ADR-025: Multi-Frame Architecture (Scavenger Paths)

Status

Accepted

Amends

  • ADR-001: M10 Threaded Rod Skeleton (adds alternative frame paths)
  • ADR-021: Dual-Rod Motion System (adds V-slot motion path)

Context

The Scavenger Reality

Amalgam requires two donor printers for a complete build (ADR-005, ADR-021). The most common donors in 2026 fall into two categories:

Donor Type Examples Frame Motion
Rod-based Anet A8, Wanhao i3, Prusa clones Acrylic/sheet 8mm smooth rods + LM8UU
V-slot Ender 3, CR-10, Voxelab Aquila 2020/2040 aluminum V-slot + POM wheels

The original specification (ADR-001 + ADR-021) assumed rod-based donors: - M10 threaded rod frame (buy new, ~$30-45 AUD) - Dual 8mm smooth rods scavenged from donors

The Problem with V-Slot Donors

Ender 3 is arguably the most common donor printer in 2026. Under the original spec, an Ender 3 scavenger must: 1. Discard the aluminum extrusions (no use in M10 frame) 2. Buy 8× stainless smooth rods (~$43 AUD) 3. Buy 22× IGUS bearings (~$30 AUD) 4. Total added cost: ~$73-90 AUD

This violates the scavenger philosophy: “use what you have.”

The Insight

The MDF base serves two functions: 1. Mass damping — absorbs vibration from the frame 2. Squaring jig — precision assembly reference

With MDF as the foundation, the frame type becomes a variable rather than a constant. Whether M10 threaded rods or aluminum extrusions, the MDF provides: - Geometric reference for squaring - Mass to dampen resonance - Consistent Z-axis mounting

The M10 + MDF combination is particularly elegant because: - M10 is adequately stiff (not overkill like M12) - MDF compensates for any remaining flexibility - The system is heavy enough to dampen, light enough to move - MDF is the precision reference (no need for machinist skills)

Decision

We adopt a Multi-Frame Architecture with three supported paths, all sharing a common MDF base:

Amalgam Scaffold (M10 + Smooth Rods)

Component Specification
Frame M10 threaded rod skeleton
Base MDF squaring jig + damping mass
Motion Dual 8mm smooth rods, vertical stacking
Bearings LM8LUU (X), LM8UU (Y/Z)
Best donors Anet A8, Wanhao i3, Prusa clones

Character: The heritage path. RepRap Darwin tribute with industrial “tractor” aesthetic.

Why choose Scaffold: - RepRap heritage (inspired by the original Darwin) - Industrial “tractor” aesthetic — threaded rod is honest, visible engineering - Hardware-store availability - Forces scavenger mindset (rod-based donors) - Differentiated from commercial designs - M10 frame adds mass for vibration damping

Amalgam Mill (Extrusion + V-Slots)

Component Specification
Frame Aluminum extrusion (2020/2040)
Base MDF squaring jig + damping mass
Motion V-slot rails + POM wheels
Bearings POM V-wheels (3-4 per carriage)
Best donors Ender 3, CR-10, Voxelab Aquila

Character: The zero-waste path. Uses everything from your Ender donors.

Why choose Mill: - Ender 3 is the most common donor - Uses ALL scavenged parts (zero waste) - Proven motion system (Ender 5 style) - ~$70-90 cheaper than forcing smooth rods - Cleaner aesthetic than threaded rod

Amalgam Lathe (Extrusion + Smooth Rods)

Component Specification
Frame Aluminum extrusion (2020/2040)
Base MDF squaring jig + damping mass
Motion Dual 8mm smooth rods, vertical stacking
Bearings IGUS drylin (recommended) or LM8UU
Best donors Ender 3 + purchased rods, or mixed donors

Character: The precision upgrade. Best motion quality, still under budget.

