Power Architecture
How power is distributed across an OpenSkyhawk cockpit. The firmware/hardware contracts cover the logic-side rails in detail; high-current actuator boards are not yet designed and are marked TBD below.
Distribution
A PC ATX power supply feeds the main bus with 12 V and 5 V, distributed between controller groups over Molex Mini-Fit Jr connectors. Each board generates its own 3.3 V locally — the 3.3 V rail is never distributed across the bus.
| Rail | Source | Used for |
|---|---|---|
| 12 V | ATX PSU, main bus | LED backlight strings; servos, when fitted, via a local 5/6 V buck (never off logic 5 V); input to local 5 V buck on high-current boards |
| 5 V | ATX PSU, main bus | Input to the local 3.3 V regulator; stepper driver motor supply (VM) |
| 3.3 V | Generated on each board | STM32 / RP2040 logic, MCP23017, ADS1115, SN65HVD230 |
Local 3.3 V regulation
Every MCU and breakout board carries an AMS1117-3.3 LDO (SOT-223) that drops the bus 5 V to 3.3 V. The 1.7 V drop is acceptable at the boards' logic load (≤ ~175 mA). Local decoupling is required on every board: 100 nF + 10 µF per rail, placed close to each IC.
For the rare board that needs significant 5 V current, an AP63205WU switching buck (SOT-23-6) converts 12 V → 5 V on-board. This is not used on standard MCU or breakout boards — only on future high-5V-current boards.
Never use a linear regulator for 12 V → 5 V
The 12 V → 5 V step is done with a switching buck (AP63205WU) where needed, never an LDO — the drop would dissipate too much power. The AMS1117 LDO is only for the small 5 V → 3.3 V step.
Board power budget — logic + LED boards
Standard logic + LED boards are designed to stay within ≤ 500 mA at 12 V input (a servo-driving board runs higher — see Servos below):
| Rail | Typical | Max expected |
|---|---|---|
| 12 V → LEDs | 54–180 mA | ~360 mA (large panel) |
| 12 V → AP63205 input (if fitted) | ~100 mA | ~150 mA |
| Total 12 V per board | ~160–280 mA | ~510 mA (large-panel edge case) |
| 5 V → stepper driver VM | 15–30 mA | 50 mA |
| 3.3 V → STM32 + MCP23017 + CAN | ~125 mA | ~175 mA |
Across a full ~20-board cockpit that's roughly 3 A at 12 V today (LED-only), plus ~2 A at 5 V — a small fraction of what the ATX supply delivers. No servos are fitted yet; when they are, powering them off 12 V (below) adds ~1–1.5 A per console → ~6–8 A cockpit-wide, so leave headroom. The build uses a fully-modular GOLDEN FIELD NX650, heavily over-provisioned; a 350–450 W single-rail ATX would be ample.
Servos — a 12 V load (planned guideline)
No servos in the cockpit yet — this is how they should draw power when added. Run them off the 12 V line through a local 5/6 V buck, never logic 5 V: five to six servos per console draw ~2–3 A running (up to ~6 A stall), which on the shared 5 V logic rail means brownouts and I²C glitches. On 12 V the same load is ~1–1.5 A per console (the buck roughly halves rail current) and stays isolated from logic; add a local bulk cap + polyfuse. A servo-driving board therefore exceeds the ≤ 500 mA logic+LED figure above — budget servos separately as a 12 V load, captured by that console's 12 V current monitoring.
Source protection — fused per-console injection
The ATX feed splits into per-console segments, and each segment is injected through its own fuse per rail. A console's fuse is sized to that console's draw and protects every board downstream of it on that rail — a fault in one console opens only that console's fuse, sparing the rest of the cockpit and its harness.
- One fuse per rail per console, at the injection point (12 V and 5 V). GND and CAN are never fused — they stay common and continuous as the bus reference and return.
- Sized to consumption; copper sized to the fuse. The rating sits above the segment's peak draw (with margin), and the fused power traces are then sized to clear it — 5 mm of 1 oz copper carries ~7 A, comfortably above the 5 A / 4 A blanket. (A 1 mm trace is ~2.5 A — that's a local sub-1 A tap only, never the fused trunk.) A console draws little (~1 A at 12 V), so slow-blow fuses in the low single amps are typical — slow-blow because LED strings and buck inputs inrush at power-on.
- Granularity follows consumption. A console is the default injection zone; a power-hungry group can get its own dedicated PDU — 12 V/5 V straight from the PSU, its own fuse sized to just that group — so it is isolated and does not load the shared console feed. The PSU sources ~54 A at 12 V (ample headroom); the per-zone fuse exists to protect the downstream harness and contain a fault, not to ration the supply.
- The PSU's own OCP protects the PSU, not your harness — a single-rail ATX only trips at tens of amps, long after a thin lead would cook. The per-zone fuse is what actually protects the wiring.
- A planned power-distribution CAN node adds live per-console current telemetry and active eFuse breaking on top of — not instead of — these baseline fuses.
Exact values live in the hardware source
Fuse ratings, trace widths, and injection copper are specified in the base-boards hardware source; this page is the overview.
Cross-tier notes
- CAN transceiver (SN65HVD230) runs directly from the 3.3 V rail — no level shifter.
- RP2040 SimGateway is bus-powered from USB. If it's co-located with STM32 CAN hardware, share GND only — do not tie the RP2040 module's 3.3 V to the STM32 board's 3.3 V.
- Stepper driver motor supply (VM) runs from the 5 V rail. The driver is a DRV8833 dual H-bridge (four logic inputs per motor).
TBD — not yet specified
Marked TBD because it isn't in source material yet
- Actuator boards (solenoids, large steppers, servos) are not yet designed. Servos' power sourcing is already decided (12 V → local buck, above); the rest — flyback protection, dedicated supply rails, driver selection — is open. Each will need its own power-budget analysis before PCB work. The established boundary is that logic + LED boards stay ≤ 500 mA at 12 V; actuator boards are sized to their specific loads and require separate design review.