Memristor cnn

sebishniraula·memristor-cnn·simulation/connectivity.net

This repository ships a schematic but no board layout.

About the Memristor cnn PCB

Memristor cnn is an open source RP2040/RP2350 PCB design by sebishniraula, published on GitHub. It is a board, with 86 components from 42 distinct parts.

Its main chip is the RaspberryPi_Pico, from the RP2040/RP2350 family. Other key parts include the TLV172IDCK, DAC80508ZRTET, ADS1115IDGS, MAX333ACPP+ and TLV9004PWR. By type, the board carries 25 capacitors, 25 resistors, 13 test points, 9 ICs, 3 diodes and 1 module.

From the project

Memristor based convolution neural network to predict digits(eventually)

These scripts reproduce the prototype's math and validate its exported wiring. They do not operate hardware or implement CNN training.

Use Python 3.10+ from the repository root. No pip dependencies are needed.

For the analog DC cases, install the ngspice command-line program and run:

The script writes simulation/results/dc-results.json; that generated directory is ignored by Git. --netlist selects another KiCad S-expression netlist and --output changes the JSON destination. The expected pins and component roles are specific to this design.

Main components on the Memristor cnn

TLV172IDCKDAC80508ZRTETADS1115IDGSMAX333ACPP+TLV9004PWRRaspberryPi_Pico

Memristor cnn bill of materials (BOM)

86 components, 42 distinct parts.

QtyPartRefs
4
TLV172IDCK
U1, U2, U8, U9
1
DAC80508ZRTET
U4
1
ADS1115IDGS
U5
2
MAX333ACPP+
U6, U7
1
TLV9004PWR
U10
1
RaspberryPi_Pico
A1
1
POWER: +5V / GND / -5V
J1
1
WRITE ARM
JP1
1
EXT CELL A1
MR1
1
EXT CELL A2
MR2
1
EXT CELL A3
MR3
1
EXT CELL A4
MR4
1
EXT CELL B1
MR5
1
EXT CELL B4
MR6
1
EXT CELL B3
MR7
1
EXT CELL B2
MR8
1
GND
TP1
1
+3V3
TP2
1
+5VA
TP3
1
-5VA
TP4
Show 22 more
QtyPartRefs
1
REF_BUF
TP5
1
PROG_CMD
TP6
1
PROG_LIMITED
TP7
1
PROG_BUS
TP8
1
OUT_A
TP9
1
OUT_B
TP10
1
ADC_A
TP11
1
ADC_B
TP12
1
ADC_BASE
TP13
3
BAT54S
D1, D2, D3
16
100n
C1, C3, C5, C10, C11, C12, C13, C14 +8
1
4.7u
C2
1
220n
C4
2
10u / 16V
C6, C7
2
100p C0G
C8, C9
3
10n
C18, C19, C20
1
10k
R1
2
4.7k
R2, R3
7
10k 0.1%
R5, R6, R7, R8, R11, R14, R17
4
1k
R9, R12, R15, R18
3
30k 0.1%
R10, R13, R16
8
100k
R19, R20, R21, R22, R23, R24, R25, R26

Memristor cnn design files

The KiCad project lives in the sebishniraula/memristor-cnn repository on GitHub; these links point at the commit this page was built from.

Memristor cnn: common questions

What microcontroller does the Memristor cnn use?

The Memristor cnn is built around the RaspberryPi_Pico, from the RP2040/RP2350 family.

How many components are on the Memristor cnn?

86 components, from 42 distinct parts. The full bill of materials is listed on this page.

Where can I download the Memristor cnn design files?

From the sebishniraula/memristor-cnn repository on GitHub, which has a netlist; the links under Design files point to each one.

Can I use the Memristor cnn design in my own project?

The repository has no license, so all rights stay with sebishniraula. Use it as a reference, and ask the author before reusing the design.

Can I test firmware for the Memristor cnn without the hardware?

Yes. HardLabs builds a simulation of the board from its netlist and BOM, so you can run and debug RP2040/RP2350 firmware against it before you order a PCB.

Similar boards