I have been learning to control LED matrix panels with CircuitPython code running on an Adafruit Matrix Portal S3. So far, I have a few P4 64x32 panels from Ali Express and a P2.5 64x32 panel from Adafruit. They all have 5 address lines and work fine with the Matrix Portal. The P4 panels have RUL2038S driver chips and the P2.5 panel has ChipOne ICN2037BP driver chips.
I then bought 2 x P2 128x64 panels on Ali Express but when they got here, I discovered that they only have address lines A, B, C and no HUB75 output connectors. They have ChipOne ICND2038S LED driver chips which I could not find any information about on GitHub, but I did find the ICN2038S on a list of LED driver chips that were a problem to control.
I naively thought that if I could find 128x64 panels with 5 address lines, they would work with the Matrix Portal S3. I found a vendor on Ali Express that had P2 128x64 panels with 5 address lines and HUB75E input and output connectors, so I bought 4 of them. When they arrived, I could not get them to do anything. These panels use Shenzen Fuman FM6363C driver chips. After many hours of scouring GitHub and various forums, I don’t think I have any chance of making these panels work with the Matrix Portal S3. I learned that people have cracked some of these complex LED driver chips and can control them with a Raspberry Pi. I found one guy who got a panel with FM6363C chips to work but with a different controller.
Before I send these panels back, does anyone here know if it is possible to control these panels with a Matrix Portal S3? It’s uses an ESP32-S3-WROOM-1 MCU. I have been working with CircuitPython but could switch to Arduino IDE if necessary. I have learned (the hard way) that most current higher-density panels now use various “Gen-3” LED driver chips that have complex timing signal requirements. I am beginning to think that I have no chance of finding a 128x64 panel that I can connect to my Matrix Portal S3 and make it go. Any insights are appreciated.
I would avoid Arduino IDE with the Matrix Portal S3. I recently ported a couple thousand lines of code to one and it crashed a lot during development. Usually when inserting Serial.printfs to debug.
Espressif has released a new version of the Arduino library but I no longer have the 64x64 matrix I was working with so can’t comment. It seems all their development is going towards Python these days.
Thanks for the tip. I prefer CircuitPython but am willing to switch to Arduino IDE if there are libraries that work with the panels I bought. I am beginning to think there is no such thing as a P2 128x64 panel that uses ICN2037BP or RUL2038S which are chips that I know work with the Matrix Portal S3.
Adafruit sells P2 64x64 panels made by Shenzen Coreman. In the product photo I can see that they use ICN2037BP driver chips. They should work with the Matrix Portal S3 but they are $50 each. I found what I thought were the exact same panels from Coreman on Ali Express for $12.50 each. Even with shipping from China, they are still less than half the cost of the Adafruit panels. Coreman told me that they used to sell Adafruit P2 64x64 panels with ICN2037BP chips but the last few batches have used FM6124 chips. They said these chips are functionally identical. Do you know if the ICN2037BP and FM6124 driver chips are the same?
If FM6124 is the same as ICN2037BP then I could take a chance and order the P2 64x64 panels from Coreman through Ali Express, but run the risk of them getting here and not working like the last 6 panels I bought on Ali Express. Coreman also just told me they have P2 128x64 panels that use ICN2037BP / FM6124. If this is true, that is what I am really looking for. I am assuming that if the panels have 5 address lines and use ICN2037BP or FM6124 that they will work with the Matrix Portal S3. Only one way to find out I guess.
I don’t know the internals of the matrix and I no longer have it. Was a birthday present.
The only issue I had with the 2023 panel from adafruit was that Red and Blue were reversed. Was an easy fix once I figured out why my screens were weird. Maybe the chip change explains the reversal.
