Skip to the content
4 layers30 × 30 mmNeuromorphic ir sensor compact, the printed circuit board in 3D

Neuromorphic ir sensor compact

RyanHailey21/neuromorphic-IR-hand-assembled

All through-hole re-implementation of the Rev A neuromorphic IR event sensor: no SMD parts, solderable with a basic iron.

Size
30 × 30 mm
Copper layers
4
Thickness
1.6 mm
Parts
121
Nets
62
Pads
465
Tracks, vias
1,301, 212
Pours
3
Routed
Yes
KiCad
KiCad 10.0 (file 20260206)
Created
27 Sep 2026
Updated
27 Sep 2026

Overview

Neuromorphic ir sensor compact is an open-source printed circuit board by RyanHailey21, from the RyanHailey21/neuromorphic-IR-hand-assembled repository on GitHub. It is a 4-layer board of 30 × 30 mm with 121 parts on 62 nets, routed with 1,301 tracks and 212 vias.

It is built around the SN74LVC1G123DCUR, BLM15AG601SN1D11 and AO3400A, with the TSAL6200 among its other parts. Its parts include 38 capacitors, 35 resistors, 15 ICs, 5 diodes and 1 transistor. It belongs with the sensors and measurement boards and the FPGA and CPLD boards of this library.

Main components

4

README

README.md

Neuromorphic IR Event Sensor — Rev A-THT (hand-assembly)

All through-hole re-implementation of the Rev A neuromorphic IR event sensor: no SMD parts, solderable with a basic iron. Same architecture as the SMD design spec: 4 photodiodes → TIA → temporal adaptation → ON/OFF comparators → one-shots → 8 asynchronous 3.3 V spike outputs for the Tang Nano 20K.

Top

  • Board: 100 × 100 mm, 2 layers, 1.6 mm FR4, 0.3 mm min track, GND pour on both layers
  • Status: ERC 0 · DRC 0 · schematic↔PCB parity 0 · 0 unconnected
  • Fab files: production/neuromorphic-ir-sensor-tht-revA-THT-gerbers.zip (upload as-is to JLCPCB/PCBWay, bare board only)
  • BOM: production/neuromorphic-ir-sensor-tht-BOM.csv · Schematic: docs/schematic.pdf · Assembly drawing: docs/assembly_top.pdf

Changes from the SMD spec

FunctionSMD Rev ARev A-THTEffect
TIA ×4OPA3812× MCP6022-I/P (dual, DIP-8)RF/CF unchanged (18k / 47 pF); ~188 kHz BW, large phase margin
Comparators ×84× TLV32024× MCP6542-I/P (dual, DIP-8)~4 µs response, irrelevant vs 2.2 ms adaptation
One-shots ×88× 74LVC1G1234× 74HC123 (dual, DIP-16)18k / 12 nF ≈ 100 µs (K≈0.45) — tune on bench
VREFREF3312 1.250 VLM385BZ-1.2 1.235 V + 4.7kThresholds ≈ +40.6 / −39.7 mV
LED switchAO3400APN2222A + 1k base~22 mA LED current with 82 Ω
PhotodiodesVBPW34FASROSRAM BPW 34 F (daylight filter)Same die, THT

The FPGA connector follows the spec (keyed 2×5: GND, 3V3, SPIKE_ON/OFF × 4). The AER / 2×10 pinout in the old README had no driving circuitry and was dropped.

Connectors

RefPins
J1 FPGA (2×5 box)1 GND · 2 3V3 · 3 ON_TL · 4 OFF_TL · 5 ON_TR · 6 OFF_TR · 7 ON_BL · 8 OFF_BL · 9 ON_BR · 10 OFF_BR
J2 DEBUG1 GND · 2 VREF · 3–6 VPHOTO_TL/TR/BL/BR (via 100 Ω)
J3 LED_EN1 3V3 · 2 LED_EN · 3 GND — jumper 1-2 = LED on, 2-3 = off, or drive pin 2 from FPGA
J4 RAW_PROBE1 GND · 2–9 RAW_ON/OFF_TL, TR, BL, BR
J5 EVENT_PROBE1 GND · 2 3V3 · 3 3V3A · 4 VREF · 5 VTH_ON · 6 VTH_OFF · 7–10 VEVENT_TL/TR/BL/BR

Assembly notes

  • Buy 6× DIP-8 and 4× DIP-16 sockets; fit ICs last.
  • Solder lowest parts first: resistors → ceramic caps → sockets → TO-92 → electrolytics → headers → photodiodes/LED.
  • Polarity: photodiode cathode (marked side) → square pad; LED1 cathode (flat) → square pad; electrolytic + marked on silk; U1/Q1 flat side per silk.
  • R18–R21 (1 MΩ hysteresis) are DNP — fit only if comparators chatter.
  • Tuning parts (adaptation R/C, one-shot R/C, thresholds, RLED) can be fitted in machined-pin socket strips for tool-free swapping.
  • Keep the photodiode cluster area clear for the M12 lens housing.

