Acceleration and Input Shaping Explained

Acceleration and its Klipper-specific counterpart, input shaping, both exist to answer the same question: how fast can the toolhead change direction before the frame's own flex turns that change into a visible ripple in the print? Bambu answers it with a fast, rigid frame and vibration calibration; Klipper printers answer it with measured, per-axis compensation.

Why acceleration causes ringing

Every time the nozzle changes direction sharply — the corner of a square, a small circular boss — the frame itself flexes slightly and then oscillates back, and that oscillation gets embossed into the plastic as a faint rippled echo trailing away from the corner. Higher acceleration makes the direction change more abrupt, which makes the oscillation bigger. The fix is either less acceleration (slower, less ringing) or compensating for the oscillation in software (input shaping) so acceleration can stay high.

Bambu's approach: vibration calibration + a stiff, fast frame

The X2D's machine profile allows a maximum X/Y acceleration of 20000 mm/s² for normal printing and 20000 mm/s² while extruding — among the highest ceilings of any machine in our knowledge base — with travel-move jerk capped at 9 mm/s. That combination only works because Bambu Studio's vibration (resonance) calibration measures the specific printer's frame behavior and adjusts internally; running that calibration after any relocation or hardware change is the single highest-value fix for new ringing on a Bambu machine that used to print clean. Details and the manual speed-based fallback: the ghosting and ringing guide.

Klipper's approach: input shaping

Klipper-based printers (Voron, RatRig, and Klipper-based control boards on machines like Sovol's SV07/SV08) expose the same idea directly as a config section. SHAPER_CALIBRATE measures the real resonance frequency per axis with an accelerometer and picks a shaper type (zv, mzv, ei, and others) to cancel it out; Klipper's own worked example shows resonance frequencies clustering in the 20–100 Hz range, and a single calibration run can suggest a max_accel ceiling that differs meaningfully by shaper choice — the docs' own example lists 5900 mm/s² for the lightest-touch shaper (zv) down to 2500 mm/s² for the most aggressive one (ei) on that specific test rig — a genuine trade-off between top speed and smoothing, not a single universal number.

What to actually do about ringing

• Bambu machines: re-run vibration calibration after any move, part swap, or persistent ringing; if it's still present, lower outer wall speed 100–120 mm/s as a manual fallback. See the ghosting guide.
• Klipper machines: run SHAPER_CALIBRATE before touching slicer acceleration values at all; a shaper chosen for a false/noisy resonance reading (loose belts, a toolhead fan left running during the test, a poorly-mounted accelerometer) is worse than no shaping.
• Either way: check belt tension and frame rigidity before assuming a pure-software fix will solve mechanical looseness.