Ringing and Ghosting on Carbon Fiber Nylon Prints: Tuning It Out for Good

You bumped the speed on a carbon fiber nylon bracket and now every corner leaves a shadow. Fine ripples throw off each sharp edge, the spacing widens as you push faster, and the part looks like it came off a machine that needs a rebuild. That pattern is ringing — ghosting, echoing, rippling, same defect — and it's mechanical, not a flow problem.

Ringing Is Vibration After a Direction Change

Ringing is the toolhead, or the bed, still oscillating after the printer stops and reverses at a corner. The frame, belts and moving mass set the frequency in Hz; the input shaper decides whether the printer keeps commanding that frequency. Spacing between ripples is travel distance per oscillation — wider gaps just mean you're printing faster, same vibration.

CF nylon shows it clearly. The matte surface rakes light across ridges you can feel with a fingernail — an honest material to tune against.

Measure It, Don't Guess It

On Klipper: bolt on an ADXL345 or LIS2DW accelerometer, run TEST_RESONANCES AXIS=X and again for Y, then feed the CSVs into Klipper's shaper calibration script. MZV suits most machines; EI is for a broad, messy resonance peak. Keep max_smoothing near 0.1–0.2 — past that the shaper rounds your corners and hides the problem instead of fixing it.

If the lowest resonance comes back under 20–25 Hz, stop tuning. That machine is flexing more than input shaping can paper over. Tighten belts, stiffen the toolhead, drop acceleration, then measure again.

No accelerometer? Run an acceleration tower. Disable pressure advance, set your shaper, then run a TUNING_TOWER test with start accel 1500, STEP_DELTA 500, STEP_HEIGHT 5. Print it in the material you actually produce parts in, find the height where ripples stop, and set max_accel about 20% below that. Marlin machines rarely give you the same tools, so the wins come from belt tension, frame bolts and lighter moving mass.

The Settings That Do the Heavy Lifting

Belt tension first — pluck both belts, they should read within a couple of Hz of each other, and one that sounds flat is your ringing right there. Then screws: a loose toolhead bolt or wobbly fan duct rings worse than a slightly slack belt. Then mass. Every gram on the toolhead lowers the resonance frequency, so a lighter duct and spool holder are tuning changes. Rails that drag or a cable chain catching at the end of travel show up as ripple too.

Retune Whenever the Machine Changes

New fan duct, new hotend, heavier spool mount, different build plate — all of it moves the tuning. Re-run the test after hardware swaps instead of copying a profile off a forum. A profile tuned on somebody else's printer is a starting point, never an answer.

What We Use

We tune on our QIDI Plus5 — Klipper-based firmware makes input shaping a config change instead of a firmware swap, and the 370°C hotend with the 65°C chamber lets us test in the high-temp carbon nylon we actually sell parts in. The towers run in TINMORRY PA6-CF — stiff, matte and unforgiving about surface defects, which is what you want when you're hunting vibration.

Affiliate disclosure: some links are affiliate links — you pay the same price, we earn a small commission that keeps our testing free.