You measure (or estimate) your printer’s resonance frequency, pick a shaper (ZV, ZVD, EI, MZV…), and the firmware does the rest:
fewer ripples after corners, higher usable acceleration, cleaner surfaces.
1) What problem are we solving? The edge-echo curse
When your toolhead hits a corner, the motors can obey perfectly and you still get wobble, because the printer’s structure
stores energy and releases it as oscillation. On a print, that oscillation appears as repeating ripples after sharp features:
ringing or ghosting.
Input shaping targets exactly this moment: acceleration changes (corners, direction reversals, infill zigzags).
If you’ve ever “fixed” ringing by reducing acceleration, input shaping is the higher-IQ version: keep speed, cancel the vibration.
Input Shaping: turning “ringing” into a thud
Amplitude
Time →Unshaped command (rings)
Input-shaped command (self-cancels)
corner
2) The core idea: self-canceling motion commands
A printer axis can be approximated as a mass-spring-damper system. A sudden acceleration step excites that system at its
natural frequency, so the carriage “rings” after the command.
Input shaping fixes this by applying a small timed sequence of mini-commands (impulses) that make the vibration waves overlap
destructively. In signal-processing terms, firmware filters the original motion command by convolving it with a short impulse
sequence (the “shaper kernel”).
Shaper “kernel”: tiny timed nudges that cancel vibration
Impulse amplitude
Time →
A1
A2
A3
t1
t2
t3
Original motion ⊗ (these impulses) = shaped motion
3) Shaper families: ZV, ZVD, EI, MZV (and the trade you always pay)
Different shapers exist because printers are messy: resonance frequency shifts with belt tension, toolhead mass, temperature,
and even where the gantry is positioned. Some shapers are short and snappy; others are more forgiving if your frequency estimate is off.
Short shapers
- Pros: less added delay, sharper corner response
- Cons: more sensitive to a wrong resonance frequency
- Use when: you can measure frequency accurately and your machine is stable
Robust shapers
- Pros: better suppression even with modeling error
- Cons: more delay / smoothing (may soften tiny features)
- Use when: your resonance drifts or you can’t measure precisely
That’s the recurring theme from the control literature: robustness vs command duration.
In printer-speak: “more cancellation” usually means “more smoothing”.
4) Measuring resonance: accelerometers, sweeps, and real peaks
Modern firmware workflows often use an accelerometer mounted to the toolhead and run a frequency sweep.
The resulting graph usually has one dominant peak per axis and sometimes a few smaller “side villains”.
The goal is to tune the shaper to the dominant resonance that shows up on prints.
If you change toolhead mass (new extruder, fan duct, hotend), belt tension, wheels/bearings, or frame bracing,
it’s worth re-checking resonance. The printer’s “note” changes when you change the instrument.
5) The hidden cost: shaping time and corner behavior
Input shaping doesn’t make your printer physically stiffer. It lets you push higher acceleration before ringing becomes visible,
but it introduces a small effective delay because the command is spread over time.
- Too aggressive shaping can soften micro-features or alter corner crispness.
- Too little shaping leaves residual ringing.
- The “best” setting is the one that gives you clean walls while preserving geometry where you care.
6) Practical tuning workflow
Here’s a sane loop you can use on most hobby printers (CoreXY, bedslingers, etc.):
Input Shaping: practical tuning loop
1) Print a ringing test
Confirm the artifact + baseline
2) Measure resonances
Accelerometer sweep preferred
3) Pick a shaper
Balance delay vs robustness
4) Apply shaper + accel
Raise acceleration carefully
5) Validate with test print
Look for reduced echoing
Re-measure after hardware changes (toolhead mass, belts, frame mods)
Checklist you can copy
- Before: tighten belts, check wheels/bearings, confirm frame fasteners.
- Measure: capture resonance peaks per axis if possible.
- Start: pick a common shaper choice (often MZV/EI style) and conservative acceleration.
- Validate: test print, inspect ripples and corner sharpness.
- Iterate: adjust shaper type or accel until you hit your “clean + crisp” target.
7) When input shaping won’t save you (alone)
Input shaping is not a fix for fundamentally loose mechanics. If your belts are slack, gantry wobbles, or linear motion is sticky,
shaping can reduce ringing but it can’t manufacture rigidity. Treat it as an upgrade to a decent machine, not a spell that exorcises
every demon in the chassis.
References (starter list)
- Singer & Seering: input shaping with specified insensitivity to modeling errors (classic foundation).
- Singhose and collaborators: multi-hump and optimized shaper families (NASA-era lineage).
- Recent open-access work on specified-duration / robust shaping variants (modern perspectives).
- Klipper documentation: resonance measurement and resonance compensation (practical implementation).
- Prusa ecosystem discussions: community guidance and validation experience.
Related Articles: Creality Ender 5 Max Review (2025) — Large Format 3D Printing That Actually Works · 3D Printing Finally Enters the “It Just Works” Era in 2026 · Essentium Works with Magigoo to Launch 3D Printing Accessories