KEMI Docs
FDM 3D Printing in Practice
A comparison matrix of seven filament materials (temperatures, shrinkage, moisture), recommended line widths and layer heights per nozzle diameter, and stringing causes ranked in fixing order. Three representative sheets excerpted from the KEMI 3D Printing Reference Bundle (printable sheets); table figures are transcribed as published.
Excerpt 1
Filament Comparison
Every figure is a starting point from common datasheet ranges — the spool label always wins.
| Material | Nozzle(°C) | Bed(°C) | Encl. | Part fan | Shrink(%) | Moisture | HDT(°C) | Strength · toughness | Typical use | Difficulty |
|---|---|---|---|---|---|---|---|---|---|---|
| PLA | 190~230 | 50~65 | Not needed | 100% | 0.2~0.5 | Low | 55~60 | Stiff but brittle — it snaps | Prototypes · decor · class | Easy |
| PETG | 230~260 | 70~85 | Not needed | 30~50% | 0.4~0.6 | Medium | 70~80 | Tough, resists cracking; strings | Functional · outdoor bits | Medium |
| ABS | 240~270 | 90~110 | Advised | 0~20% | 0.6~0.8 | Medium | 90~100 | Tough; acetone smoothing | Cases · car interior | Hard |
| ASA | 240~270 | 90~110 | Advised | 0~20% | 0.4~0.7 | Medium | 95~105 | ABS-grade plus real UV resistance | Permanent outdoor parts | Hard |
| TPU95A | 220~250 | 35~60 | Not needed | 20~50% | 0.5~1.5 | High | — | Flexible, wear-resistant, damping | Gaskets · bumpers · grips | Hard |
| PA(nylon) | 260~300 | 80~110 | Advised | 0~30% | 1.0~2.0 | Very high | 70~150 | Toughest, best wear resistance | Gears · hinges · wear parts | Very hard |
| PC | 260~300 | 100~120 | Required | 0~10% | 0.6~0.8 | Very high | 110~140 | Top stiffness and heat resistance | Hot, highly loaded parts | Very hard |
Enclosure Not needed = open frame is fine · Advised = warping and cracking without one · Required = effectively fails without one. TPU is chosen by Shore A hardness, not HDT.
Choosing guide
Start from what the part must survive, then work back to the material.
| Requirement | First choice | Alternative | Why |
|---|---|---|---|
| Indoor decor | PLA | PETG | Easiest to print and dimensionally stable; no load, no heat. |
| Outdoor use | ASA | PETG | UV is the problem, not strength. ABS embrittles, PLA sags. |
| Car interior | ASA · ABS | PC | A parked car passes 60 °C; PLA and PETG lose their shape. |
| Food contact | Process first | PETG | Bacteria trapped between layers matter more than the grade. |
| Flexible parts | TPU 95A | TPU 85A | Pick by Shore A hardness — lower is softer. Direct drive. |
| Load-bearing | PA · PC | PETG | Add perimeters before changing material. |
| Sliding · gears | PA | PETG | Nylon is self-lubricating — but only if you keep it dry. |
Watch out
- · ABS, ASA and PC need an enclosure AND ventilation. An enclosure stops warping but also traps VOCs and ultrafine particles — sealing it without exhaust or filtration is worse for you.
- · PA, PC and TPU absorb water fast. A spool left out gives rough surfaces, popping and weak parts — drying is a precondition, not an option.
- · Part cooling works opposite to shrinkage. PLA likes maximum cooling; ABS and PC crack if you cool them.
Excerpt 2
Nozzle Size & Layer Height
Layer height = 25 – 75 % of nozzle diameter
| Nozzle Ø | Line width | Layer height (25–75% of Ø) | Relative print time (0.4 mm = 1) | Smallest detail | Typical use |
|---|---|---|---|---|---|
| 0.2 mm | 0.20 – 0.25 mm | 0.05 – 0.15 mm | approx. 3 – 4× | approx. 0.2 mm | Miniatures, fine detail |
| 0.4 mm | 0.40 – 0.50 mm | 0.10 – 0.30 mm | 1× (baseline) | approx. 0.4 mm | General purpose — stock profiles |
| 0.6 mm | 0.60 – 0.75 mm | 0.15 – 0.45 mm | approx. 0.5× | approx. 0.6 mm | Functional and larger parts |
| 0.8 mm | 0.80 – 1.00 mm | 0.20 – 0.60 mm | approx. 0.3× | approx. 0.8 mm | Large models, draft speed |
Relative print time reflects only the drop in layer and line count — the real ceiling is the hotend's maximum volumetric flow (mm³/s).