Why choose Lathe: - Superior motion system — smooth rods + IGUS bushings outperform POM wheels - IGUS is self-lubricating, maintenance-free, quieter than linear bearings - No eccentric nut adjustments, no flat spots developing over time - Combines modern frame (extrusion) with proven motion (smooth rods) - Two Ender 3 donors @ $50-60 each + ~$75 for rods/IGUS = still under $250 AUD - Best choice if you prioritize motion quality over zero-waste

The Shared Foundation

All paths share: - MDF base — squaring jig + mass damping - Z-drop architecture — bed moves in Z only (ADR-023) - Triple-Z leveling — three independent Z-motors (ADR-005) - Cartesian X-Y — head on X, gantry on Y (not Core-XY) - Toolhead — Pitan extruder + E3D V6 (ADR-019, ADR-004) - Electronics — Dual-MCU Klipper or MKS SKIPR (ADR-012)

Who Is Amalgam For?

Got one donor? Just add Klipper to it. Amalgam isn’t for you.

Got two matching donors? Perfect. Pick your frame path: - Two smooth-rod donors (Anet A8, Wanhao, Prusa clones) → Scaffold (buy M10 rods) - Two V-slot donors (Ender 3, CR-10, Aquila) → Mill (zero-waste) or Lathe (buy rods for better motion) - Mixed donors (one extrusion + one smooth-rod) → Lathe (use extrusions + scavenged rods)

Want to buy new? Just buy a Bambu A1 Mini. This project is for scavengers.

Cost Breakdown (Two Donors, Matched)

Item Scaffold Mill Lathe
Two matching donors $100-120 $100-120 $100-120
MDF base $15-20 $15-20 $15-20
Frame material M10 rods: $30-45 $0 $0
Motion (rods/bearings) $0 $0 ~$75*
Pitan gear $2 $2 $2
Klicky microswitch $2 $2 $2
Misc (wires, bolts) $40 $40 $40
Total ~$190-230 ~$160-185 ~$235-280

*Lathe motion cost: ~$43 for 8× stainless rods + ~$30 for IGUS bushings

Optional: Add MKS SKIPR (~$130) for cleaner single-board electronics: ~$290-410

All paths achieve the <$300 AUD target (Lathe is at the upper end but still qualifies).

Choosing Between Mill and Lathe (Ender Donors)

If you have two Ender 3 donors, you have a choice:

Factor Mill Lathe
Cost $160-185 $235-280
Motion quality Good Better
Maintenance POM wheels wear, need adjustment IGUS is maintenance-free
Waste Zero (use everything) Buy rods + bearings
Philosophy Maximum scavenger Pragmatic upgrade

Recommendation: If budget is tight, Mill is excellent. If you want the best motion and can spend ~$75 more, Lathe is the superior choice.

Consequences

Benefits

  1. True scavenger philosophy — use what you have
  2. Ender 3 support — the most common donor is fully supported (Mill or Lathe)
  3. Three distinct paths — Scaffold (heritage), Mill (zero-waste), Lathe (precision)
  4. Same print quality — Klipper + Pitan + E3D V6 equalizes all paths
  5. Simple requirement — two donors, pick your path
  6. Parametric flexibility — config.py branches on donor type

Trade-offs

  1. More documentation — three paths to explain
  2. More CAD work — brackets differ per path
  3. Testing burden — must validate all three paths
  4. Potential confusion — users must choose a path

What This Enables

  • Wizard improvement: “What donor(s) do you have?” drives path selection
  • Conditional docs: Quarto shows relevant steps per path
  • STL generation: build123d outputs correct parts for chosen path
  • Fair comparison: Cost estimates accurate for each donor type

Differentiation from Other Projects

Project Frame Motion Kinematics Cost
Amalgam Scaffold M10 threaded rod + MDF Smooth rods Cartesian, Z-drop <$300 AUD
Amalgam Mill Aluminum extrusion + MDF V-slots Cartesian, Z-drop <$300 AUD
Amalgam Lathe Aluminum extrusion + MDF Smooth rods Cartesian, Z-drop <$300 AUD
Voron Legacy Aluminum extrusion Smooth rods Core-XY ~$600 USD
Mercury One Ender 5 conversion V-slots Core-XY Ender 5 + $200+
Ender 5 Aluminum extrusion V-slots Cartesian, Z-drop ~$350 AUD new

