Back to Skills

prepare-print-model

pjt222
Updated Yesterday
3 views
17
2
17
View on GitHub
Othergeneral

About

This skill prepares 3D models for FDM/SLA printing by handling export, mesh repair, wall-thickness analysis, support generation, and slicing. Use it to validate STL/3MF files for printability, troubleshoot slicing failures, and optimize part orientation for strength or surface finish. It's designed for the workflow between CAD/modeling software and the final 3D print.

Quick Install

Claude Code

Recommended
Primary
npx skills add pjt222/agent-almanac -a claude-code
Plugin CommandAlternative
/plugin add https://github.com/pjt222/agent-almanac
Git CloneAlternative
git clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/prepare-print-model

Copy and paste this command in Claude Code to install this skill

Documentation

備印模型

匯出並優化 3D 模型以為加成製造。此技能涵自 CAD/建模軟體匯出至網格修、可印性析、支持生、切片器配之全流。確模型為流形、有足壁厚、為強度與印質適向。

用時

  • 自 CAD 軟體(Fusion 360、SolidWorks、Onshape)或 3D 建模具(Blender、Maya)匯出模型以 3D 印
  • 送切片器前驗既 STL/3MF 文件可印
  • 解模型切片或印敗
  • 為強度、表面終光、最少支持材優部位向
  • 備有特定強度或公差所需之機械部
  • 保可印性而換模格式(STL、3MF、OBJ)

  • source_model:CAD 文件或 3D 模型文件之徑(STEP、F3D、STL、OBJ、3MF)
  • target_process:印程型(fdmslasls
  • material:欲印材(如 plapetgabsstandard-resin
  • functional_requirements:載向、公差所需、表面終光所需
  • printer_specs:建體、噴嘴徑(FDM)、層高之能
  • slicer_tool:目標切片器(curaprusaslicerorcaslicerchitubox)

1. 自源軟體匯出模型

匯出 3D 模型以宜印之格:

為 FDM/SLA

# If starting from CAD (Fusion 360, SolidWorks)
# Export as: STL (binary) or 3MF
# Resolution: High (triangle count sufficient for detail)
# Units: mm (verify scale)

# Example export settings:
# STL: Binary format, refinement 0.1mm
# 3MF: Include color/material data if using multi-material printer

得:模型文件以宜解析匯(機械部 0.1mm 弦容差,有機形 0.05mm)。

敗則:察模型已全定(無構造幾何)、無缺面、諸組件可見。

2. 驗網格完整

察網格為流形且可印:

# Install mesh repair tools if needed
# sudo apt install meshlab admesh

# Check STL file for errors
admesh --check model.stl

# Look for:
# - Non-manifold edges: 0 (every edge connects exactly 2 faces)
# - Holes: 0
# - Backwards/inverted normals: 0
# - Degenerate facets: 0

常見問題

  • 非流形邊:多面共邊,或邊唯一面
  • :網格表之缺
  • 反法線:模型內外反
  • 交面:自交幾何

得:報示 0 誤,或誤可修。

敗則:自動或手修網格:

# Automatic repair with admesh
admesh --write-binary-stl=model_fixed.stl \
       --exact \
       --nearby \
       --remove-unconnected \
       --fill-holes \
       --normal-directions \
       model.stl

# Or use meshlab GUI for manual inspection/repair
meshlab model.stl
# Filters → Cleaning and Repairing → Remove Duplicate Vertices
# Filters → Cleaning and Repairing → Remove Duplicate Faces
# Filters → Normals → Re-Orient all faces coherently

若自動修敗,返源軟體修建模誤(重合頂點、開邊、疊體)。

3. 察壁厚

驗所擇程之最小壁厚:

程之最小壁厚

ProcessMin WallRecommended MinStructural Parts
FDM (0.4mm nozzle)0.8mm1.2mm2.4mm+
FDM (0.6mm nozzle)1.2mm1.8mm3.6mm+
SLA (standard)0.4mm0.8mm2.0mm+
SLA (engineering)0.6mm1.2mm2.5mm+
SLS (nylon)0.7mm1.0mm2.0mm+
# Check wall thickness visually in slicer:
# - Import model
# - Enable "Thin walls" detection
# - Slice with 0 infill to see wall structure

# For precise measurement, use CAD software:
# - Measure distance between parallel surfaces
# - Check in critical load-bearing areas

得:諸壁達所擇程之最小厚。薄壁標待察。

敗則:返 CAD 增壁,或:

