Amalgam: X-Gantry Design Decision

⚠️ Partially Outdated: This document explores Cartesian vs. Core-XY trade-offs in the context of old tier system (Tier 0/1/2/3). Current canonical decisions are in: - ADR-025: Multi-Frame Architecture (Scaffold, Mill, Lathe all use Cartesian, not Core-XY) - ADR-000: Engineering Philosophy (“Tractor” principles favor Cartesian simplicity)

The core insight (Cartesian is simpler for scavenger builds) remains valid, but framework is outdated.

Engineering Decision Record

Project: Amalgam 3D Printer
Component: X-Axis Gantry System
Decision: Parallel-Rail Sled (“The Plough”)
Status: Accepted

Executive Summary

After evaluating multiple X-gantry architectures including Cross-Rod (Eustathios/HercuLien), CoreXY variants, and the classic Cartesian sled, we selected the Parallel-Rail Sled design (“The Plough”) for Amalgam.

Key Drivers: - Simplicity over speed - Scavenger-friendly construction - Maintainability by non-experts - Authentic RepRap Darwin lineage - Leverages M10 rod stiffness advantage

Design Context

Amalgam Philosophy

Amalgam is designed as: - A “Tractor with a Racecar Brain” (heavy frame, smart software) - A scavenger/junkstrap build using donor printers - Quality-focused rather than speed-focused - Maintainable by people with basic mechanical skills - A spiritual successor to the original RepRap Darwin (2007)

Hardware Constraints

Component Specification
Frame M10 threaded rod on 36mm laminated MDF
X-Rails M10 smooth rods (scavenged from photocopiers)
Span ~380mm
Extruder Pitan (Titan clone) direct drive
Hotend E3D V6
Controller Klipper on dual MCU
Z-System Triple-Z steppers with kinematic joints
Y-System Dual Y-steppers with auto-squaring

Alternatives Considered

Option A: Cross-Rod Gantry (Eustathios/HercuLien Style)

Description:
Two rods for X and two for Y intersect at the toolhead. The toolhead “floats” at the crossing point, often called a “spider” carriage. This is sometimes called “Ultimaker-style.”

    ════════════════════════  Y-Rod 1
           ╲         ╱
            ╲       ╱
             [HEAD]
            ╱       ╲
           ╱         ╲
    ════════════════════════  Y-Rod 2
    
    ║                      ║
    ║      X-Rods          ║
    ║   (perpendicular)    ║

Pros: - Extremely light toolhead (only hotend moves in X-Y) - Excellent for high-speed printing (500mm/s+) - The “gold standard” for high-end rod-based DIY printers - M10 rods would provide essentially zero sag at 380mm

Cons: - Complex belt routing (crossed belts or dual belt systems) - Requires precise geometry—belt path errors cause skew - “Spider” carriage needs tight tolerances - Harder to scavenge parts (specific pulley positions required) - Klipper config more complex ([hybrid_corexy] or custom kinematics) - More difficult to troubleshoot and repair - Not in the Darwin/Mendel lineage

Verdict: Rejected—too complex for scavenger build philosophy.

Option B: CoreXY / H-Bot

Description:
Two motors work together via a continuous belt loop. Both motors spinning the same direction = Y movement. Opposite directions = X movement.

    [Motor A]═══════════════[Motor B]
              ╲           ╱
               ╲ Belts  ╱
                ╲     ╱
                 [HEAD]
                ╱     ╲
               ╱       ╲
              ╱         ╲
    ═══════════════════════════════

Pros: - Both motors are stationary (not on gantry) - Very light moving mass - Excellent acceleration capability - Popular design with extensive community support

Cons: - Requires rigid frame (usually aluminum extrusion) - Belt routing is complex and must be precise - Misaligned belts cause “racking” (diagonal prints) - Not compatible with threaded-rod frame philosophy - Requires specific pulley stacks and belt lengths - Harder to build from scavenged parts - Completely different lineage from RepRap Darwin

Verdict: Rejected—incompatible with M10 threaded rod frame design.

Option C: Parallel-Rail Sled (“The Plough”) ✓ SELECTED

Description:
Two horizontal M10 smooth rods spaced 50-80mm apart. A sled carriage rides on linear bearings between them. Single X-motor drives belt to move sled left-right. The entire X-assembly moves on Y-rails.

