I’ve had a few people contact me about the SACO LED matrix I posted info on here: https://erykzagorski.com/index.php/2020/07/26/saco-led-matrix-reverse-engineering/

I have noticed that there is a bit of wrong information on this post, mainly the diagram I created to show how the LEDs are connected to the drivers. I have created some new diagrams to better explain how this is logically connected.

Update on hardware

The FETs are not connected to the ‘rows’ of the matrix as I showed in my first post, they are actually connected to the ‘columns’ in groups of 4 like this:

The TLC5940’s are connected to each set of 4 columns and will clock data out onto the LEDs like so if all FETs are on:

And like this if one FET is on:

Update on software

I have also been asked about the software I used to control the matrix. My initial tests to control the panel were done with an Arduino UNO and this TLC5940 library https://playground.arduino.cc/Learning/TLC5940/

I had to modify this library to spit out 3 serial data streams instead of one as well as update the number of TLC drivers connected in series. My arduino code also had to switch the relays and display the other columns of LEDs.

To get the spiral demo I showed in my original post working I had to use a FPGA to drive the panel along with an ESP32 to clock image data into the FPGA over SPI. I eventually had 6 of these panels working together using the FPGA. If you are interested in the verilog code for this please reach out to me by e-mail.