For usage with the smart-leds crate.
An embedded-hal driver for ws2812 leds using spi as the timing provider.
It provides three variants:
The normal usage
Your spi peripheral has to run betwee 2MHz and 3.8MHz & the SPI data is created on-the-fly. This means that your core has to be reasonably fast (48 MHz should suffice).
Prerendered
If your core is too slow (for example, the AVR family), you may want to use this. It creates all the data beforehand & then sends it. This means that you have to provide a data array that's large enough for all the spi data.
Hosted
Intended for device like the Raspberry Pi or other Linux SBCs. Similarly to prerendered it creates all the data beforehand, but sends it using a single call.
Do you use the normal variant? Does your spi run at the right frequency?
Your CPU might be too slow, but this can also depend on the HAL implementation
& your Iterator chain. Using the prerendered
variant might help. For many
SPI peripherals, the clock generations is way less sophisticated than e.g.
the UART peripheral. You should verify it runs at an acceptable frequency, by
either studying the datasheet & the hal code or using a logic analyzer. An
fx2 based one, commonly available under $10 works great for this.
If the first led is always on, no matter what data you put in, your spi is probably not setting the mosi line to low on idle (You can check with a multimeter). It may also be a timing issue with the first bit being sent, this is the case on the stm32f030 with 2MHz.
You could try using the mosi_idle_high
feature, it might help.
If this does not help you can try different embedded_hal::spi::Mode
parameters. For example on
stm32f411 (as used for example on the Black Pill board) you have to set phase
to
Phase::CaptureOnSecondTransition
.
Is your device fast enough? Is your iterator fast enough? Taking too long may completely screw up the timings for the normal version. Try the prerendered variant.
Is everything white? This may stem from an spi peripheral that's too slow or one that takes too much time in-between bytes.
Is your first LED the wrong brightness/color while the rest of your LED's do what you expect? This is due to low voltage of data line, or too much of a voltage difference between Vin and Din voltages, making some "high" bits read as "low" bits to the chip. Due to the chips circuitry, these voltages are regulated as they are passed on to the next LED in the line, which is why the other LED's perform as expected. (For more on exactly what is going on here, see Hackaday | Cheating at 5V WS2812 Control to Use 3.3V Data)
Are you using the --release
compiler flag?
The timing of each byte passed over SPI is very sensitive, and running code compiled
without full optimizations can throw off your timing. Always use either --release
flag with your cargo <command>
, or alternatively set [profile.dev] opt-level = "3"
To ensure timing matches what your programmed. A dead giveaway of this is when all
pixels go full brightness for every color.
Does it mostly work, but sometimes some of the LEDs get randomly colored?
As the sending process needs to be without gaps, interrupts and other parallel tasks might lead to intermittent issues. Either disable interrupts during the sending process or ensure that the internal buffer of your HAL (if it exists) is large enough.
When opening an issue about wrong/strange data, it would help if you include your code (of course) and a capture of MOSI & SCK from an oscilloscope/a logic analyzer.
Licensed under either of
at your option.
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.