    [Y-Motor]═══════════════════[Y-Motor]
         ║     Y-smooth rods      ║
         ║                        ║
    [X-End]══════════════════[X-End]
         ║     X-smooth rods      ║
         ║       (60mm gap)       ║
         ║    ┌───────────┐       ║
         ║    │ THE PLOUGH│       ║
         ║    │ Pitan+V6  │       ║
         ║    └───────────┘       ║

Pros: - Simple—one motor per axis, straightforward kinematics - Extremely scavenger-friendly (any smooth rods, any bearings) - Easy to understand, maintain, and repair - Tolerant of frame imperfections (nuts adjust everything) - Native Klipper support (simple [stepper_x] config) - Direct lineage from Darwin → Mendel → Mendel i2 - M10 rods provide massive stiffness at 380mm span - Pitan/V6 can be centered between rods (balanced) - All components visible and accessible

Cons: - Higher moving mass on Y-axis (entire X-assembly moves) - Higher moving mass on X-axis (Plough + extruder + motor) - Lower theoretical maximum speed than CoreXY - X-motor rides on gantry (adds moving mass)

Verdict: Accepted—best fit for Amalgam philosophy.

Detailed Comparison Matrix

Factor Cross-Rod CoreXY The Plough
Simplicity ⭐⭐ ⭐⭐ ⭐⭐⭐⭐⭐
Scavengeability ⭐⭐ ⭐ ⭐⭐⭐⭐⭐
Maintainability ⭐⭐⭐ ⭐⭐ ⭐⭐⭐⭐⭐
Frame Tolerance ⭐⭐⭐ ⭐⭐ ⭐⭐⭐⭐⭐
Klipper Config ⭐⭐⭐ ⭐⭐⭐⭐ ⭐⭐⭐⭐⭐
Max Speed ⭐⭐⭐⭐⭐ ⭐⭐⭐⭐⭐ ⭐⭐⭐
Moving Mass (X) ⭐⭐⭐⭐⭐ ⭐⭐⭐⭐⭐ ⭐⭐⭐
Moving Mass (Y) ⭐⭐⭐⭐⭐ ⭐⭐⭐⭐⭐ ⭐⭐
Rigidity ⭐⭐⭐⭐ ⭐⭐⭐ ⭐⭐⭐⭐⭐
Darwin Heritage ❌ ❌ ✅
Repairability ⭐⭐⭐ ⭐⭐ ⭐⭐⭐⭐⭐

Why Moving Mass Penalty is Acceptable

The M10 Advantage

At 380mm span with M10 smooth rods: - Deflection under load is measured in microns - Natural frequency is very high (won’t resonate at print speeds) - The “bounce” that plagues M8 systems doesn’t exist

The moving mass penalty of The Plough is almost irrelevant when rails don’t flex.

Klipper Compensation

Klipper’s Input Shaping can compensate for the higher moving mass: - ADXL345 accelerometer measures actual resonance - Software shapes motor commands to avoid exciting those frequencies - Pressure Advance handles extruder response

Result: Quality prints at moderate speeds (150-250mm/s) with excellent surface finish.

Quality Over Speed Philosophy

Amalgam prioritizes: 1. Print quality (layer consistency, surface finish) 2. Reliability (prints complete successfully) 3. Repairability (fix it yourself) 4. Understanding (know how it works)

Speed is explicitly not a priority. A well-tuned Plough at 150mm/s produces better prints than a poorly-tuned CoreXY at 500mm/s.

The Darwin Lineage

The Plough design maintains authentic RepRap heritage:

RepRap Darwin (2007)
    │
    ├── Box frame with threaded rods
    ├── Sled-style X-carriage
    └── Simple Cartesian kinematics
         │
         ▼
RepRap Mendel (2009)
    │
    ├── Triangular frame (still threaded rods)
    ├── Sled X-carriage between parallel rods
    └── Same kinematic philosophy
         │
         ▼
Mendel i2 (2012)
    │
    ├── Refined sled design
    ├── Better bearing arrangements
    └── Still threaded rod construction
         │
         ▼
Amalgam (2025)
    │
    ├── M10 threaded rod frame (upgraded)
    ├── M10 smooth rod rails (upgraded)
    ├── "The Plough" sled (evolved)
    ├── Klipper brain (modernized)
    └── Same fundamental philosophy

The Eustathios/HercuLien and CoreXY come from different branches:

Ultimaker (2011)
    │
    └── Cross-rod / floating head
         │
         ▼
H-Bot → CoreXY (2012+)
    │
    └── Crossed belt systems
         │
         ▼
Eustathios / HercuLien (2014+)
    │
    └── Rod-based CoreXY variants

Amalgam should remain true to its Darwin ancestry.

