Go Down

Topic: ws2812b 5V power up rate and noise affecting the leds with solutions (Read 252 times) previous topic - next topic

gy5000

I have a simple strip of 120 ws2812b leds connected to an arduino nano powered by a separate 5V supply.The supply is rock solid on 5V. I am using the fastled library.
I am using a MOSFET switch controlled by the nano to power them up or down.
There is no data transmitted to the leds at power up so they are all effectively off.

I have had a problem with random flashes and strange behaviour on power up.

After trying many capacitors to help I started looking at the power up rate. I have an oscilloscope.
The MOSFET switches the supply almost instantaneously (micro seconds).
I started slowing it down using some capacitors on the MOSFET gate. As I did the strange behaviour reduced.
Now the leds are powered up over 5 millisecs and the leds do not flash or glitch. This might be overkill and a shorter time may work but 5 miliisecs looks good. Data can then be sent and the leds be controlled.

The second discovery is their susceptibility to EMI. I have a washing machine motor nearby I ended up changing the power/data lead to a shielded cable and again my problems were solved.

Comments welcome.

PaulRB

Quote
There is no data transmitted to the leds at power up so they are all effectively off.
No, sorry, as I understand it, they will be completely random! And if you deny them power with the MOSFET, they can't receive data, I suspect.

gy5000

I do not see that behaviour.

My experience is they normally power up blank unless the supply is applied with a high slew rate then the leds may well be random.
Secondly I found the leds may not respond to the data if powered rapidly. They appear to be in a fault state that only a power reset fixes.

My finding is to power the led with a slew rate over 5ms and then supply data to control the led.
The led will start blank and then be controllable.

Go Up