nerve_ada_board

Nerve dev board using off-the-shelf modules.

Note: This design is now no longer under active development. The final changes to the main branch are published as release tagged v0.1.0-beta, see the release notes: 1.1 v0.1.0-beta.

💡Hint: To clone this repository use:

git clone --recursive

---

Table of Contents

• nerve_ada_board
  • 1 Release Notes
    • 1.1 v0.1.0-beta
      • 1.1.1 LSE Crystal
      • 1.1.2 External Clock Crystal Tuning
      • 1.1.3 GPS
  • 2 Overview
    • 2.1 Function Connectors
    • 2.2 Switches & Jumpers
    • 2.3 Plug-in Modules
    • 2.4 Breakout Connectors
  • 3 Development
    • 3.1 Worflow
    • 3.2 Submodules
  • 4 Production History
    • 4.1 v0.1.0-alpha-7b18788

---

1 Release Notes

1.1 v0.1.0-beta

Following the initial validation of components and software on this PCB design, development has shifted to a more production-oriented board: nerve_pcb. While the base components remain similar, many have been modified, with further changes expected.

During the development of the new design, several hardware and PCB issues were identified. Although this README.md will no longer be updated, the following sections document the known issues as of the final release.

1.1.1 LSE Crystal

The current LSE crystal on the STM32F446RE is the ABS25-32.768KHZ-6-T.

This is not the intended component; the correct crystal should have been the ABS25-32.768KHZ-T, which has a more appropriate load capacitance specification of 12.5 pF. The ABS25-32.768KHZ-6-T variant, with a 6 pF load capacitance, was used inadvertently.

1.1.2 External Clock Crystal Tuning

The crystal circuits use calculated capacitors that have not yet been fully verified or fine-tuned. Specifically, the 32.768 kHz LSE crystal is likely using incorrect capacitors with a value of 1.5 pF.

1.1.3 GPS

The GPS module (SAM-M10Q) had not been verified to be functional as of the release. However, this is likely due to software limitations rather than issues with the hardware or PCB design.

---

2 Overview

Development board for the STM32 based nerve controller firmware.

2.1 Function Connectors

These connectors are fixed by hardware (PCB traces or the connector itself).

Connector	Ref	Description
Tag-Connect TC2050	J1	Programming/debug connector
Hinge microSD Card	J2	Portable storage, see Molex product video
USB-C 5 V Power	J3	Power only USB-C, primary 5 V source
CR1220 Battery	BT1	MCU and GPS battery source 3.3 V supply
CAN1 (Transceiver U7)	J7	Pin 1: CAN1 High, Pin 2: CAN1 Low
CAN2 (Transceiver U8)	J8	Pin 1: CAN2 High, Pin 2: CAN2 Low

2.2 Switches & Jumpers

User controllable hardware and/or firmware driven inputs.

Switch/Jumper	Ref	Description
MCU NRESET Switch	SW1	Generic 6 mm TH button, push to reset
MCU PA0 Switch	SW2	Generic 6 mm TH button, designed for SYS_WKUP0
MCU PC13 Switch	SW3	Generic 6 mm TH button, designed for SYS_WKUP1
SDIO Card Detect Jumper	J4	Close for card inserted (pull-down on close)
BOOT0 Jumper	J5	Open for run flash memory (pull-down on open)

2.3 Plug-in Modules

Off-the-shelf plug-in boards, primarily from Adafruit.

Manufacturer Part Number	Ref	Manufacturer	Hardware Interface	Quantity	Description
4816	U3	Adafruit Industries LLC	2.54 mm pitch TH	1	BMP390 Module
Digi XBee-PRO 900HP	U5	Digi	2 mm pitch TH	1	Long Range 900 MHz OEM RF Module
4754	U6	Adafruit Industries LLC	2.54 mm pitch TH	1	BNO085 Module
5708	U7, U8	Adafruit Industries LLC	2.54 mm pitch TH	2	TJA1051T/3 Module

2.4 Breakout Connectors

Simple breakout connectors, driven fully or partially by firmware. See schematic and firmware docs for details.

Connector	Ref	Logic	Description
WS2812B Extender	J6	5 V	Pin 1: PWM, Pin 2: Ground, use separate supply (preventing over current draw) and common ground
SPI Breakout	J9	3.3 V	Pin 1: CS, Pin 2: SCK, Pin 3: MISO, Pin 4: MOSI
GPIO Breakout	J10	5 V	Pin 1, 2, 3: GPIO 1, 2, 3
PWM Breakout	J11	5 V	Pin 1, 2: PWM 1, 2
UART Camera	J12	5 V	Pin 1: 5 V, Pin 2: UART RX, Pin 3: UART TX, Pin 4: Ground, design for low current camera

---

3 Development

3.1 Worflow

Developed using KiCad with KiBot GitHub Actions workflow, kibot.yaml.