The Plough: Design Specifications

Rod Spacing

60mm center-to-center (horizontal, top-down)

Rationale: - NEMA 17 motor is 42.3mm wide - 4mm plastic walls on each side = 8mm - ~10mm clearance for LM10UU housings - Pitan/V6 nests perfectly between rods

Bearing Arrangement

LM10UU linear bearings with split-clamp housings

  • Minimum 2 bearings per rod (4 total)
  • Spaced as far apart as carriage allows
  • Split-clamp design (no zip ties)
  • “Long wheelbase” prevents pitching during acceleration

Extruder Positioning

Pitan centered between rods, nozzle at geometric center

  • Weight balanced left-right
  • No torque trying to twist carriage
  • Belt attachment at carriage center (or balanced dual-attachment)

Belt Path

Single GT2 6mm belt, center-line routing

  • Belt runs parallel to X-rods
  • Centered at 30mm mark (halfway between 60mm-spaced rods)
  • Minimizes “cocking” force during acceleration

Structural Optimization

“Hollow Bone” Principle

The Plough should be designed with: - 4-6 perimeter walls (not high infill) - Structural ribs and webs - Cut-away weight reduction where possible - 3-5mm thick structural members

Why Not Solid?

Solid plastic: - Heavier (bad for moving mass) - Cools unevenly (warps during printing) - Wastes material

Ribbed structure: - Lighter - Cools evenly - Looks “engineered” - Actually stronger in bending

Integration with Other Systems

Dual Y-Steppers

The Plough design works perfectly with dual Y: - Each Y-motor pulls one side of the X-assembly - Klipper auto-squares on every home - No mechanical coupling needed between Y-motors

Triple Z with Kinematic Joints

Z-axis independence means: - Plough doesn’t need to compensate for bed tilt - Klipper’s z_tilt_adjust handles leveling - X-gantry only needs to be parallel to itself

Rubber Gasket Motor Isolation

X-motor on the Plough benefits from Strategy B sandwich mount: - Reduces vibration transmitted to gantry - Quieter operation - Klipper Input Shaping handles residual resonance

Rejected Optimizations

Carbon Fiber X-Rods

Considered: Replacing M10 steel with CF tubes for weight reduction.

Rejected: - Not scavenger-friendly (must purchase) - M10 steel is already stiff enough - Adds complexity to bearing interface - Against “hardware store” philosophy

Pancake X-Motor

Considered: Using shorter NEMA 17 to reduce moving mass.

Rejected: - Reduced torque - May not be available from donor printer - Standard NEMA 17 is fine for quality-focused printing

Bowden Extruder

Considered: Moving extruder off gantry to reduce mass.

Rejected: - Worse print quality (retraction tuning nightmare) - Against quality-over-speed philosophy - Direct drive is better for flexible filaments - Pitan is the reference spec

Future Upgrade Path

If users want more speed later, they can:

  1. Lighter Plough: Redesign with more aggressive weight reduction
  2. Pancake Motor: Swap X-motor for shorter variant
  3. Higher Acceleration: Increase Klipper accel values
  4. Better Belts: Upgrade to Gates GT3 belts

But the fundamental geometry—sled on parallel rails—scales well and doesn’t need to change.

Conclusion

The Parallel-Rail Sled (“The Plough”) is the correct choice for Amalgam because:

  1. It’s simple — One motor, one belt, two rods. Anyone can understand it.

  2. It’s scavenger-friendly — Photocopier rods, donor bearings, hardware store parts.

  3. It’s maintainable — All parts visible, accessible, replaceable with basic tools.

  4. It’s authentic — Direct descendant of RepRap Darwin and Mendel designs.

  5. It leverages M10 — The stiffness advantage negates the moving mass penalty.

  6. It prioritizes quality — Reliable prints over maximum speed.

  7. It’s a Tractor — Heavy, reliable, fixable with a wrench.

Amalgam should feel like a machine tool, not a race car.

Decision Record

Date Decision Rationale
2025 Selected Parallel-Rail Sled Best fit for scavenger philosophy
2025 Rejected Cross-Rod Too complex, wrong lineage
2025 Rejected CoreXY Incompatible with threaded rod frame
2025 Set 60mm rod spacing Optimal for Pitan nesting
2025 Specified split-clamp bearings Better than zip-ties
2025 Confirmed center-line belt Minimizes carriage twist

References

  • RepRap Darwin (2007) — Original box-frame threaded rod printer
  • RepRap Mendel (2009) — Triangular frame evolution
  • Mendel i2 (2012) — Refined sled carriage
  • Eustathios — Cross-rod gantry reference design
  • HercuLien — Large-format cross-rod printer
  • Ingentis — Modernized Cartesian sled design
  • V-Core (early versions) — Rod-based Cartesian reference