Ringing leaves repeated surface echoes around sharp features. A repeatable sample gives you a way to follow those echoes through the checks ahead, beginning with the mechanics and moving on to firmware tuning. Keep that first print: you'll want to compare each change with where you started.
Read the calibration print
Use the test model and procedure documented for your firmware. For Klipper’s manual procedure, that model is ringing_tower.stl[1]; follow the guide’s orientation so its X and Y marks identify the correct axes. Record speed, acceleration and material alongside the sample, and photograph both axes under the same lighting. Those notes will help you see which conditions stayed the same when comparing a later print.
For a properly configured constant-speed ringing measurement, frequency is speed divided by ripple spacing. As an illustration, 60 mm/s divided by 1.5 mm gives 40 Hz. Check the measurement conditions before using the result: acceleration during that section of the move, or spacing measured along the wrong axis, makes it unsuitable for this calculation.
Check mechanical condition
Inspect frame fasteners, carriage play and belts using the machine's maintenance procedure, including its specified belt-tension method. The belt span, mass and motion system matter to that specification, which is why a general 50–60 Hz target is too broad to use across printers.
Reprint after correcting an observed defect. Keep the old sample; otherwise it is easy to mistake a lighting change for an improvement.
Use the actual input-shaping feature
Klipper's resonance-compensation guide[2] provides manual calibration and discusses the trade-off between vibration reduction and smoothing. Its accelerometer procedure[3] is a separate supported route to measuring the axes and choosing settings.
Marlin's ZV input shaper uses M593[4] when the required support is compiled in; its Fixed-Time Motion system uses M493[5] when that feature is enabled. Linear Advance and junction-deviation tuning are not substitutes for those input-shaping features.
Choose the shaper and frequency through the procedure for your own printer. A setting copied from another machine skips that measurement, and a relevant mechanical change gives you a reason to recalibrate before relying on the old result.
Tune motion within the machine's limits
Compare lower acceleration or the firmware's documented cornering control if the surface still needs improvement. The controls and units differ between firmware implementations; “jerk,” junction deviation and square-corner velocity are not interchangeable numbers.
As the ripples improve, give small features and dimensions the same attention. Keep any shaping effects in that comparison before changing the model's scale; enlarging the whole part is too broad an adjustment to make automatically.
Validate on a real part
Keep a final tower and print representative geometry at the intended settings. Record the accepted quality and time trade-off.
Keep the samples with the settings you settle on, so you can return to the comparison when the printer needs attention again. If another surface defect remains, treat it as a separate question rather than assuming the clean face on the tower has settled it.