Regenerating (headless)

Everything is generated from .companion/scripts/design.py (netlist source of truth). With KiCad closed:

bash
python .companion/scripts/gen_schematic.py --force
"C:/Program Files/KiCad/10.0/bin/python.exe" .companion/scripts/gen_pcb.py --force
python .companion/scripts/setup_rules.py
# autoroute (kicad-companion autoroute_board / Freerouting), then:
"C:/Program Files/KiCad/10.0/bin/python.exe" .companion/scripts/finalize_pcb.py
python .companion/scripts/setup_rules.py

Close KiCad before regenerating: it rewrites the .kicad_pro from memory on exit and drops the net classes.

Open items

  • Re-run SPICE with Microchip's MCP6022 / MCP6542 vendor models (not yet in the repo).
  • Bench-trim the 74HC123 pulse width (K varies with supply voltage).
  • The TIA input nets (photodiode + RF + CF) are autorouted at 22–36 mm total length, well above the spec's 2–8 mm photodiode→op-amp target. The op-amps sit back from the cluster to leave room for the lens housing. Worth a manual tidy-up in KiCad if noise turns out to be a problem on the bench.

Parts

121
QtyValueFootprintReferences
22100nFCapacitor_SMD:C_0201_0603MetricC3, C11, C13, C15, C17, C18, C19, C20, C21, C22, C24, C26, C28, C30, C32, C34 +6
9100kResistor_SMD:R_0201_0603MetricR17, R19, R21, R23, R25, R27, R29, R31, R35
8150Resistor_SMD:R_0201_0603MetricR18, R20, R22, R24, R26, R28, R30, R32
81nFCapacitor_SMD:C_0201_0603MetricC23, C25, C27, C29, C31, C33, C35, C37
8SN74LVC1G123DCURPackage_SO:VSSOP-8_2.3x2mm_P0.5mmU8, U9, U10, U11, U12, U13, U14, U15
418k 0.1%Resistor_SMD:R_0201_0603MetricR5, R7, R9, R11
41MResistor_SMD:R_0201_0603MetricR13, R14, R15, R16
422kResistor_SMD:R_0201_0603MetricR6, R8, R10, R12
447pFCapacitor_SMD:C_0201_0603MetricC10, C12, C14, C16
4BPW 34 FASOptoDevice:Osram_BPW34S-SMDD1, D2, D3, D4
4TLV3202AIDGKRPackage_SO:VSSOP-8_3x3mm_P0.65mmU4, U5, U6, U7
310uFCapacitor_SMD:C_0402_1005MetricC1, C2, C38
210k 0.1%Resistor_SMD:R_0402_1005MetricR2, R3
2M2MountingHole:MountingHole_2.2mm_M2H1, H2
2OPA2381AIDRBRPackage_SON:VSON-8-1EP_3x3mm_P0.65mm_EP1.65x2.4mmU2, U3
1100Resistor_SMD:R_0201_0603MetricR34
11uFCapacitor_SMD:C_0201_0603MetricC4
1301k 0.1%Resistor_SMD:R_0402_1005MetricR4
13V3ATestPoint:TestPoint_Pad_D1.0mmTP1
1499k 0.1%Resistor_SMD:R_0402_1005MetricR1
1600R@100MHzInductor_SMD:L_0402_1005MetricFB1
182Resistor_SMD:R_0402_1005MetricR33
1AO3400APackage_TO_SOT_SMD:SOT-23Q1
1FPGAConnector_PinHeader_1.27mm:PinHeader_2x05_P1.27mm_Vertical_SMDJ1
1GNDTestPoint:TestPoint_Pad_D1.0mmTP2
1LED_ENJumper:SolderJumper-3_P1.3mm_Bridged12_Pad1.0x1.5mmJP1
1LED_ENTestPoint:TestPoint_Pad_D1.0mmTP6
1RAW_OFF_BLTestPoint:TestPoint_Pad_D1.0mmTP20
1RAW_OFF_BRTestPoint:TestPoint_Pad_D1.0mmTP22
1RAW_OFF_TLTestPoint:TestPoint_Pad_D1.0mmTP16
1RAW_OFF_TRTestPoint:TestPoint_Pad_D1.0mmTP18
1RAW_ON_BLTestPoint:TestPoint_Pad_D1.0mmTP19
1RAW_ON_BRTestPoint:TestPoint_Pad_D1.0mmTP21
1RAW_ON_TLTestPoint:TestPoint_Pad_D1.0mmTP15
1RAW_ON_TRTestPoint:TestPoint_Pad_D1.0mmTP17
1REF3312AIDCKRPackage_TO_SOT_SMD:SOT-323_SC-70U1
1TSAL6200LED_THT:LED_D5.0mm_IRGreyLED1
1VEVENT_BLTestPoint:TestPoint_Pad_D1.0mmTP13
1VEVENT_BRTestPoint:TestPoint_Pad_D1.0mmTP14
1VEVENT_TLTestPoint:TestPoint_Pad_D1.0mmTP11
1VEVENT_TRTestPoint:TestPoint_Pad_D1.0mmTP12
1VPHOTO_BLTestPoint:TestPoint_Pad_D1.0mmTP9
1VPHOTO_BRTestPoint:TestPoint_Pad_D1.0mmTP10
1VPHOTO_TLTestPoint:TestPoint_Pad_D1.0mmTP7
1VPHOTO_TRTestPoint:TestPoint_Pad_D1.0mmTP8
1VREFTestPoint:TestPoint_Pad_D1.0mmTP3
1VTH_OFFTestPoint:TestPoint_Pad_D1.0mmTP5
1VTH_ONTestPoint:TestPoint_Pad_D1.0mmTP4