  • 換較小噴嘴(FDM)
  • 用「察薄壁」切片器設
  • 為原型受減強度

4. 定印向

擇向以優強度、表面終光、支持用:

向決矩

為強度

  • 向之使層線垂於主載向
  • 例:拉力下之托架 → 立印使層沿載軸疊

為表面終光

  • 最大/最顯之面平於床(最少階梯)
  • 關鍵尺對 X/Y 平面(高於 Z 之精)

為最少支持

  • 減 >45°(FDM)或 >30°(SLA)之懸
  • 可時平面置床

載向析

If part experiences:
- Tensile load along axis → print with layers perpendicular to axis
- Compressive load → layers can be parallel (less critical)
- Bending moment → layers perpendicular to neutral axis
- Shear → avoid layer interfaces parallel to shear direction

得:擇向附明確之強度、終光、支持權衡之因。

敗則:若無向滿諸需,依序分優:功能強度 → 尺精 → 表面終光 → 支持最少。

5. 生支持結構

為懸配自動或手支持:

支持角閾

  • FDM:垂之 45°(部分橋至 60° 可)
  • SLA:垂之 30°(橋能少)
  • SLS:無需支持(粉床支)

支持型

樹支(FDM,建議):

  • 與模較少接點
  • 易除
  • 宜有機形
  • 配:枝角 40-50°,枝密中

線支(FDM,傳統):

  • 對大懸更穩
  • 多接點(除難)
  • 配:格模式,密 15-20%,介面層 2-3

重支(SLA):

  • 為重部之較厚接點
  • 表面留痕之險
  • 配:接徑 0.5-0.8mm,密依部重

介面層

  • 支與模間加 2-3 介面層
  • 減表面痕
  • 略易除
# In slicer (PrusaSlicer example):
# Print Settings → Support material
# - Generate support material: Yes
# - Overhang threshold: 45° (FDM) / 30° (SLA)
# - Pattern: Rectilinear / Tree (auto)
# - Interface layers: 3
# - Interface pattern spacing: 0.2mm

得:支生於諸逾閾角之懸,預覽示無浮幾何。

敗則:若自動支不足:

  • 於關鍵區加手支強制
  • 增薄懸近之支密
  • 若支不可行,分模並段印

6. 配切片器配

設程適之參:

FDM 層高

  • 草稿:0.28-0.32mm(速、層可見)
  • 標準:0.16-0.20mm(質速平衡)
  • 細:0.08-0.12mm(順、緩)
  • 規:層高 = 噴嘴徑之 25-75%

SLA 層高

  • 標準:0.05mm(平衡)
  • 細:0.025mm(微縮、高細)
  • 速:0.1mm(原型)

諸程之關鍵參

FDM

layer_height: 0.2mm
line_width: 0.4mm (= nozzle diameter)
perimeters: 3-4 (structural), 2 (cosmetic)
top_bottom_layers: 5 (0.2mm layers = 1mm solid)
infill_percentage: 20% (cosmetic), 40-60% (functional)
infill_pattern: gyroid (FDM), grid (basic)
print_speed: 50mm/s perimeter, 80mm/s infill
temperature: material-specific (see select-print-material skill)

SLA

layer_height: 0.05mm
bottom_layers: 6-8 (strong bed adhesion)
exposure_time: material-specific (2-8s per layer)
bottom_exposure_time: 30-60s
lift_speed: 60-80mm/min
retract_speed: 150-180mm/min

得:配以程適之默設,依特定材/模需修。

敗則:若不確參,自切片器之「標準質」默配始於所擇材,再迭。

7. 逐層預覽切

察切之 G-code 問題:

# In slicer:
# - Slice model
# - Use layer preview slider to inspect each layer
# - Check for:
#   * Gaps in perimeters (indicates thin walls)
#   * Floating regions (missing supports)
#   * Excessive stringing paths (reduce travel)
#   * First layer: proper squish and adhesion
#   * Top layers: sufficient solid infill

預覽之紅旗

  • 實區之白缺:壁過薄於當前線寬
  • 大距之行:增回抽或加 z-hop
  • 首層不擠:Z-offset 下調 0.05mm
  • 頂層稀:增頂實層至 5+

得:預覽示連周長、適填、淨行、無顯缺。

敗則:調切片器設並重切。常修:

