Ripples Repeating After Corners (Ghosting and Ringing)

Likely causes, most common first

1. Acceleration (and jerk) too high for the machineMost common
   Polymaker names high acceleration and jerk as the number one reason, and shows the same part printed with lower values and almost no ghosting. Prusa notes the vibration grows with print speed.
2. Input shaping off, not tuned, or out of dateVery common
   Input shaping cancels the printer's own resonance. Klipper's docs say to measure again if ringing returns, since belt tension and other changes shift the frequency; OrcaSlicer suggests once a year or after moving the printer or replacing belts, motors or frame parts.
3. Loose beltsCommon
   Prusa lists loose X or Y belts as a cause and gives belt tension steps for each of its printers. Klipper names loose or springy belts alongside a frame that is not rigid enough.
4. Unstable surface under the printerCommon
   Prusa lists an unstable surface as a cause and says to place the printer on a firm surface. A wobbly table moves with the printer and adds its own vibration.
5. Loose or misaligned parts, or a heavy toolheadOccasional
   Klipper lists misaligned mechanical parts and heavy moving mass. Polymaker notes a heavier carriage rings more, and that bed-moving machines are more prone than CoreXY or delta designs.

Fix it, one change at a time

1. Check the printer stands firm. Put it on a solid, level surface that does not wobble when you push the frame.
2. Check belt tension on X and Y with the method in your printer maker's manual. Prusa describes a deep, low tone when the belt is plucked as correct and a high pitch as too tight; some printers have a belt test in the menu. Polymaker warns that over-tight belts can make ghosting worse, so do not just crank them.
3. Check for play. With the printer off and cool, gently try to rock the toolhead and the bed. Tighten anything loose per the maker's guide.
4. Run the printer's resonance calibration if it has one. Many current printers measure resonance automatically with a built-in accelerometer; OrcaSlicer notes this is usually more accurate than a manual test. Re-run it after you change belts or move the printer.
5. On Klipper without an accelerometer, use the ringing tower. Klipper's procedure prints a tower while raising acceleration, measures the ripple spacing on the X and Y faces, and calculates a frequency for each axis (see the formula below). The measured values go in the [input_shaper] section, and Klipper suggests trying the MZV shaper first, then EI if ringing remains. Re-tune pressure advance afterward, as Klipper recommends.
6. On Marlin or RepRap firmware with input shaping, follow the printer maker's instructions or the OrcaSlicer input shaping test, which writes the values with the firmware's own command. Restore your normal acceleration afterward: the test deliberately uses very high values.
7. Set a sensible maximum acceleration. Klipper explains that higher acceleration with input shaping smooths fine detail, so it picks the lower of the acceleration where ringing becomes acceptable and the one where detail starts to blur. Without input shaping, lower acceleration and jerk in the slicer until the test is clean.
8. Accept the trade. Lower acceleration costs some print time. Polymaker's example cleaned up the part for a modest increase in time.

Confirm it is fixed

Print a small box with embossed letters or a sharp-cornered test on two adjacent faces, with a glossy, opaque filament (OrcaSlicer recommends one for its tests because it makes ringing easier to see). Shine a light at a low angle along the wall. The waves after the corners and letters should be gone or barely visible on both the X and Y faces. If ringing returns weeks later, Klipper notes belt tension and other changes can shift the frequency; measure again.

Why it happens

Every time the toolhead or bed changes direction sharply, the moving mass tries to keep going. The belts stretch a little, the frame flexes, and the whole system springs back and forth for a moment at its natural frequency. Klipper's documentation describes ringing as exactly this: mechanical vibration in the printer during quick changes of direction. The nozzle is still printing while it shakes, so the vibration is drawn into the wall as a fading series of waves.

Because the ripples come from a natural frequency, their spacing depends on how fast the nozzle was moving. That is what lets Klipper's ringing tower measure the frequency from a print:

Input shaping uses that frequency. It reshapes each motion command so the vibration it would start is canceled out by the next part of the move. It does not slow the printer down by itself, but Klipper notes it smooths sharp detail more as acceleration rises, which is why the maximum acceleration still matters.

Mechanical fixes come first. Klipper's docs say a frame that is not rigid enough, loose or springy belts, misalignment and heavy moving parts should be checked and fixed before tuning. Input shaping tuned on loose belts drifts out of tune as soon as they are tightened. For how acceleration and speed relate, see Print Speed and Acceleration.