Reading Print Failures: A Visual Diagnostic Guide for FDM

A failed FDM print isn’t random. Every artifact — stringing, ringing, layer shifts, warping, elephant’s foot, zits, under-extrusion — has a physical cause, and in most cases the print itself tells you what went wrong. Experienced hobbyists can pick up a bad print, look at it for three seconds, and tell you the cause and the fix. That skill isn’t mystical. It’s pattern recognition.

This post is the pattern reference. For each major failure mode, we’ll cover what it looks like, what’s physically happening, what the root causes are, and the fixes — ordered by likelihood, not by alphabet. The goal is that by the end you can look at a print and diagnose it instead of guessing.

The mental model: what can go wrong

Before the specific failure modes, a quick framework. Every FDM print depends on four interacting subsystems:

1. Motion — the printer’s ability to move the toolhead accurately in XY and Z. Belts, pulleys, rails, steppers.
2. Extrusion — the drive gear pushing filament into the hotend, the nozzle extruding at the right volume. Flow rate, temperature, filament quality.
3. Adhesion — the print sticking to the build plate, and each layer bonding to the previous. Bed temp, surface prep, Z-offset.
4. Thermal management — the part cooling correctly. Part cooling fan, ambient temperature, enclosure use.

Most failures fall cleanly into one of these buckets, which is the diagnostic shortcut. Is the motion off (shifts, ringing)? Is the extrusion off (stringing, blobs, gaps)? Is adhesion off (curling, lifting, first-layer gaps)? Is thermal off (warping, weak layers, drooping)? Pick the bucket, then work through the causes in that bucket.

First-layer failures

First-layer problems are the single largest category of failure. If the first layer goes wrong, the rest of the print is doomed — the print either lifts off during printing or has a rough, weak base. These are covered in depth elsewhere on this blog; the short version:

• Gaps between first-layer beads → Z-offset too high. Lower by 0.03-0.05mm.
• Translucent, thin first layer → Z-offset too high or insufficient flow.
• Nozzle dragging, plastic extruding around the nozzle → Z-offset too low. Raise by 0.03-0.05mm.
• Print lifts off mid-way → bed temp too low, contaminated surface, or a warped plate.
• Part doesn’t stick at all → check bed cleanliness (IPA wipe), then Z-offset, then bed temperature.

For everything else, keep reading.

Stringing

What it looks like: thin plastic threads strung between parts of the model, especially across gaps in the geometry. Think “cotton candy” or “spider webs” in the print.

What’s happening: during travel moves (when the nozzle moves from one point to another without extruding), molten plastic oozes out of the nozzle. The thread it leaves gets laid down on the print. This is especially bad with PETG and nylon; PLA is relatively tolerant.

Root causes, most common first:

1. Wet filament. Moisture inside the filament vaporizes at the nozzle, pushing plastic out during travel even when retracted. If you hear popping during printing, dry the filament. This is by far the most common cause of stringing that doesn’t respond to retraction tuning.
2. Retraction too low. When the nozzle moves, the extruder should retract filament slightly to relieve pressure. Too little retraction = too much oozing. Typical retraction values: direct drive 0.5-1.5mm, Bowden 2-5mm.
3. Temperature too high. Hotter plastic is runnier and oozes more. Drop nozzle temperature by 5°C and reprint. If stringing improves, keep lowering until other quality suffers, then back off 5°C.
4. Travel speed too slow. Fast travels give less time for ooze to land on the print. Most slicers default to 100-150mm/s travel; increase to 200-250mm/s if your printer can handle it.
5. Filament quality. Some brands are worse than others. If your tuning is right and stringing persists on one spool, the filament may be formulated worse than the baseline you’re used to.

The calibration test: print a “stringing test” or “retraction tower” — a small model with pillars that travel between them. Print variants at different retraction distances and temperatures. The version with the cleanest gaps wins.

Ringing / Ghosting

What it looks like: faint echoes of a sharp edge appearing as ripples in the surface past that edge. Usually seen around letters, square cutouts, or abrupt geometry changes. Looks like the surface is vibrating behind the feature.

What’s happening: when the toolhead changes direction suddenly, the whole printer frame resonates slightly. That vibration shows up in the deposited plastic as a wavy surface — the nozzle was vibrating relative to the already-cooled layers.