  • 薄壁缺 → 啟「察薄壁」或減線寬
  • 橋差 → 減橋速至 30mm/s,增冷
  • 拉絲 → 增回抽距 +1mm,減溫 -5°C

8. 匯出 G-code 並驗

存切之 G-code 附描述名:

# Naming convention:
# <part_name>_<material>_<layer_height>_<profile>.gcode
# Example: bracket_petg_0.2mm_standard.gcode

# Verify G-code:
grep "^;PRINT_TIME:" model.gcode  # Check estimated time
grep "^;Filament used:" model.gcode  # Check material usage
head -n 50 model.gcode | grep "^M104\|^M140"  # Verify temperatures

# Expected first layer temp:
# M140 S85  (bed temp for PETG)
# M104 S245 (hotend temp for PETG)

印前清單

  • 床平且清
  • 正材已載且乾
  • 溫合材所需
  • 首層 Z-offset 已校
  • 餘料/樹脂足
  • 印時宜監督謀

得:G-code 文存附嵌元數據、溫已驗、印時/料估合理。

敗則:若印時過(>12 時),考:

  • 增層高(0.2 → 0.28mm 省約 30% 時)
  • 減周長(4 → 3)
  • 減填(40% → 20% 為非結構)
  • 若大小不關鍵,縮模

  • 模型自源軟體以正單位(mm)與比匯出
  • 網格完整已驗:流形、無孔、法線正
  • 壁厚達所擇程之最小(FDM ≥0.8mm、SLA ≥0.4mm)
  • 印向為強度、終光、支持權衡優
  • 諸 >45°(FDM)或 >30°(SLA)之懸生支
  • 切片器配附宜層高與參
  • 逐層預覽已察,無缺或浮區
  • G-code 匯出附驗溫與合理印時
  • 印前清單已完(床平、材載等)

  1. 略網格修:非流形網可切而印敗,含缺或畸層
  2. 忽壁厚:薄壁(< 最小)將有缺,劇減強度
  3. 誤強度向:拉力部以層平於載向印致弱層離面
  4. 支不足:低估懸角致垂、拉絲、或全敗
  5. 首層忽:90% 印敗發於首層——Z-offset 與床附為要
  6. 網之溫:每印機/材組獨;常以塔試校溫
  7. 層高之過細:小於 2× 層高之細徵不能適解
  8. 不預覽切:切片器或作未料決(薄壁缺、怪填);印前常預覽
  9. 材吸濕:濕料(尤 Nylon、TPU、PETG)致層附差、拉絲、脆
  10. 支之過信:附支之大懸重部仍可垂——先試小模

  • select-print-material:依機、熱、化所需擇宜材
  • troubleshoot-print-issues:若備之模仍敗,診斷並修印敗
  • Model with Blender(未來技能):自零立優印之 3D 模
  • Calibrate 3D Printer(未來技能):E-步、流率、溫塔、回抽調

GitHub Repository

pjt222/agent-almanac
Path: i18n/wenyan/skills/prepare-print-model
0
agentsagentskillsai-assisted-developmentclaude-codeskillsteams

Related Skills

llamaguard

Other

LlamaGuard is Meta's 7-8B parameter model for moderating LLM inputs and outputs across six safety categories like violence and hate speech. It offers 94-95% accuracy and can be deployed using vLLM, Hugging Face, or Amazon SageMaker. Use this skill to easily integrate content filtering and safety guardrails into your AI applications.

View skill

cost-optimization

Other

This Claude Skill helps developers optimize cloud costs through resource rightsizing, tagging strategies, and spending analysis. It provides a framework for reducing cloud expenses and implementing cost governance across AWS, Azure, and GCP. Use it when you need to analyze infrastructure costs, right-size resources, or meet budget constraints.

View skill

quantizing-models-bitsandbytes

Other

This skill quantizes LLMs to 8-bit or 4-bit precision using bitsandbytes, achieving 50-75% memory reduction with minimal accuracy loss. It's ideal for running larger models on limited GPU memory or accelerating inference, supporting formats like INT8, NF4, and FP4. The skill integrates with HuggingFace Transformers and enables QLoRA training and 8-bit optimizers.

View skill

dispatching-parallel-agents

Other

This Claude Skill dispatches multiple agents to investigate and fix 3+ independent problems concurrently. It is designed for scenarios involving unrelated failures that can be resolved without shared state or dependencies. The core capability is parallel problem-solving, assigning one agent per independent problem domain to maximize efficiency.

View skill