Amalgam is unique in combining: - RepRap threaded-rod heritage (Scaffold path) - Scavenger economics (all paths) - Modern Klipper intelligence (all paths) - MDF squaring jig (all paths)

Physics Comparison

Frame Stiffness (Bending)

Frame Member Bending Stiffness (EI) Relative
M10 Steel Rod ~98,000 1×
M12 Steel Rod ~204,000 2×
2020 Aluminum ~483,000 5×
2040 Aluminum ~3,850,000 39×

Aluminum extrusion is stiffer, but: - M10 is “adequate” for 70-120mm/s target speeds - MDF base compensates with mass damping - Klipper Input Shaping cancels resonance - M10 has the “tractor” aesthetic

Motion Comparison

System Deflection @ 360mm Pros Cons
Dual 8mm smooth rods + LM8UU 0.033mm Precision, proven Requires rod donors or purchase
Dual 8mm smooth rods + IGUS 0.033mm Precision, maintenance-free, quiet Must purchase (~$75)
V-slot + POM wheels ~0.05mm Zero cost from donors Wheel wear, periodic adjustment

Smooth rods with IGUS (Lathe) offer the best motion quality. V-slot with POM wheels (Mill) is acceptable and zero-cost from Ender donors. Input Shaping compensates for minor differences.

Implementation Notes

Config Parameters

# Frame path selection
FRAME_TYPE = "SCAFFOLD"  # or "MILL", "LATHE"

# Derived settings
if FRAME_TYPE == "SCAFFOLD":
    FRAME_MATERIAL = "M10_THREADED"
    MOTION_SYSTEM = "SMOOTH_ROD_DUAL"
    BEARING_TYPE = "LM8UU"
elif FRAME_TYPE == "MILL":
    FRAME_MATERIAL = "ALUMINUM_EXTRUSION"
    MOTION_SYSTEM = "V_SLOT"
    BEARING_TYPE = "POM_WHEELS"
elif FRAME_TYPE == "LATHE":
    FRAME_MATERIAL = "ALUMINUM_EXTRUSION"
    MOTION_SYSTEM = "SMOOTH_ROD_DUAL"
    BEARING_TYPE = "IGUS"

Wizard Flow

  1. “What donor printer(s) do you have?”
  2. → Detect donor type (rod-based, V-slot, mixed)
  3. → Recommend path (Scaffold, Mill, or Lathe)
  4. → Generate appropriate STLs, BOM, Klipper config

Documentation Structure

docs/guides/
├── scaffold-build.qmd    # M10 + smooth rods (heritage)
├── mill-build.qmd        # Extrusion + V-slots (zero-waste)
├── lathe-build.qmd       # Extrusion + smooth rods (precision)
└── common/
    ├── z-axis.qmd        # Shared Triple-Z assembly
    ├── toolhead.qmd      # Shared Pitan + E3D V6
    └── electronics.qmd   # Shared Klipper setup

Quality Consistency

All paths achieve the same print quality because:

  1. Same toolhead: Pitan + E3D V6 + CHT nozzle
  2. Same firmware: Klipper with Input Shaping, Pressure Advance
  3. Same Z-system: Triple-Z kinematic leveling
  4. Same calibration: ADXL345 resonance measurement

The frame and motion type affect resonance frequency, not quality. Input Shaping measures and compensates.

References

  • ADR-001: M10 Threaded Rod Skeleton
  • ADR-005: Triple-Z Kinematic Leveling
  • ADR-021: Dual-Rod Motion System
  • ADR-023: Z-Drop Architecture
  • ../philosophy.md: “Tractor with a Racecar Brain”
  • RepRap Darwin: Original threaded-rod heritage
  • Voron Legacy: Extrusion + smooth rod validation
  • Ender 5: V-slot Cartesian Z-drop reference
  • AI Conversation: “Multi-Frame Architecture Discussion” (2025-01-28)