Root causes:

1. Printing too fast. Lower your print speed. Ringing is directly proportional to acceleration and velocity.
2. Insufficient input shaping / resonance compensation. Modern Klipper, Bambu firmware, and recent Marlin support resonance compensation. If your printer has it and isn’t calibrated, the accelerometer calibration (a one-time 10-minute process) eliminates most ringing.
3. Loose belts or pulleys. A belt that’s slightly loose resonates more. Tension belts to the manufacturer’s spec (usually 100-120Hz when plucked).
4. Frame flex. Lightweight frames (Ender 3 clones, cheap CoreXY builds) ring more than rigid frames. Often the limit of the hardware. Heavier printers (Prusa XL, Voron 2.4, Bambu X1C) have less frame flex.
5. Decoupled bed. If the bed rocks slightly on a bed-slinger at high Y accelerations, the print surface moves relative to the nozzle and you get ghosting even with a rigid frame.

Fix priority: if you have input shaping available, calibrate it. If not, slow down and check belts.

Layer shifts

What it looks like: a sudden horizontal offset somewhere in the print. Everything below the shift is one geometry; everything above is offset by 1-5mm. Sometimes catastrophic (the top of the print peels off mid-layer), sometimes a single misregistration.

What’s happening: one of the stepper motors missed steps. It commanded “move X units” but the pulley didn’t turn the right amount. Once a shift happens, subsequent layers are printed on top of the shifted position — so the shift is permanent and propagates upward.

Root causes:

1. Mechanical obstruction. The toolhead bumped into a previous layer (warping, blob, shifted part) and the stepper couldn’t move. Inspect for ridges or blobs on the print at the shift height.
2. Belt tension too loose. Belt skips under acceleration. Tighten.
3. Belt tension too tight. Overloads the motor; motor can’t provide enough torque. Loosen slightly.
4. Stepper driver overheating. Running too hot, the driver thermally protects and skips steps. Check cooling.
5. Speeds or accelerations too high for the machine. Simply asking too much. Drop print speed by 25% and try again.
6. Mechanical damage. Bent linear rail, loose eccentric nut, damaged pulley teeth. Inspect carefully.

Fix priority: check for mechanical obstructions first, then belt tension, then slow down. If shifts happen repeatedly on the same axis at the same speed, you’ve found a mechanical limit.

Warping

What it looks like: corners of the print lift off the bed, especially on large flat parts. In severe cases, the whole bottom curls up and the print fails mid-way. Characteristic of ABS, ASA, nylon, and PC; rare for PLA.

What’s happening: plastic contracts as it cools from melt temperature to solid temperature. A large flat part has significant contraction force; the bottom layer is held by adhesion, but adjacent layers above cool and contract and pull the corners up. Once a corner lifts, the nozzle might collide with it, causing further damage.

Root causes:

1. Ambient temperature too cold. High-temp filaments need a warm environment. An enclosure helps enormously for ABS/ASA; without one, large prints are very difficult.
2. Bed temp too low. The bed should be near or above the filament’s glass transition during printing to prevent contraction. ABS at 100-110°C, ASA similar, PC at 110-120°C.
3. Insufficient first-layer adhesion. If the bottom doesn’t stick fiercely, the contraction pulls it loose. Glue stick, ABS slurry, Magigoo, or a raft all help.
4. Drafts. A draft of cool air at the print causes uneven cooling and extra warping. Enclose or shield.
5. Wrong filament choice. If you’re printing 200mm flat parts in ABS without an enclosure, no tuning fixes it. Use PETG, or get an enclosure.

Fix priority: enclosure and bed temp are the big wins. Brims and rafts are symptom treatments; they help but don’t address the underlying thermal problem.

Elephant’s foot

What it looks like: the bottom layer of the print is visibly wider than the layers above. The part has a small “flange” at the base, often a millimeter or two beyond the nominal geometry.

What’s happening: the first few layers are printed while the bed is still very hot. The plastic is soft and the squish of subsequent layers pushes the bottom plastic outward. As the plastic cools, the lower layers get progressively squished by their own weight plus the nozzle’s force.

