Nonuniform tracker angles

On single-axis tracker sites, rows or sections of rows sit at different tilts, even though they still respond to commands. Monitoring shows the tracker positions do not match the commanded angle or the calculated ideal angle, and the affected rows give less normalized current than neighboring rows. Losses repeat with sun position and are worst at low sun angles. Compare morning, solar-noon and afternoon windows, because a row can look fine at one angle and fail at another. Tracker twist is a defect in which part of a row no longer holds a common plane with the rest because of torque tube, drive, bearing, damper or coupler problems. Confirm it by measuring the angle at the drive end, center and free end under the same command. Terrain undulation losses can happen even when every row follows its command correctly. In that case the losses cluster at motor block boundaries, grade breaks and swales, because the backtracking model assumes a simpler slope than the real ground. Check the mechanical geometry first, using drone shadow lines and field angle checks, before blaming the controls. Controller offsets, bent torque tubes, foundation settlement and loose clamps can all produce similar signatures.

<!-- review notes

• sentence 2: support check: partial - u10 gives the mismatch with commanded/ideal angle and u13 gives lower normalized current in affected rows, but no passage joins them as one monitoring observation. -->