PCB Net Routing Issues
0:26
Thought for a moment
You’re right to be frustrated. At this point, I can still help diagnose nets, verify schematic connectivity, clean stale traces, run DRC, and export once clean — but I currently cannot draw arbitrary manual PCB copper traces/vias from this chat session unless Flux exposes that manual routing action to the AI session.
To let me finish it automatically, Flux would need to add AI-accessible support for:
- drawing a trace between two pads on a named net
- choosing layer, width, and path points
- placing vias tied to a net
- routing differential pairs with spacing/length constraints
- editing/deleting individual trace segments reliably
You do not need to add this yourself unless you are developing Flux. As a user, the practical path is:
- Manually route the listed nets in the PCB editor using the Trace/Route tool.
- Route in this order:
- FSYNC → GND
- EP pads → GND vias/pour
- LOUT+ → speaker connector pin 1
- RTS/CTS/RXD → TXS0104E ↔ SIM7600G
- USB D+/D− last, carefully as a short matched pair
- After each net, run DRC and confirm that airwire disappears.
- Once you finish, ask me to “run DRC and export Gerbers,” and I can validate and generate the manufacturing files if clean.
For USB D+/D− specifically: keep the two traces parallel, same length, close together, avoid vias, don’t cross gaps in the ground plane, and keep them away from LTE/GPS/RF and switching regulator areas.
Manual route steps
List unrouted nets
USB routing guidance
Check stale traces
Routing checklist
Finished working
Refine this doc
Ask about this doc
Project Specification: AR11 AI Driver Safety Device
Source
Uploaded file:
AR11_Hardware_Spec_v1.pdfProduct Overview
AR11 is a dashboard-mounted AI driver safety device for vehicles. It monitors the driver and road simultaneously using two cameras, GPS, a G-sensor, local AI processing, and optional 4G/WiFi cloud connectivity. It is intended for African roads, vehicles, and climate conditions and must operate offline for core safety functions.
Major Functional Blocks
Processing and AI
- Main processor: Raspberry Pi Compute Module 4, 8GB RAM, 32GB eMMC.
- AI accelerator: Google Coral Edge TPU, specified as M.2 A+E, 2230 form factor.
- Local AI features: drowsiness, yawn, head nod, gaze/distraction, seatbelt, phone use, face recognition, fatigue scoring.
- OS/software: Raspberry Pi OS, Python, MediaPipe, TensorFlow Lite, Coral acceleration.
Cameras
- Driver camera: Sony IMX219 IR, 8MP, 1080p30, CSI-2, IR night vision, automatic IR cut filter, 62 degree FOV.
- Road camera: Sony IMX477, 12.3MP, 1080p60, CSI-2, 140 degree wide angle, low-light lens.
- Requirement: two 22-pin CSI-2 FPC connectors.
GPS and Connectivity
- GPS: u-blox NEO-M8N, UART to CM4, 3.3V, 23mA average, external SMA antenna.
- Cellular: SIMCom SIM7600G, LTE Cat-4 with 3G/2G fallback, nano SIM slot, UART + USB to CM4, 3.8V supply.
- WiFi/Bluetooth: provided by CM4; WiFi 802.11 b/g/n/ac 2.4GHz + 5GHz, Bluetooth 5.0.
- Antenna requirements: SMA connectors and 50 ohm RF traces for LTE/GPS.
Motion Sensing
- G-sensor: MPU6050 over I2C.
- I2C connection: SDA to CM4 GPIO2, SCL to CM4 GPIO3.
- Events include hard braking, collision, sharp cornering, hard acceleration, rollover, tamper detection, potholes.
Audio and User Interface
- Status LEDs: green, orange, red.
- Speaker: built-in 2W speaker with PAM8403 amplifier.
- Buzzer: passive 5V piezo, must use transistor driver such as 2N2222; do not drive directly from GPIO.
- Reset button required.
Storage
- OS/software: CM4 32GB eMMC.
- Recording: removable MicroSD, 64GB to 256GB.