Root causes:

1. Z-offset too low. Too much squish at the first layer amplifies elephant’s foot. Check Z calibration.
2. Bed temperature too high. If the first few layers are staying molten, they deform under pressure. Drop bed temperature by 5-10°C after the first layer (most slicers support this).
3. Insufficient cooling during the first few layers. A small part cools fast; a large base layer stays warm too long. Part cooling fan on after layer 2 or 3 helps.
4. Slicer doesn’t have elephant’s foot compensation. Bambu Studio, OrcaSlicer, and Prusa Slicer have this setting — it shrinks the first layer’s XY geometry by a configurable amount (0.1-0.3mm typical) to pre-compensate.

Fix priority: enable elephant’s foot compensation (0.15mm is a good starting value), then check Z-offset. Bed temp drop is the last resort.

Zits and blobs

What it looks like: small blobs of plastic scattered across the surface of the print, usually at seam points or layer starts. Sometimes regular-looking (every layer at the same spot), sometimes random.

What’s happening: at the seam point where the nozzle starts a new perimeter, or when it pauses and resumes, a small amount of plastic oozes out and leaves a blob. If every layer has its seam at the same spot, you get a stacked line of blobs on one side of the print.

Root causes:

1. Seam placement. Default slicer behavior is “aligned seams” (every layer at the same position) which makes one visible seam line. Change to “random” for hidden seams or “rear” for prints with an obvious back.
2. Pressure advance / linear advance not tuned. Pressure advance (Klipper, Bambu) or linear advance (Marlin) compensates for extruder pressure buildup. Without calibration, the nozzle over-extrudes at the end of each move and under-extrudes at the start.
3. Too-high nozzle temperature. Hotter filament oozes more. Try 5°C lower.
4. Retraction too aggressive. Counterintuitively, retraction during seam transitions can cause a gap followed by a blob. Tune retraction to the minimum that prevents stringing.
5. Moisture. Wet filament blobs at seams as moisture vaporizes.

Fix priority: seam placement for cosmetic improvement, pressure advance for real tuning, temperature as secondary.

Under-extrusion

What it looks like: gaps, thin spots, missing beads in the top surface. Walls have visible lines of missing material. Infill is patchy. Surfaces look “starved” — not enough plastic to fill the nominal geometry.

What’s happening: the printer is pushing less plastic than the slicer expected. Either the extruder is slipping, the nozzle is partially clogged, or the flow rate calibration is off.

Root causes:

1. Clog (partial or full). A partial clog restricts flow; full flow gets through but not enough for fast moves. Do a cold pull or swap the nozzle.
2. Extruder gear slipping. Check tension on the drive gear spring. The gear should bite into the filament without crushing it.
3. Flow rate (extrusion multiplier) too low. If your filament’s flow calibration is off, the slicer commands “extrude X” but the printer extrudes less. Calibrate flow rate — print a single-walled cube, measure wall thickness, adjust flow to match.
4. Temperature too low. If the nozzle is barely hot enough to melt filament, viscosity is too high and flow restricts. Raise nozzle temp by 5-10°C.
5. Speed too high for the volumetric throughput limit. Every nozzle has a maximum volumetric flow rate it can sustain (usually 10-25 mm³/s). Beyond that, it can’t melt fast enough. Slow down or use a higher-flow hotend.
6. Filament diameter variation. Cheap filament sometimes measures 1.70mm in one spot and 1.80mm in another. Your slicer assumes a constant 1.75mm; with thin filament, you extrude less than commanded.

Fix priority: clogs first, flow calibration second, then speed/temperature.

Over-extrusion

What it looks like: rough, bumpy top surfaces. Walls have visible ridges or squeeze-out. Overall print has a “fat” or “blobby” appearance. Dimensional accuracy is poor (prints are larger than CAD).

What’s happening: the printer is pushing more plastic than the geometry needs. Excess plastic has nowhere to go, so it bulges outward or gets dragged by the nozzle.

Root causes:

1. Flow rate too high. Most common cause. Calibrate.
2. E-steps miscalibrated. On Marlin-firmware printers, the stepper’s steps-per-mm for the extruder can drift or be set wrong. Calibrate by measuring actual extrusion length vs commanded.
3. Filament diameter assumed too low. If you’re running 1.80mm filament and the slicer thinks 1.75mm, you over-extrude by ~6%.
4. Bed too hot on small parts. First-layer over-extrusion is often elephant’s foot, not true over-extrusion.

Fix priority: flow calibration is the definitive answer. Do it once per new filament brand.

Layer separation / delamination

What it looks like: visible cracks or splits between layers, especially on tall thin walls or under load. Parts break cleanly along a layer line with minimal force.

