ADR-020: Dual-8 Scavenger Variant

Status

Superseded by ADR-021: Dual-Rod Motion System

Note: This ADR proposed dual 8mm rods as a scavenger “variant.” The analysis proved compelling enough that dual 8mm was adopted as the reference specification in ADR-021, not merely an alternative. This ADR is preserved for historical context on the engineering analysis that led to that decision.

Original Status

Proposed

Context

The reference Amalgam specification uses M10 smooth rods for XY motion (ADR-003). While M10 provides excellent stiffness at 370mm spans, M10 smooth rods are difficult to scavenge:

• Photocopiers sometimes have M10/10mm shafts, but storing/disposing of the rest is problematic
• Most donor 3D printers (Prusa clones, Anet A8, Creality) use 8mm smooth rods
• Buying new M10 smooth rods costs $60-80 AUD, pushing builds toward $300+

Analysis shows that single 8mm rods cannot achieve acceptable deflection at Amalgam’s target 370mm X-span:

Configuration	Deflection @ 370mm	Verdict
Single 8mm, 280g toolhead	0.072mm	Excessive
Single 10mm, 280g toolhead	0.030mm	Acceptable

However, the Voron Legacy achieves acceptable performance with 8mm rods by using a smaller build volume (~200mm). This raised the question: Can Amalgam maintain 235mm build volume with 8mm rods?

The Dual-Rod Insight

Beam deflection follows: δ = F × L³ / (48 × E × I)

When two parallel rods share the load: - Each rod carries F/2 - Deflection halves: δ_dual = δ_single / 2

This means dual 8mm rods can match or exceed single 10mm performance.

Decision

We define a “Dual-8 Scavenger” variant that uses dual 8mm smooth rods per axis with vertically-stacked configuration, enabling full 235mm build volume with scavenged parts from two donor printers.

Rod Configuration

X-Axis (Vertical Stack):

═══════════════════════════════  Top 8mm rod
         [Plough carriage]
═══════════════════════════════  Bottom 8mm rod

Y-Axis (Dual per Side):

Left side:              Right side:
  ║══8mm══║               ║══8mm══║
  ║ belt  ║               ║ belt  ║
  ║══8mm══║               ║══8mm══║
  LM8LUU                  LM8LUU

Deflection Analysis

Configuration	360mm span, 280g	Verdict
Single 8mm	0.066mm	Excessive
Dual 8mm (shared load)	0.033mm	Acceptable
Single 10mm	0.027mm	Acceptable

The dual-8 configuration achieves 50% less deflection than single-8, bringing it within acceptable tolerance for quality printing.

Stiffness Comparison

Second moment of area (I = π×d⁴/64):
  Single 8mm:  I = 201 mm⁴
  Dual 8mm:    I_eff = 402 mm⁴ (load shared)
  Single 10mm: I = 491 mm⁴

Effective stiffness ratio: Dual 8mm / Single 10mm = 0.82
With load sharing: Dual 8mm deflection < Single 10mm deflection

Why Vertical Stacking?

Vertical vs horizontal rod spacing provides identical sag resistance, but different moment resistance:

Orientation	Resists	Vulnerable to
Horizontal (reference Plough)	Yaw (twist L-R)	Pitch (nose-dive)
Vertical (Dual-8)	Pitch (nose-dive)	Yaw (twist L-R)

Vertical stacking on X-axis is preferred because: - Pitch forces (from Y-acceleration) dominate toolhead dynamics - Voron Legacy validates this configuration - Allows belt to run between rods (cleaner routing)

Long Bearings (LM8LUU)

Long linear bearings are recommended but not required: - Do NOT reduce rod deflection (that requires more rods or larger diameter) - DO improve carriage stability (anti-racking) - DO provide better moment resistance - DO give smoother motion (more ball contact)

Standard LM8UU bearings work; LM8LUU provides better carriage stability.