• KiBot configurations found in config.kibot.yaml.

CI/CD objectives:

1. ERC (Electrical Rule Check).
2. DRC (Design Rule Check).
3. PCB docs.
   • PCB layer docs (.pdf and .svg).
4. PCB manufacturing.
   • Gerber files (.grb).
   • Drill file (.drl and .pdf reference docs).
5. Assembly (fabrication service & hand assembly).
   • Pick-and-place assembly position file (.pos).
   • Component distributor ordering BoM (.csv).
   • General BoMs (.xlsx and .html).
   • Component placement guide (.html).
6. CAD.
   • (.step).

3.2 Submodules

This repository includes the submodule component_library, located at: component_library.

When cloning the repository, make sure to use the recursive option:

git clone --recursive

To update the submodule to its latest version in a cloned repository, use the following command:

Note: This is not recommended for this deprecated repository, as pulling updates from the submodule's latest commit may result in incompatibilities.

git submodule update --remote --merge

---

4 Production History

4.1 v0.1.0-alpha-7b18788

Commit: 7b18788.

Manufacturing & Assembly Details

Manufacturing Information:

Production date: 2024-10-22.

• Manufacturer: JLCPCB.
• Order: PCB, stencil.
• PCB specs (non-default):
  • PCB quantity: 5.
  • Surface Finish: LeadFree HASL.
  • Impedance control stackup: JLC041611-7628.
    • Outer Copper Weight: 1 oz.
    • Inner Copper Weight: 1 oz.
• Stencil specs (non-default):
  • Stencil quantity: 1.
  • Custom size: 100 mm x 100 mm (matching PCB dimensions).

Assembly Information:

Assembly date: 2024-11-09 to 2024-11-13.

• Assembled PCB quantity: 2.
• Process:
  • Stencil solder paste print by hand & soldering iron.
    • Solder paste: Chip Quik TS391LT Bi57.6Sn42Ag0.4.
  • SMD pick-and-place by hand.
  • Reflow with oven.
    • Roughly correct temperatures to solder paste docs.
    • Approximately doubled soak time due to oven limitations.

1. PCB 1 Notes:
   • Components: All placed except for GPS.
   • Damage to though holes for connector J4 mostly on bottom side.
     • First assembly was unsatisfactory.
     • Attempt to remove solder to reassemble header pins resulted in damage from aggressive force with the soldering iron.
     • Final result was functional, but the board damage remains on bottom side.
       • Suggested correction: heat soak the board, note also left in board evaluation section of this writeup.
     • Continuity tests showed no issues after completion.
   • Bridged pins on the U2 (STM32F446RE).
     • Likely too much solder paste and less than ideal definition of the printed solder paste.
     • Resolved with flux and reheat with soldering iron as well as use of solder wick.
       • Suggested correction: Cleaner swipe of the stencil with less solder paste.
2. PCB 2 Notes:
   • Components: All placed.
   • Misplacement of U2 (STM32F446RE).
     • Component shifted seemingly during reflow, some sides of the MCU had short or unconnected pins.
     • There was likely too much solder as well as potential misalignment during the pick-and-place by hand.
     • Attempts were made to reposition the component.
     • Some solder mask was removed near the pads, however no significant functional concerns were visible.
     • In order to limit potential damage to the board or components ( particularly the expensive GPS module), the component was removed.
     • No pad damage was visible.
     • A replacement component was hand soldered with a soldering iron.
       • Suggested correction: Improve solder paste printing (again) and better hand placement. If too much paste was visible retry before any reflow. Perhaps, even holding the component in position with thin strips of kapton tape.
   • Bridged pins on the U9, U10 (TXB0104QPWRQ1).
     • Same details as U2 on PCB 1 (see above).
   • U4 GPS module solder joints are not visible (pads under component), but a continuity test of nets in every combination pair showed no shorts.

Board Evaluation:

• Large planes, especially ground. It is difficult to adjust through hole components that have ground pin(s).
  • Requires heat soaked board or potentially risky soldering iron temperatures.
• Some components such as resistors or capacitors are very close to through hole header pin connectors/jumpers.
  • Not super difficult to deal with, but something to keep in mind.
• The XBee Pro 900HP can cause clearance issues when trying to press-release the nearby TC2050-IDC's plastic clips.
• The Molex 472192001 hinge microSD card slot is functional, but slightly blocked by the Adafruit TJA1051T/3 breakout board.
  • This Would be easily resolved by implementing CAN circuit/components directly onto the PCB.