Really cool, but can you still drive it with a teensy 3.1?
Originally shared by Adafruit Industries
NEW PRODUCTS – 64×32 RGB LED Matrices – 6mm pitch / 5mm pitch / 3mm pitch
Bring a little bit of Times Square into your home with these sweet 64×32 square RGB LED matrix panels. These panels are normally used to make video walls, here in New York we see them on the sides of busses and bus stops, to display animations or short video clips. We thought they looked really cool so we picked up a few boxes of them from a factory.
These matrices have 2048 bright RGB LEDs arranged in a 64×32 grid on the front. On the back there are two IDC connectors (one input, one output: in theory you can chain these together) and 12 16-bit latches that allow you to drive the display with a 1:16 scan rate.
These displays are technically ‘chainable’ – connect one output to the next input – but our Arduino example code does not support this (yet).
These panels require 13 digital pins (6 bit data, 7 bit control) and a good 5V supply, up to 4A per panel. We suggest our 4A regulated 5V adapter and then connecting a 2.1mm jack. Please check out our tutorial for more details!
Much more information on the product pages!
All in stock and shipping now! #newproducts#leds#rgb#matrix#matrices
There are forks of the SmartMatrix Library that can refresh multiple panels chained together. I’m 99% sure these new Adafruit 32x64 panels behave like two 32x32 panels chained together.
I have one of these 32x64’s and it’s working fine with Teensy 3.1 + SmartMatrix and/or FastLED. I really didn’t need to do much to “fork” SmartMatrix 2.1, just change the MATRIX_WIDTH constant. I also dropped MATRIX_REFRESH_RATE a bit, too, based on a concern I’d read elsewhere. FastLED 3.0.3’s SmartMatrix demo reveals small bugs that are probably just another constant I’ll need to change. Text scrolls over 32px, the plasma doubles rather than going full width, but the “draw a circle” code works across the full width. I’ll check it on 3.1 next.
I’m glad it’s mostly working. Fixing text scrolling will take more than just changing a constant, but it might be fixed already on the pull request I linked to above.
Thanks. I had completely missed the link to the PR in your comment! But I stumbled on it myself, and yeah, now that I’ve read through it and #7, I see that things are more complicated than just the constant change. I understand why changes are necessary for the 1-bit foreground storage. I haven’t traced into FastLED’s SmartMatrix usage yet, so I don’t quite understand why the background is getting doubled. I’ll give the fork a try and report back!
Well, no luck - with PR17 installed it has the same issues.
My initial optimism came from testing with https://github.com/pup05/RainbowSmoke. It works great when modified for 32x64 (both my simple MATRIX_WIDTH change and PR17)
When I switched to the FastLED SmartMatrix demo, it mostly worked too so I posted here. Looks like I spoke too soon.
To make sure it’s nothing to do with FastLED, I went back and added some scrolling foreground text to RainbowSmoke, and scrolls across 32px while the background image draws correctly at 64px wide.
I guess this isn’t really a FastLED issue, so maybe I can switch forums But I’m happy to help debug this further… I’ll dig deeper into PR17 to see if I can help.
There was a much earlier pull request (and so based on a much earlier version of the SmartMatrix Library) that made scrolling text work across panels. You might be able to merge some of those changes into the latest version of the SmartMatrix Library.
Thanks for the lead. Yep, I just used that PR to fix the foreground buffer! It took a little updating, but it works for my 64x32 case now. I don’t think it’ll work for every matrix dimension, though. Only panels laid side-by-side.
I’m using https://github.com/mrwastl/SmartMatrix/ because it’s merged with your latest code (vs PR17 which is a few commits beind.) I’ll check with him to see if he’ll pull my changes and update your PR17.
@Steve_Eisner - what pitch 32x64 did you get? I’m troubleshooting an issue with the 3mm pitch, it doesn’t work with the logic levels of the Teensy 3.1. Potential solutions are putting a level shifting buffer between the Teensy 3.1 and the matrix, and lowering the voltage of the panel to lower the input high voltage of the chips.
But I’m happy to help debug this further… I’ll dig deeper into PR17 to see if I can help.