver, a spirit level, and a few careful minutes on the mount. For sites setting up or replacing a weather station sensor, the RS485 & 4-20mA solar radiation pyranometer sensor for weather stations is one option to consider alongside the checks above, and whatever sensor is chosen, the mounting and verification steps here still decide whether its readings can be trusted.

    Need help choosing the right sensor?

    Our engineers can help you pick the right model, check compatibility, and get your installation right the first time.

    RS485 & 4-20mA Solar Radiation Pyranometer Sensor for Weather StationRS485 & 4-20mA Solar Radiation Pyranometer Sensor for Weather StationView product →
    Email our engineers Contact us
    Back to blog

    Leave a comment

    Please note, comments need to be approved before they are published.