More from RyanHailey21/neuromorphic-IR-hand-assembled

1
  • 2 layers100 × 100 mmNeuromorphic ir sensor tht, the printed circuit board in 3D

    Neuromorphic ir sensor tht

    RyanHailey21/neuromorphic-IR-hand-assembled

    All through-hole re-implementation of the Rev A neuromorphic IR event sensor: no SMD parts, solderable with a basic iron.

    102 parts

Similar boards

  • 4 layers160 × 100 mmLunar Hawks Rover Control, the printed circuit board in 3D

    Lunar Hawks Rover Control

    Matthew-Garcia/NASA-Lunabotics-UHCL-Lunar-Hawks

    UHCL Lunar Hawks proof-of-concept excavation rover: ROS 2, micro-ROS, MicroPython, control PCB development, CAD, and research.

    207 partsMIT
  • 2 layers50 × 70 mmNeuromorphic ir event sensor, the printed circuit board in 3D

    Neuromorphic ir event sensor

    RyanHailey21/neuromorphic-ir-event-sensor

    Neuromorphic Infrared Event Sensor (Rev A) - 4-Quadrant Temporal Contrast Analog Front-End in KiCad 10

    123 parts
  • 4 layers220 × 120 mmCandidate filled profile, the printed circuit board in 3D

    Candidate filled profile

    misko/prompt2device

    The separate vNext PCB skill and orchestrator package is available alongside the reference workflow below; existing projects keep their accepted toolchain until an explicit migration.

    569 parts3MIT
  • 2 layersEphemerkey, the printed circuit board in 3D

    Ephemerkey

    ephemerkey/ephemerkey

    The consumer half of the ephemerkey system, and the second PCB in this repo (alongside hardware/ephemerkey/, sharing the same scripts/kschgen.py engine, Makefile, and lib tables — the multi-project layout used by…

    70 parts
  • 6 layers36.5 × 36.5 mmH743 Ardupilot, the printed circuit board in 3D

    H743 Ardupilot

    OliviezGit/FC_01

    Le dossier composants contient les fiches techniques PDF, bibliothèques locales, STEP, vues de contrôle et la nomenclature France du 27/09/2026.

    156 partsCERN-OHL-S-2.0
  • 4 layers69.6 × 29.4 mmSpartan6 board, the printed circuit board in 3D

    Spartan6 board

    openconcepts-ar/accel2d

    This project aims to provide a graphics server system based on hardware accelerated graphics, and a easy way to develop the graphics primitives.

    169 parts1
  • 4 layers69.9 × 44.5 mmDoughLee2, the printed circuit board in 3D

    DoughLee2

    chl33/Dough133

    A temperature-controlled container for proofing and fermenting soughdough

    62 parts3MIT
  • 6 layers46 × 76 mmQ2-v2, the printed circuit board in 3D

    Q2-v2

    0x616c6f6e/esp32s31-mp3

    打开 q2-v2.kicadpro。这是 ESP32-S31NRV16 裸芯片版的独立工程,当前按用户要求改为 0402 阻容、电源分区铺铜并就近打过孔,信号待手动布线。不是可打板文件。原生产主板 hardware-q2 的 PCB 保持原文件哈希。

    198 parts

Questions

How big is the Neuromorphic ir sensor compact PCB?

It measures 30 × 30 mm and has 4 copper layers, 1.6 mm thick.

What parts does the Neuromorphic ir sensor compact use?

121 parts, built around the SN74LVC1G123DCUR, BLM15AG601SN1D11, AO3400A and TSAL6200. The full list of its parts is on this page, with links to each part's symbol and footprint where PCBKit's library has them.

Where can I download the Neuromorphic ir sensor compact KiCad files?

From the RyanHailey21/neuromorphic-IR-hand-assembled repository on GitHub, at the commit this page was made from: the board is compact/neuromorphic-ir-sensor-compact.kicad_pcb and its schematic compact/neuromorphic-ir-sensor-compact.kicad_sch. PCBKit opens the board in the browser, with nothing to install.

Can I use the Neuromorphic ir sensor compact design in my own project?

Its repository states no licence, so its author keeps all rights: ask RyanHailey21 before reusing the design.