Input Shaper Calibration: Eliminate Ghosting and Print Faster

Ghosting, those faint ripple patterns on flat surfaces near sharp edges, is the primary quality limiter when you push print speed. Every printer has resonance frequencies where the frame and toolhead vibrate in response to sudden direction changes. At low speeds, these vibrations are too small to see. At 100+ mm/s, they produce visible artifacts. Input shaping actively counteracts these vibrations by adding calculated compensating movements to the toolhead's trajectory, and the result is dramatically cleaner prints at speeds that would otherwise produce ugly ringing.

If you're running Klipper, input shaper calibration is the single highest-impact upgrade you can make. Here's the complete process from hardware to tuned results.

What You Need

An ADXL345 accelerometer. This is a small, cheap sensor ($5-10) that measures acceleration in three axes. Klipper uses it to measure your printer's actual vibration response during a test routine, then calculates the optimal input shaper parameters from real data instead of guessing.

You have two mounting options:

• SPI connection to Raspberry Pi: Wire the ADXL345 directly to the Pi's GPIO SPI pins. This is the traditional method, reliable, well-documented, requires soldering five wires.
• USB accelerometer (KUSBA, Cartographer, etc.): A self-contained USB board that plugs directly into the Pi. No soldering, no GPIO wiring. Costs $15-25 but saves significant setup time.

Mounting the Accelerometer

The accelerometer must be rigidly mounted to the toolhead for XY calibration. Any flex between the sensor and the toolhead introduces noise that corrupts the readings. Most Voron and aftermarket toolheads have dedicated ADXL345 mounting points. For other printers, use a printed mount that screws directly to the toolhead carriage, not taped, not zip-tied, screwed.

For bed-mounted Z-axis testing (relevant for bed-slinger printers), mount the accelerometer to the center of the print bed with the same rigid mounting requirement.

Configuration

Add the accelerometer configuration to your printer.cfg. For SPI connection:

[adxl345]
cs_pin: rpi:None

[resonance_tester]
accel_chip: adxl345
probe_points:
 150, 150, 20 # Center of your bed

For USB accelerometers, the configuration varies by device, check the manufacturer's documentation for the correct config block.

Verify the connection with ACCELEROMETER_QUERY. You should see acceleration values for all three axes. If you get errors, check wiring, SPI configuration, and ensure the Pi's SPI interface is enabled in raspi-config.

Running the Calibration

Home the printer and move the toolhead to the center of the bed. Then run:

SHAPER_CALIBRATE

Klipper will vibrate the toolhead through a range of frequencies on both X and Y axes (separately). You'll hear the printer make a rising-pitch buzzing sound, that's normal. The test takes about 2-3 minutes per axis.

After the test completes, Klipper outputs recommended shaper types and frequencies for each axis. The output looks like this:

Recommended shaper_type_x = mzv, shaper_freq_x = 55.4
Recommended shaper_type_y = ei, shaper_freq_y = 41.2

Understanding Shaper Types

Klipper supports several input shaper algorithms, each with different trade-offs between vibration suppression and maximum achievable acceleration:

• ZV (Zero Vibration): The simplest shaper. Allows the highest acceleration but provides the least vibration suppression. Good for printers with very clean resonance peaks.
• MZV (Modified Zero Vibration): Better suppression than ZV with slightly lower max acceleration. The best choice for most printers with a single dominant resonance frequency.
• EI (Extra-Insensitive): Strongest vibration suppression, handles broad or multiple resonance peaks. Limits acceleration more than MZV but produces the cleanest results on printers with messy resonance profiles.
• 2HUMP_EI and 3HUMP_EI: Specialized shapers for printers with multiple strong resonance peaks. Rarely needed on well-built machines.

Applying and Saving Results

Apply the recommended settings:

SET_INPUT_SHAPER SHAPER_TYPE_X=mzv SHAPER_FREQ_X=55.4
SET_INPUT_SHAPER SHAPER_TYPE_Y=ei SHAPER_FREQ_Y=41.2
SAVE_CONFIG

Klipper also recommends maximum acceleration values for each axis. Use the lower of the two (usually Y, which has more mass) as your max_accel in printer.cfg. This ensures you stay within the input shaper's effective range on both axes.

Validation: Before and After

Print a test cube or ringing test model at high speed (150+ mm/s with 5,000+ mm/s² acceleration) before and after input shaper tuning. The difference is dramatic, ghosting ripples that were clearly visible before calibration should be completely eliminated or reduced to the point of invisibility. If some ghosting persists, check your accelerometer mounting rigidity and re-run the test.

The Speed Unlock

With input shaping properly calibrated, you can increase acceleration from typical Marlin values (1,000-3,000 mm/s²) to 5,000-10,000 mm/s² while maintaining clean print quality. That acceleration increase is where the real speed gains come from, not peak velocity, but how fast the toolhead reaches and decelerates from that velocity. A printer doing 150 mm/s with 7,000 mm/s² acceleration finishes prints dramatically faster than one doing 150 mm/s with 1,500 mm/s² acceleration, because it spends far less time accelerating and decelerating.

Input shaping turns "I'd love to print faster but the quality degrades" into "I'm printing twice as fast and the quality improved." It's the rare calibration that gives you both speed and quality simultaneously.