Driving RGB LEDs with nRF52

[andrey]

In my case LEDs (N<10) would be animated (think pulsing).

My only concern is that the BLE core will limit the possibility to control the LEDs in real time, thus causing flickering and other artifacts. Perhaps DMA could help with that?

Has anyone had any experience with that?

[Ole Bauck]

Without knowing exactly what you aim to do I would recommend using the PWM peripheral on nRF52. This allows you to play sequences with different duty cycle. The lowest period you can get is 32767/125KHz ~= 0.26s. But this can be extended if you pad the sequence with some zeros (or ones).

I can also recommend to use programmable LEDs, like the WS2812B. These only require power, ground and signal, so no additional transistor needed. You can search the forum for code to control these (uses software to toggle gpio) or take a look at this to use the PWM peripheral for controlling the LEDs (this project is not yet finished, but I have tested it).

[andrey]

WS2812B work fine with this port even with softdevice running. Glitch-free in connection mode with the loosest timing that iOS allows:

#define MIN_CONN_INTERVAL                MSEC_TO_UNITS(180, UNIT_1_25_MS)
#define MAX_CONN_INTERVAL                MSEC_TO_UNITS(200, UNIT_1_25_MS)
#define SLAVE_LATENCY                    4
#define CONN_SUP_TIMEOUT                 MSEC_TO_UNITS(5000, UNIT_10_MS) 

[Ole Bauck]

Nice, I have not seen this before.

I guess it uses the same principle as the one I linked to which uses PWM, but I don't see how using SPIM is better, either the timing will be incorrect or it will use much more RAM. I have not looked at exactly how it works, this is based on the description. Also it says: "It may not work with all WS2812B based array, because current version is not meet with the official spec". Do you know what he means about this? Is it that the timing is not correct?

[andrey]

I'm not sure. And I have to qualify my previous statement -- I'm using a single LED, not an array.