- MicroSD slot required, with ESD protection.
Power Requirements
- Primary input: 12V vehicle input from fuse box.
- Main regulator: buck converter, 12V to 5V/3A.
- Backup battery: 3.7V 2500mAh LiPo, approximately 30 minutes runtime after engine off.
- Battery charger: TP4056 charging circuit.
- Typical power: 8-12W.
- Peak power: 15W with 4G active.
- Standby/sleep: 500mW.
- Protection required:
- Reverse polarity protection.
- Ignition spike protection/capacitor bank.
- Overvoltage protection.
- Overcurrent protection.
- ESD protection on external connectors.
- Power isolation requirement: keep 12V and 5V power areas separated; no mixing of 12V and 5V areas.
PCB Requirements
- Board size: 95mm x 70mm.
- Layers: 4 minimum: signal, power, ground, signal.
- Thickness: 1.6mm.
- Copper: 1oz outer, 0.5oz inner.
- Finish: ENIG.
- Solder mask: black.
- Silkscreen: white.
- Trace widths: 0.2mm signal minimum, 0.8mm power minimum.
- Vias: 0.3mm drill, 0.6mm pad minimum.
- Impedance control: 50 ohm for USB and camera traces.
- Operating temperature: -20C to +85C; automotive-grade components.
Key PCB Design Instructions
- CM4 connector: two official Hirose DF40C-100DS-0.4V connectors.
- Coral M.2 slot: A+E key, 2230.
- Camera connectors: two 22-pin CSI-2 FPC connectors.
- GPS UART: NEO-M8N TX to CM4 UART0 RX GPIO15; NEO-M8N RX to CM4 UART0 TX GPIO14.
- 4G UART: SIM7600G to CM4 UART1 GPIO8/GPIO10 with hardware flow control.
- MPU6050 I2C: SDA to GPIO2, SCL to GPIO3.
- USB-C: USB 2.0 only; connect D+/D- to CM4 USB lines with proper termination.
- Thermal: add thermal vias under CM4 module and SIM7600G.
- ESD: TVS diodes on USB-C, SIM, SD card, and other external connectors.
- SIM: ESD protection such as PRTR5V0U2X or similar on all SIM lines.
- LED resistors: calculate for 3.3V GPIO; spec suggests 100 ohm for bright LEDs.
Mechanical and Enclosure Requirements
- Main body: 110mm x 80mm x 35mm aluminum enclosure.
- Material: aluminum alloy 6061, matte black anodized.
- IP rating: IP54.
- Operating temperature: -20C to +85C.
- Mount: dashboard suction cup, 80mm industrial suction base, adjustable 100-150mm arm.
- Camera structure: two adjustable hinged camera sections with ball joints.
- External access: SMA antenna ports, USB-C service port with rubber plug, SD slot with rubber plug, sealed 12V cable gland.
- PCB mounting: four M3 threaded standoffs.
Manufacturing Deliverables Requested by Spec
- Gerber files.
- BOM in JLCPCB-compatible format with LCSC part numbers where possible.
- CPL/component placement list for assembly.
- STEP model for enclosure verification.
- Schematic PDF.
- KiCad project files for future modifications.
Important Mismatch With Current Flux Project Context
The current Flux project was named and described as a USB-C powered WiFi/BLE environmental temperature/humidity sensor. The uploaded PDF instead specifies an automotive AI driver safety device using 12V vehicle power, Raspberry Pi CM4, dual cameras, LTE, GPS, LiPo backup, speaker, MicroSD, and an aluminum enclosure. Before schematic work starts, the project intent should be confirmed and the project name/description should likely be updated to AR11 AI Driver Safety Device if this PDF is the intended target.
Current Schematic Status
As of extraction, the Flux schematic contains 0 components and 0 nets. No schematic or layout changes have been made from this uploaded spec.
Welcome 👋
Flux helps you build PCBs faster with an AI teammate!
Create your account to collaborate, stay updated, fork your own version, and get instant answers from our AI agent.