Someone asked about the Aliexpress panel and adapter he was trying with a Pi Zero.
(I got a couple for $1.56 ea in a monthly Aliexpress sale)
A most important 7 year old citation from the 3 string passive adapter:
if the LED panels have 74HCT245 (as opposed to just 74HC245) in their input stage, then they can deal properly with the 3.3V logic levels. ( I suspect the majority of panels require active interface boards to function).
After scouring the internet, I did find a hazy schematic diagram. After rechecking the wiring it does indeed appear to match the rpi-rgb-led-matrix connection standard.
@MaxMatrix : Do you have a link to the amazon one? I see the aliexpress one is https://www.aliexpress.us/item/3256806177226619.html?gatewayAdapt=glo2usa4itemAdapt
Has anyone tried it? Ultimately I usually need a 3 channel board, but having a mini board that works with a mini rPi is my new goal.
Has anyone made a solution that works with a compute board and is as flat/small as possible (compared to my current rpi3 with electrodragon 3 plugged into it)?
I would love to contact the maker to see if they’d be willing/able to make a 3 channel version. I’m also going to reach out to electrodragon to ask them the same.
I am using this exact board with a 64x64 matrix display… Problem is, it just displays 2 bars of multiple pixels width that are flickering in red and green. Is that really because the logic of the raspbi is at 3.3v and I dont have an active adapter? Shouldnt it show nothing then when I run one of the test samples from the project? Any help would be appreciated.
I tried the voltage at 4.5V and 3V for the panel and while it surprisingly still displays that weird flickering bar thing it does just that. I just cut my losses and bought a 20€ active adafruit adapter card. Oh well.
Thanks sharing your experience. I also have one on the way from China, I will try it with my panels, but indeed passive may not work well with some or maybe all panels
But to be clear, the adafruit adapter only has one port, whereas this one has 2. It’s not about saving money but I’d love a 2 or 3 port adapter that is that size.
I just received mine and I get the same bars than you. I tried all values of --led-gpio-mapping outside of regular, and none work.
The same rPi and panels work fine if I put the electrodragon board back in.
I don’t understand why it would be a level shifter problem if I lower the voltage to the panels to 3.3V (which I did and it didn’t help). @MaxMatrix do you have any idea about that? I didn’t verify the pin numbers like you did, but if you say they match --led-gpio-mapping=regular, we should get 2 channels without issues and without problems related to level shifters
It was very much about saving money for me Interestingly, you got bars in a different color than mine? Are they flickering or static? Mine were obnoxiously flickering.
I can also only plug it in one way around because the USB-ports on my raspbi block the orientation, I was wondering what would happen if I just plugged it in the other way round.
Err, rPi expansion boards only go in one direction. If you reverse them you can damage both the board (not here because the board has no components or power unless you added power), and the rPi (if the board has power and you plug it backwards, it will send 5V to non 5V tolerant GPIOs and break them).
Hopefully you put the board the same was as I did in my picture.
Yes, the bars flashed a little bit.
My 2nd panel was a 64x32, so ABCD only and it failed the same way. That said when the E line is working the pattern is very easy to recognize. I’ve had to solder the 8-E jumper on the ED3 boards many times and it doesn’t look like that when that jumper is missing.
Here, honestly it just looks like garbage IO being sent, like the wrong pins being mapped
Well, you may have been on to something. I did some basic checks on the board, and with the board barrel plug pointing up the 2 pins closest to HUB75 (pins 1 and 3) are shorted together and connected to the middle pin of the barrel connector. (the 2 I point to with my multimeter)
Now if we look at the rPi pinout the top 2 right pins are both 5V (pins 2 & 4), but on that board, it is the top 2 left pins that are connected together and to the barrel plug, and I assume that plug is meant to power both the rPi and the panels.
So if I get this right, my board is routed wrong and the pins on left and right may be reversed (they are bringing what I think is 5V to the top left 2 pins when it’s the top right 2 pins).
If that’s correct, that would explain why the rest doesn’t work either.
I did another test to be sure: LED 1 ping 2 is blue1 goes to 26, 26 on rPi is GPIO7 which is pin 26 on the right, I traced it and it went to pin 25 instead of 26 (same line, but left vs right).
So it looks like this board was meant for an rPi where the header is connected under it, which in turn will swap all the left-right pin orders
And guess what? I was right. I had a spare rPi without headers, I put through pin headers to get pins on the other side (didn’t solder) and tried again.
Now things work, except address E is missing, so you only get ABCD, not ABCDE which means 64x32 and not 128x64 (although with a couple of patch wires, that could be fixed I think)
Note how the adapter is under the rPi. You cannot do this on any rPI unless you solder a connector on the wrong side:
@hechaogm just for completeness, I made a test with P1.875 ABC panels ( --led-row-addr-type=3 ) and they worked fine.
So the good news is that all my panels work fine with 5V power and 3.3V signalling from rPi. The only issue was this board was designed and shipped without even being tested,
I didn’t solder the rPi plug on it but I think if they had soldered it on the other side, the board would work if it’s plugged away from the rPI with the HUB75 connectors pointed down but away from the board.
I wonder if I can message the seller to just ask if they want to send me a fixed version, It cant hurt to try and technically the product is not functioning properly