Consequences

Benefits

• Full scavengeability: 2 donor Prusa clones provide all 8-10 smooth rods needed
• Maintains build volume: 235×235×250mm preserved
• Lower cost: Eliminates $60-80 smooth rod purchase
• Proven physics: Dual-rod load sharing is well-understood beam mechanics
• Voron-validated: Vertical stacking proven in Voron Legacy

Trade-offs

• More rods: 6 XY rods vs 4 in reference design
• More bearings: 12-16 LM8UU vs 8 LM10UU
• Redesigned brackets: X-ends and Y-pucks need dual-rod mounts
• Tighter toolhead mass budget: Keep under 280g for best results
• Slightly less stiff: 82% of M10 stiffness (but deflection is better due to load sharing)

What This Enables

• True 2-donor build: Motors, boards, rods, bearings all from donors
• Sub-$150 hardware cost: When combined with scavenged electronics
• MKS SKIPR path: $130 AUD for SKIPR + CANbus + ADXL still hits <$300 total

BOM Implications (Generic)

Scenario A: Two Prusa Clone Donors (Recommended)

• Parts scavenged:
  • 12× 8mm smooth rods (need 8-10)
  • 24× LM8UU bearings (need 12-16, keep spares)
  • 10× NEMA17 motors (need 6-7 for triple-Z)
  • 2× control boards (use best one, keep spare)
  • 2× heated beds (use best one)
  • 2× power supplies
• Parts to buy:
  • M10 threaded rod for frame skeleton (~$45 AUD)
  • MKS SKIPR + EBB36 CANbus + ADXL345 (~$130 AUD)
  • E3D V6 clone hotend (~$15 AUD)
  • Belts, pulleys, misc hardware (~$30 AUD)
• Cost implication: ~$220-250 AUD total
• Build volume: Full 235×235×250mm

Scenario B: Two Anet A8 Donors

• Parts scavenged: Similar to Scenario A
• Note: Anet beds are smaller (220×220), may need bed upgrade
• Cost implication: ~$250-280 AUD (if bed upgrade needed)

Scenario C: Mixed Donors (Prusa + Ender)

• Challenge: Ender 3 uses V-slot rollers, not smooth rods
• Parts scavenged from Ender: Motors, board, bed, PSU
• Parts scavenged from Prusa: Smooth rods, bearings
• Cost implication: May need to buy additional rods (~$30 AUD)

Scenario D: Single Donor + Purchased Rods

• Not recommended: Defeats scavenger philosophy
• If necessary: Buy 8mm precision ground rods (~$40-50 AUD)
• Cost implication: Higher than dual-donor path

Implementation Notes

Rod Count Summary

Axis	Reference (M10)	Dual-8 Scavenger
X	2 rods	2 rods (stacked)
Y	2 rods	4 rods (2 per side)
Z	2 rods	2 rods
Total	6 rods	8 rods

Bearing Count Summary

Axis	Reference	Dual-8
X carriage	4× LM10UU	4× LM8UU (or LM8LUU)
Y gantry (per side)	2× LM10UU	4× LM8UU
Z	4× LM10UU	4× LM8UU
Total	12× LM10UU	16× LM8UU

Toolhead Mass Budget

For acceptable deflection at 360mm span: - Target: ≤280g total toolhead mass - Pitan + pancake NEMA17: ~200-250g ✓ - Pitan + standard NEMA17: ~280-320g (borderline) - Wade + standard NEMA17: ~400g+ ✗ (too heavy)

CAD Implications

New parametric parts needed: - x_end_dual8.py: X-end brackets for dual vertical rods - y_puck_dual8.py: Y-pucks for dual stacked rods per side - plough_vertical.py: Plough carriage for vertical rod stack

Config parameter:

SMOOTH_ROD_CONFIG = "DUAL_8"  # vs "SINGLE_10" for reference

Klipper Configuration

No changes to kinematics—still Cartesian with: - [stepper_x]: Single motor, single axis - [stepper_y]: Dual motors with [stepper_y1] for auto-squaring - [stepper_z], [stepper_z1], [stepper_z2]: Triple-Z

Cost Analysis: Hitting $300 Target

Component	Reference Spec	Dual-8 Scavenger
Frame (M10 threaded)	$45	$45
Smooth rods	$70 (M10)	$0 (scavenged)
Bearings	$30 (LM10UU)	$0 (scavenged)
Motors	$60	$0 (scavenged)
Heated bed	$35	$0 (scavenged)
Electronics	$130 (SKIPR+CAN)	$130 (SKIPR+CAN)
Hotend	$15	$15
Belts/pulleys/misc	$30	$30
Total	~$415	~$220

The Dual-8 Scavenger path achieves significant cost reduction while maintaining the full 235mm build volume—well under the $300 AUD target.

References

• ADR-003: Precision Smooth Rods (reference M10 specification)
• ADR-005: Triple-Z Independent Kinematic Leveling (requires 2 donors anyway)
• Voron Legacy: Validates 8mm vertical stacking
• Beam Deflection Theory: Simply-supported beam equations
• Conversation: “Voron Legacy Comparison” (2025-01-27)