weird things happening with 64x64 matrix and rpi4

[carmonkeylab]

hi, i am using an rpi4 and a 64x64 matrix. i used to be running Raspbian lite and everything was good. then i decided to switch to dietpi since it was lightweight. i installed the repo but when i run sudo ./demo -D0 --led-rows=64 --led-cols=64 --led-slowdown-gpio=5, this happens (yes, i broke a small part of the matrix) sometimes it does that and sometimes it does something similar to what it does in the pic. also the photo coloring is a little off. i have suspected it was probably a IDC cable problem so i am using going to buy a new one. the reason i suspect it is a cable problem is because sometimes when the cable wiggles a little the display shows nothing. is there any other way to trouble shoot the matrix and make sure it is the cable and not anything else.

[carmonkeylab]

ok, but the weird thing is that it was working and then stopped working

[carmonkeylab]

Red is ICND2038S a24268ja Blue is mw245b Green is ruc7258 Purple is 74hc048

i will also send better quality pictures

[carmonkeylab]

[carmonkeylab]

i will check the continuity

[carmonkeylab]

no problem with continuity

[carmonkeylab]

Do you have the original image to jump back to?

You are using Pi4 which is pretty fast compared to the other Pi models. It could be GPIO slowdown of 5 is too low with dietPi. You will need to hack the code to fix that if it ends up being the case.

no i do not have the original raspbian lite image but i put a new fresh one which results in the same thing happening

[carmonkeylab]

i measured the distance between the bonnet and the boards and gave the pics so you can tell me if that is good or if it is supposed to go down more

[hdm]

Hello! I am running into the same issue. Using the 64x64 with the RGB bonnet on an RPi 4B+ with the default gpio mapping shows the same two green lines as the original poster. Switching to the adafruit-hat-pwm configuration gets closer, but its still wrong:

[hdm]

Testing the same hw/code/pi image on a 3B+ has a similar green bar in adafruit-hat-pwm configuration and is completely broken (but in a different way) with the regular gpio mapping. Maybe something changed with the RGB panels?

[hdm]

It may be worth noting that the 64x64 has an E address and it looks like this library doesn't have explicit support for that yet. Edit: E is supported, thanks!

[hdm]

Thank you for the reply! It looks like this may be a known (if not common) defect in the RGB bonnet boards. Reading through the Adafruit support forums I found someone with a similar issue:

• https://forums.adafruit.com/viewtopic.php?f=47&t=170770&p=838423

I posted a similar entry at:

• https://forums.adafruit.com/viewtopic.php?f=47&t=171713

The "green lines" issue from the original poster matches the behavior I see if the gpio pinout is set to "regular" instead of "adafruit-hat" (or the pwm variant). My guess is that if they switch to --led-gpio-mapping=adafruit-hat they will see a similar issue to my photo above.

[hdm]

Thanks for the reply! I went down the list with a multimeter after booting the Pi 4, but with no demo (or other LED program) running:

A/B/C/D/E: All zero as expected R1: 3.14 R2: 0 G1: 0 G2: 3.99 B1: 0 B2: 0

It sounds like G2 should not be almost 4v while idle.

Edit: These tests were from the top of the bonnet

[hdm]

Thanks again for the reply. I wasn't sure which pins on the Pi mapped to G2 and found something interesting while trying to determine this. It appears that the G2 pin is shorted to quite a few other pins on the board, including the 5V line. A continuity test on G2 hits on about 6 pins. I tested w/o the Pi connected and still see a green line and up to 5V high on the G2 pin (before it goes to the chip).

Broken PCB?

[hdm]

Adafruit sent me a new bonnet and this resolved the issue.