ADR-007: CAN Bus Architecture

Status

Accepted

Context

The Amalgam has multiple electrical components that need to communicate: - Main board (motion control) - Toolhead (extruder, hotend, fan) - Z-probe (BLTouch/PINDA) - Future expansions (ERCF, toolhead boards, bed heater, etc.)

In 2026, there are two main communication architectures: - Traditional wiring: Every component wired directly to main board with individual cables - CAN bus: Network-based communication with shared twisted-pair wiring

Traditional 3D printers use direct wiring, but high-end systems (Voron 2.4, toolhead boards) use CAN bus for reduced wiring and improved reliability.

Decision

We choose CAN bus support as an optional upgrade path for Amalgam, with traditional wiring as the baseline.

Architecture Options

Tier 1-2: Traditional Wiring (Baseline)
- All stepper motors wired to main board
- Toolhead components wired with multicore cable
- BLTouch wired directly to board
- Simple, proven, works with all boards
Tier 3-4: CAN Bus (Optional Upgrade)
- CAN-connected toolhead board (e.g., BTT EBB36)
- CAN bus for expansions (ERCF, bed heaters, etc.)
- Reduced wiring (4 wires: VCC, GND, CAN-H, CAN-L)
- Requires CAN-capable main board

Why Optional CAN Bus?

Wiring reduction: 20+ wires reduced to 4 for toolhead
Noise immunity: Differential signaling is electrically robust
Expandability: Easy to add CAN-connected devices
Modern standard: Used in Voron, RatRig, high-end systems
Future-proof: Supports toolhead boards, smart peripherals

Consequences

Benefits (CAN Bus)

Reduced cable clutter: 4 wires vs 20+ for toolhead
Improved reliability: Fewer connections = fewer failure points
Noise resistance: CAN bus immune to EMI from steppers
Flexible mounting: CAN nodes can be placed anywhere
Daisy-chain: Multiple devices on single bus
Hot-swappable: Some CAN boards support hot-plug

Trade-offs (CAN Bus)

Cost: Additional CAN board(s) ~$30-50 AUD
Complexity: Requires CAN-capable main board
Configuration: More complex Klipper setup
Debugging: Hardware issues harder to diagnose
Compatibility: Not all boards support CAN

Benefits (Traditional Wiring - Baseline)

Simple: Works with all boards, no special hardware
Proven: Decades of community knowledge
Cheap: No extra boards required
Easy debugging: Each wire individually testable
Universal: Works with any board/donor

Trade-offs (Traditional Wiring)

Cable clutter: Wiring looms can be messy
More connections: More potential failure points
EMI sensitive: Long cable runs susceptible to noise
Fixed architecture: Adding devices requires running new cables

BOM Implications (Generic)

Scenario A: Traditional Wiring (Recommended for Tier 1-2)

Parts needed:
- Main board (MKS SKIPR, BTT SKR 3, etc.)
- Standard stepper motor cables (4-pin)
- Multicore toolhead cable (4-6 conductors)
- BLTouch cable (3-5 pin)
- Hotend cable (heater, thermistor, fan)
Cost implication: Included in main board price
Donor compatibility: All donors
Board requirement: Any board with 6+ drivers
Experience: Simple, proven setup

Scenario B: CAN Toolhead Only (Tier 3)

Parts needed:
- Main board with CAN support (MKS SKIPR, BTT SKR 3, BTT Octopus)
- CAN toolhead board (BTT EBB36, EBB42)
- CAN bus cable (twisted pair, 4 wires)
- 24V power for toolhead board
- Reduced toolhead cabling (just 4 CAN + 24V)
Cost implication: Medium (+$35-50 AUD for EBB36)
Donor compatibility: All donors (board upgrade)
Board requirement: MUST have CAN port
Benefits: Cleaner toolhead wiring, noise immunity

Scenario C: Full CAN System (Tier 4)

Parts needed:
- Main board with CAN support
- CAN toolhead board (BTT EBB36/EBB42)
- CAN ERCF controller (Happy Hare CAN)
- CAN bed heater controller
- Multiple CAN bus segments (terminators as needed)
- CAN bus cables throughout
Cost implication: High (+$100-150 AUD for all CAN boards)
Donor compatibility: All donors (board upgrade)
Board requirement: Multiple CAN ports or CAN hub
Benefits: Minimal wiring, highly modular, future-proof

Scenario D: Hybrid CAN (Recommended Path)

Parts needed:
- Main board with CAN support (MKS SKIPR)
- CAN toolhead board (future upgrade)
- Traditional wiring for now
- Upgrade toolhead to CAN later
Cost implication: Medium (buy CAN-capable board, add EBB36 later)
Donor compatibility: All donors
Board requirement: CAN-capable board
Experience: Traditional start, upgrade to CAN as needed

