跳转到内容
YBO 技术文档

20261014-2350-air-hub-docs-follow-hardware Bring the Air Hub documents in line with the current schematics and boards, a 9–36 V supply and fixed core boards

  • status: completed
  • createdAt: 2026-10-14 23:50
  • approvedAt: 2026-10-14 23:58 (owner: "按这个执行，然后提交一下")
  • relatedTask: 20261014-2350-air-hub-docs-follow-hardware

Context

章节标题：“Context”

Documents: spec R1.7, design appendix R1.3, reference-board page R0.9, guide R1.3 (uncommitted work of task 20261014-0022 included). Hardware baseline: /srv/ybo/switch/handover/air-hub-2026-10-13/ README, BOMs and pin maps as refreshed on 2026-10-14 (hardware repo commit 6169219; the working tree has later uncommitted changes that are not used).

Differences found between documents and hardware:

  1. Supply: documents say 12/24 V nominal, 8–32 V continuous (spec §2.4, GW-05, §5, design and reference-board diagrams). The owner now sets 9–36 V. The schematic input stage was sized for 32 V: TVS 5.0SMDJ33A has a 33 V stand-off (below 36 V, it would conduct in normal operation), and its 53 V clamp sits under the −60 V reverse FET SQD50P06-15L. A TVS with 36 V stand-off (5.0SMDJ36A class, clamp about 58 V) keeps the clamp below the FET and below the LMR36520 65 V rating; this is a hardware change, recorded as an open item.
  2. Core boards: documents still list alternates (Forlinx FET3506J-S, MYiR MYC-YR3506J for Pro; Ebyte ECK41-E1N2-N for Air Hub). The hardware is drawn for the Vanxoak land pattern only: HD-RK3506G-128F256LW (Air Hub) and HD-RK3506J-256F256LI (Pro).
  3. G board temperature: documents say −20…+70 °C; the vendor data sheet (handover core-board-evb-check.md) gives −20…+80 °C.
  4. Mini is already on ESP32-C6-MINI-1-N4 in hardware (since 2026-10-13, rev B). The reference-board page §7.4 and guide §8 written in task 20261014-0022 call the Mini schematic and PCB "WROOM-1, superseded"; that is wrong. Mini pins also differ: PAIR on IO4, system LED on IO9, radio LED on IO8, IO15 spare with pull-down.
  5. Carrier pins (Pro / Air Hub): seven GPIOs moved off LCD pins (MG21_RESET_N 121, C6_EN 122, C6_BOOT 81, WDT_WDI 116, LED_CAN 115, CMT_FCSB 117, CMT_GPIO3 118); 22 Ω series resistors on RMII, SDMMC CMD/D0–D3 and the watchdog reset; core-board land pattern redrawn from the vendor DXF (35.4 mm pad pitch, pin 1 top-left counter-clockwise) and a 30 × 31 mm slot under the core board.
  6. Reset and keys: no separate RESET key; a long press (about 5 s) of PAIR resets through 1 MΩ / 4.7 µF, 74LVC1G14 and 2N7002 (Pro: TPS3823 MR_N; Mini: module EN); holding PAIR after reset restores factory settings. Pro has PAIR, RECOVERY, MASKROM and four LEDs; Mini has PAIR and three LEDs.
  7. Boards: 6 layers (JLC06161H-3313), Pro rev C 110 × 102 mm, Mini rev B 90 × 80 mm; documents say 4 layers and rev B. Common-mode chokes are wired per data sheet (windings 1–4 / 2–3); the CAN choke is TDK ACT45B; the EFR32MG21 32.768 kHz crystal is 7 pF (ABS07); USBLC6 channels swapped.
  8. Verification state: ERC and netlist check 0 mismatches; DRC 0 unconnected / 0 clearance on both boards; SPICE subcircuit run (Pro 23 pass / 17 warn / 0 fail, Mini 21 / 4 / 0); EMC rule check 72 items (20 errors) on Pro, 64 (30 errors) on Mini.
  9. Diagrams: the 19 schematic and PCB images in bizplan/product/diagrams/ date from 2026-10-13 early morning (Mini images are the WROOM-1 board).

Proposal

章节标题：“Proposal”
  1. Spec R1.8: supply 12/24 V nominal, 9–36 V continuous in §2.4, GW-05, §5 and the brief; core boards fixed to one Vanxoak part number each in §1, §2.1, §2.2 (H12, H17), §2.3, §2.4 and AIR-O-03 (other boards kept only as evaluation records); G board −20…+80 °C in H17, AH-19, §2.4 environment and AIR-O-03; new open item in AIR-O-04 for the input stage at 36 V (TVS stand-off, clamp versus reverse FET and buck rating).
  2. Design appendix R1.4: §2.1 scheme I and §2.2 / §2.5 core-board rows to the fixed parts and +80 °C; power rows and §3 diagram to 9–36 V.
  3. Reference-board page R1.0, rewritten against the handover package: §2.1 / §2.2 diagram and pin table (moved GPIOs, series resistors, land pattern and slot), §2.3 C6 rows, §2.4 Mini diagram and pin table (IO4 / IO8 / IO9 / IO15), §3 power tree at 9–36 V, §4 circuit notes (long-press reset, chokes, crystal, USBLC6, series resistors), §5 part list from the BOM files, §7 board facts (6 layers, rev C / rev B, verification results, EMC check counts) with §7.4 corrected to the MINI-1 board, §8 open items (36 V input stage, vendor confirmations: board outline 37.9 mm, stencil thickness, the seven re-purposed GPIOs).
  4. Guide R1.4: §8 handover note corrected (Mini material is current); D0 gate lists the 36 V input-stage change.
  5. Diagrams: replace the PCB renders with the handover pcb_top.png / pcb_bottom.png and re-render the schematic sheets from the exported SVGs (conversion only, in a labelled sibling container; no KiCad run, no hardware file touched).
  6. Platform evaluation §4 / §8.6 and the temperature decision record: G board −20…+80 °C, core boards fixed. Changelog, task and plan status.

Governance files (product plan, open questions) are being edited by task 20261014-1417; only a one-line supply/temperature correction is made there if still needed after that task, otherwise left to it.

Risks

章节标题：“Risks”
  • The hardware working tree has uncommitted changes after the handover refresh; documents follow the handover package and may lag the next hardware commit.
  • The 36 V input-stage parts are a recommendation in documents only until the hardware owner changes the schematic.

Scope

章节标题：“Scope”

About 10 documentation files and 19 diagram images; no KiCad project files.

Alternatives

章节标题：“Alternatives”
  • Keep 8–32 V and only sync the rest (rejected by the owner's 9–36 V decision).
  • Leave diagrams as they are and only fix text.

Annotations

章节标题：“Annotations”
  • 2026-10-14: owner set the supply to 9–36 V and fixed the core boards.