Thanks. I bought a P2.5 64x32 panel from Adafruit that has green and blue reversed. There is a tiny note about this “slight oddity” at the very bottom of the product description that I did not see. I was not very happy when I discovered this. They said just flip them in the code, which is easy to do, but then that code is not correct for “normal” panels. I bought the Adafruit panel to test the Matrix Portal S3 with because I wanted to start with a panel known to work with the Matrix Portal, and so I could ask Adafruit support for help if I had problems. It annoyed me that the cheap 64x32 panels from Ali Express worked fine and the 1 panel I bought from Adafruit had the colors swapped.
I downloaded spec sheets for the ICN2037BP and FM6124… Best I can tell, they are identical. I can live with a color flip as long as all the panels are the same. I will probably buy the P2 64x64 or P2 128x64 panels from Coreman on Ali Express and pray that they work with the Matrix Portal. I already have 2 panels with ICND2038S and 4 panels with FM6363C chips that do not work with the Matrix Portal S3 that I am going to return.
Thank you. I am reading this now. I ended up returning the 2 P2 128x64 LED panels that only had address lines A, B, C, ICND2038S driver chips, and HUB75 inputs only. Even I I could have gotten them to work with the Matrix Portal S3, I need to daisy-chain them so the missing HUB75 output connectors was a deal killer. I also sadly returned the 4 x P2 128x64 panels that had 5 address lines but used the FM6363C LED driver chips which the Matrix Portal S3 does not work with. I am still searching for P2 128x64 panels that have 5 address lines and use the ICN2037BP or FM6124 driver chips. Shenzen Coreman said they had them and sent a picture of the back but now they are saying they only have panels with 3 address lines. Maybe I can make these work with the info at the link you provided. If not, it may be time to give up on the Matrix Portal S3 and get a Raspberry Pi. Thanks again.
I agree that HUB75 connectors do not seem to adhere to a rigid pinout. I have encountered multiple variations that were quite confusing until I realized that some panels use different pinouts that I guess correspond to specific receiver cards. Its bad enough that on Ali Express you may get panels with 5 address lines, or 3, or 2, and LED driver chips that may or may not work with a specific controller. I thought I had found the elusive P2 128x64 panels with FM6124 driver chips and 5 address lines from Coreman LED but they just sent me a pic of a panel with address lines A, B, C and said they don’t have the panels they sent pics of a few days ago. I am close to giving up.
5 address lines are not a strict requirement. As you have already been told here and on another site, working with 3 address lines is very simple and is described, for example, in the post by @Irakli_Jibladze1 above. So, given that 128x64 panels has a little options at all, don’t limit yourself to just 5-line panels
I noticed your username and realized that you have responded to some of my posts on github, so you already know my situation. Most of what I have found that works with panels that only have 2 or 3 address lines seems to be written for a Raspberry Pi or STM32 board. I did not think the Adafruit Matrix Portal S3 would work with these libraries, but I may be wrong about this. The Matrix Portal S3 has an ESP32-S3 MCU and I found pin mappings for it in mrcodetastic’s and your ESP32-HUB75-MatrixPanel-DMA documentation.
Regarding what @Irakli_Jibladze1 posted, I did read it and sort of understand what he did, but not clearly enough to implement it. I am fairly new to working with these HUB75 LED panels and find a lot of what you guys are doing is over my head. The 64x32 panel I bought from Adafruit with ICN2037BP chips and from Ali Express with RUL2038S chips work fine with the Matrix Portal S3. The panels I got with ICND2038S and FM6363C chips did not work. Based on what I learned about the FM6363C, I don’t think I had any chance of getting those panels to work with the Matrix Portal S3. I never found the ICND2038S listed anywhere but I did find the ICN2038S listed on the Pixelmatrix GitHub site under incompatible chipsets and figured the ICND2038S was probably similar to the ICN2038S.
I think I need to take your advice and give up on my search for 128x64 panels with 5 address lines and ICN2037 or FM6124 chips and focus on learning about how to control panels with more complex driver chips. I do appreciate your patience with a newbie, and any insights you care to provide. Thanks.