Scenario E: Legacy Board (Tier 1)

Parts needed:
- Salvaged board (Creality 4.2.2, Anet, etc.)
- All traditional wiring
Cost implication: Very Low ($0)
Donor compatibility: Use donor board
Board requirement: Legacy board (no CAN)
Limitation: Cannot upgrade to CAN without board replacement

Implementation Notes

CAN Bus Topology

Main Board (CAN Controller)
  |
  +-- CAN Termination (120Ω)
  |
  +-- Toolhead CAN Board (EBB36)
  |
  +-- CAN Hub (if needed)
        |
        +-- ERCF CAN Controller
        +-- Bed Heater CAN Board
        +-- Other CAN peripherals
  |
  +-- CAN Termination (120Ω)

MKS SKIPR CAN Support

Native CAN Port: Yes
CAN Protocol:  CAN FD (backward compatible with CAN 2.0)
Max Nodes:     64 (theoretical)
Recommended:   Start with 1-3 nodes

CAN Cable Requirements

Type:           Twisted pair (CAN-H and CAN-L)
Shielding:      Optional but recommended (EMI)
Termination:    120Ω resistor at each end
Length:         Up to 40m (NEO-DARWIN: < 2m typical)
Wire gauge:     22-24 AWG

CAN Toolhead Board (EBB36 Example)

Drivers:        2x TMC2209 (for dual gear or direct drive)
Sensors:        Thermistor, ADXL345 (onboard)
Outputs:        Heater, Fan, Part fan
Inputs:         BLTouch, Probe
Power:          24V (5V generated onboard)
Communication:  CAN bus
Size:           36mm × 36mm

Traditional vs CAN Wiring Comparison

Traditional Toolhead Wiring:
- Extruder motor:  4 wires
- Hotend heater:   2 wires
- Hotend therm:    2 wires
- Part fan:        2 wires
- Hotend fan:      2 wires
- BLTouch:         3-5 wires
- Total:           15-17 wires + bulky cable management

CAN Toolhead Wiring:
- CAN bus:        4 wires (VCC, GND, CAN-H, CAN-L)
- 24V power:      2 wires (from main PSU)
- Total:          6 wires (clean loom)

Board CAN Compatibility

MKS SKIPR:     ✓ Native CAN (recommended)
BTT SKR 3:     ✓ Native CAN
BTT Octopus:   ✓ Native CAN
Creality 4.x:  ✗ No CAN
Anet A8:       ✗ No CAN
Prusa boards:  ✗ No CAN (proprietary)

Klipper CAN Configuration

# MCU config for CAN-connected toolhead
[mcu canbus]
canbus_uuid: [UUID from EBB36]

# Extruder on CAN toolhead
[extruder]
step_pin: canbus_gpio:26
dir_pin: canbus_gpio:22
enable_pin: !canbus_gpio:10
...

CAN UUID Discovery

# On MKS SKIPR or Pi host
~/klippy-env/bin/python -m klippy.canbus_query
# Press BOOT button on CAN board to get UUID

CAN Bus Health Monitoring

Klipper command: CAN_BUS_STATUS

Monitors:
- Bus errors
- Node status
- Message rates
- Error counters

Normal: < 10 errors/hour
High:   > 100 errors/hour (check terminators, noise)

Termination

Required: 120Ω resistor at BOTH ends of CAN bus
- One on main board
- One on furthest node
- Remove terminators from middle nodes

Diagnostic: Measure resistance between CAN-H and CAN-L
- With power off: 60Ω (two 120Ω in parallel)
- With terminators: Correct
- Without terminators: Open circuit

Recommended Path for Amalgam

Tier 1-2:
  - Use traditional wiring
  - Save money
  - Focus on frame mechanics first

Tier 3 (MKS SKIPR):
  - Buy CAN-capable board
  - Start with traditional wiring
  - Plan for CAN toolhead upgrade

Tier 4:
  - Full CAN system from start
  - Toolhead CAN board
  - Clean wiring loom
  - Future ERCF CAN integration

Common CAN Issues

No communication: Check terminators, verify 24V power, check UUID
High error rate: Check for loose connections, add shielding, reduce cable length
Node not detected: Press BOOT button to get UUID, check canbus_uuid in config
Intermittent faults: Add CAN bus shield, separate from stepper cables

References

BTT EBB36 Documentation
Klipper CAN Bus Guide
Voron 2.4 CAN Bus Setup
docs/AI-Conversations/ [Relevant conversations about CAN bus architecture]