What’s happening: layers aren’t bonding to each other. The plastic from layer N+1 isn’t melting enough of layer N to create a solid weld between them.

Root causes:

1. Nozzle temperature too low. Hotter plastic bonds better to previous layers. If you’re at the low end of the filament’s recommended range, raise 5-10°C. This is the #1 cause.
2. Part cooling fan too aggressive. Over-cooling freezes each layer before the next can bond to it. For ABS/ASA/nylon, run the fan at 0-30% (or off entirely for nylon). For PLA, 100% is fine in most cases but can over-cool tall thin features.
3. Layer height too high. Layer height should be ≤75% of nozzle diameter. 0.4mm nozzle = max 0.3mm layer; 0.32mm and below is more reliable.
4. Wet filament. Moisture disrupts bonding.
5. Draft of cold air. Uneven cooling causes uneven layer bonding.

Fix priority: temperature first. Raise until delamination stops or print quality suffers.

Pillowing

What it looks like: top surface of the print has a wavy, lumpy texture. Looks like the print has “sunk in” at the top. Often accompanied by visible infill patterns showing through the top surface.

What’s happening: the top layers aren’t thick enough to span the infill gaps. Each top layer sags slightly into the infill cells, and with few top layers, the surface shows the infill pattern.

Root causes:

1. Not enough top layers. Default is often 3; for fine print quality, use 5-6. A 0.2mm layer height with 5 top layers = 1.0mm solid top.
2. Infill density too low. With 10% infill, top layers have to span large gaps. Raise infill to 15-20% for better top surfaces.
3. Insufficient top-layer cooling. Top layers are effectively bridges over the infill. They need strong part cooling to hold shape before they sag.
4. Nozzle temp too high on top layers. Hotter plastic sags more. A slight temperature drop for top layers (some slicers support this) reduces pillowing.

Fix priority: top layers count, then infill density, then cooling.

Stringing on the inside (internal ooze)

What it looks like: fine strings of plastic visible when you look at the interior of a print, especially at infill crossings or in hollow sections. Sometimes the internal strings affect external quality (zits pushed out by internal ooze).

What’s happening: during travel moves across hollow sections, the nozzle oozes just as it does externally — but inside the print, where it’s less visible.

Root causes: same as external stringing (wet filament, retraction, temperature). Internal stringing matters most if it affects dimensional accuracy of internal features or manifests as external zits.

The big-picture diagnostic

When a print fails, here’s the checklist, in order:

1. Is the filament dry? 50% of quality problems are moisture.
2. Is the first layer correct? If not, nothing else matters.
3. Is the nozzle in good condition? A worn or clogged nozzle limits everything downstream.
4. Is the flow rate calibrated for this filament? Per-filament calibration is the tightest feedback loop.
5. Is the temperature at the right point for this filament? Check the spool or the filament maker’s recommendation.
6. Is the print speed reasonable for the geometry? Tall thin parts need slower speeds than wide flat parts.
7. Is there a specific artifact (stringing, ringing, layer shift)? Match to the sections above.

Skipping ahead to step 6 when step 1 is the problem wastes hours. Most experienced hobbyists triage in this order reflexively, and you’ll train yourself into the same reflex within a few months.

The calibration regimen

A reasonable per-filament calibration for a new spool:

1. Dry the filament for 4-8 hours at 60-70°C (for non-TPU/nylon).
2. Print a 20mm calibration cube to verify nothing’s fundamentally wrong.
3. Print a flow calibration test (Bambu Studio / Orca have these built in). Set flow rate.
4. Print a temperature tower if quality feels off. Set nozzle temp.
5. Print a retraction test if stringing is obvious. Set retraction.
6. Save the filament profile in your slicer with these settings.

Total time: 1-2 hours for a new brand. After the first spool of a given brand, you usually reuse the profile.

The short version

Print failures are signals, not surprises. Each pattern — stringing, ringing, layer shifts, warping, zits — points to a specific physical cause and a limited set of fixes. The fastest way to get good at FDM is to stop treating failures as “the printer is broken” and start treating them as “the printer is communicating something.” Dry the filament, watch the first layer, calibrate flow, and the vast majority of problems disappear before they start.

Everything else is tuning. And tuning — the process of making a printer better and better at a specific job — is half the fun of the hobby, once you’ve stopped fearing the failed print.