Why the 25 – 75 % rule holds
Upper limit 75% — thicker layers stop the nozzle's flat face from pressing the extrusion onto the layer below, so interlayer bonding falls away sharply. Lower limit 25% — thinner layers make the nozzle drag on solid plastic, leaving too little volume for stable flow.
Worked example — layer count and time
A 60 mm part at 0.20 mm → 60 ÷ 0.20 = 300 layers Same part at 0.28 mm → 60 ÷ 0.28 = 214.3 → 215 layers 215 ÷ 300 = 0.717 → about 28% fewer layers Acceleration and travel time do not scale with layer count, so the real saving lands below 28%.
Abrasive filament + brass nozzle = a consumable — carbon/glass-filled, glow-in-the-dark and wood fill widen a brass orifice visibly within tens of print hours. Line width and flow drift from the slicer's values, a decay easily blamed on a setting. Use hardened steel or better for abrasives.
Excerpt 3
Stringing & Oozing
Retraction relieves pressure — the pressure pushing melt out of the nozzle does not vanish the moment extrusion stops. Distance differs 5–10× by drive type, speed barely — set distance first.
Retraction starting values
| Extruder type | Dist. (mm) | Speed (mm/s) | Z-hop (mm) | Notes |
|---|---|---|---|---|
| Direct (short path) | 0.4–1.0 | 30–40 | 0–0.2 | 0.8 a common start |
| Direct (general) | 1.0–2.0 | 30–45 | 0–0.2 | Gearbox sits well above the hotend |
| Bowden (short tube) | 3–5 | 30–45 | 0.2–0.4 | Tube ~300 mm — longer tube, longer pull |
| Bowden (long tube) | 5–7 | 30–50 | 0.2–0.4 | Tube 500 mm+ — past 7 mm the cause is elsewhere |
Z-hop is a nozzle-collision guard, not a stringing fix — it adds time with the nozzle parked, so oozing can get worse.
Fix causes in this order — top to bottom
| Rank | Cause | Action | How to confirm |
|---|---|---|---|
| 1 | Wet filament | Dry the whole spool, retest | Popping sounds, steam, bubbles in the strand |
| 2 | Nozzle too hot | Drop 5 ℃ at a time, temp tower | Strings get visibly thinner each step |
| 3 | Too little retraction | Raise the starting value by 0.2–0.5 mm | Compare bands on a retraction tower |
| 4 | Travel speed too low | Raise travel to 150–300 mm/s | Strings only in the slow travel sections |
| 5 | No coasting / wiping | Enable wipe, add coasting sparingly | Strings start where extrusion ended |
| 6 | Residue on the nozzle | Brush the hot nozzle with brass | Thick, irregular strings in an old color |
Watch out
- · No retraction value fixes wet filament — moisture boils in the nozzle and pushes material out. Dry first.
- · Over-retraction causes clogs and heat creep. No gain from more distance? Roll it back.
- · Large retraction on TPU = feed failure — flexible filament buckles instead of moving. Use travel speed instead.
- · Oozing ≠ stringing — oozing is leakage while parked, a string is that leak stretched.
This document is an excerpt
The full PDF edition of the FDM 3D Printing Reference covers filament storage and drying, first layer and bed adhesion, infill and supports, a print-defect diagnosis flowchart, warping and extrusion problems, calibration procedures and how to read calibration towers, plus design rules and strength design for 3D printing — all as printable sheets and log sheets.
Have a model to print?
Upload an STL and see FDM and SLA estimates from its volume. Single enclosures and jigs welcome.
3D printing quote →FAQ
Should I use PLA or PETG?
For indoor decor and prototypes with no load or heat, PLA is enough — easiest to print and dimensionally stable. For functional parts and outdoor bits that must resist cracking, pick PETG, keeping in mind it is the most stringing-prone material.
What layer height should I use?
The recommended layer height is 25–75% of the nozzle diameter. Above 75% the nozzle's flat face can no longer press the extrusion onto the layer below and interlayer bonding falls away sharply; below 25% the nozzle drags on solid plastic and flow becomes unstable. For a 0.4 mm nozzle: 0.10–0.30 mm.
What do I fix first when I get stringing?
The order matters: ① wet filament (dry, retest) → ② nozzle temperature (drop 5 °C at a time) → ③ retraction distance → ④ travel speed → ⑤ coasting/wiping → ⑥ nozzle residue. Skip 1 and 2 and you keep raising a retraction value that was never the problem — until you cause a clog.
This document is an excerpt of KEMI's own reference material. Figures vary widely by manufacturer, grade (fillers, additives) and slicer/firmware version — always defer to the spool label and datasheet. Bambu Lab and Bambu Studio are trademarks of Bambu Lab; Prusa and PrusaSlicer of Prusa Research. All trademarks mentioned are the property of their respective owners.