Hi JP!
Thanks to you and other members who have published posts about it, I have also started my project with the ESP
It consists in a 16*16 APA102 square panel controlled through SPI by an ESP-07 (very similar to ESP-12) using Arduino for ESP8266 (in fact at this moment is only 16*10, pending to connect the rest)
I made a small B4A drawing app in which you can draw directly the pixels and change its color and global brightness. Each time you make a change, it is sent using UDP to the ESP-07. It works fine!
About the HSPI speed:
It "seems" that it can be programmed to run much faster, at even 40MHz (if the hardware you are addressing allows that ) NOT TESTED
Also, you can transmit some buffers (very limited in size) NOT TESTED. So, you can split the big buffer into multiple chunks and send one when the previous has just been sent.
The drawback is that, at least with the arduino IDE, (perhaps I am wrong) there are not SPI interrupts, This would be ideal for the above point. Anyway, it can still be done by polling the SPI state in the main loop.
So, next step will be to measure the max real SPI speed that can be achieved.
RGBLed considerations (if someone wants to do imaging)
I made the "prototype" panel cutting 16 pieces, each one with 16-led, from a 5m 60led/meter APA102C. Then I put these pieces one under the other, so that the result is a square 16*16 panel.
The scan is from left to right for the odd rows and from right to left for the even ones because this is the way the rows are connected between them. So, the B4A app makes this pre-processing before sending the frame to the ESP-07
Color: The LED colours are far from "neat". Red is near orange, and less bright than the others. So, a little processing has to be done with the image to improve it
Pitch: with a 60leds/m strip, the distance between leds is 1,67cm. I have seen some panels with 10mm pitch, which, given the led size, seems better for images. But also, a higher density means higher cost.
Hope to be able to post a video when it is finished!
