[BUG] Some weird problems, Arduino (SainSmart) crashes (3 Viewers)

[Dragy]

That is a good idea. Unfortunately this is not something I can test right now because I don't have a spare led.

But I am sure the Arduino IS crashing sometimes, because it executes the setup() again at this point.

 

[Lightning303]

But I am sure the Arduino IS crashing sometimes, because it executes the setup() again at this point.

I see. What about power consumption? Is your powersupply big enough? With 64 leds you would need a power supply that provides 3,84A. The arduino could cut out if there is no current to drive it left.

Or maybe the voltage is dropping to much sometimes (fast and bright changes) so the arduino cuts out and after a while reboots again. Do you have a capacitor between power and ground? This cap can make sure something like this does not happen. If this is the cause of the problem that is.

https://learn.sparkfun.com/tutorials/ws2812-breakout-hookup-guide

 

[Dragy]

Yes, the power supply is sufficient, 4A, but may handle even more because I think it is of good quality.

I now have a resistant on the data line, but it doesn't work. Also, I'm using another Arduino board now, the Arduino Nano w/ ATmega328. Still the same problem with this board, so I'm sure there's something wrong in the leds, software, drivers or power supply. I will test further, but if anyone here has a good idea, please tell because I like to know!

 

[Lightning303]

What about the capacitor i mentioned earlier?

 

[Dragy]

Yes, I was pretty sure (before) that wasn't going to solve the problem, because it doesn't crash with sudden changes in brightness when using Live view or something. Instead it crashes with small changes in using the sliders, not giving sudden power drops or requests.

I think the problem lies in the software, drivers or leds instead. I will test the separate led thing soon!

 

[HomeY]

Yes, I was pretty sure (before) that wasn't going to solve the problem

It's still highly recommended to put a 1000 uF capacitor on the +5V powerline, since the WS2812B are very sensitive LEDs.
http://www.tweaking4all.nl/hardware/arduino/arduino-ws2812-led/

I think i've burned about 10 of them while testing The good thing is that you only need to 'cut' of the first LED, since the rest is still good (if it happens)

 

[Dragy]

Okay, the capacitor did not solve the problem, it's still the same. I even encountered flickering of the leds at the beginning, but not anymore.

I am now using my Arduino Uno to debug. I have 1 WS2812B led on pin 13 and one "debug" led on pin 10. I am not using an external power supply in this setup, so I can rule out a lot of factors. The result is that the problem also occur with this setup. The leds get stuck when playing with the static color. What I can see is that the Arduino still gets serial code (RX led flashes) when playing with it, but the RGB led does not change. I am also testing this code:

#include <FastLED.h>

// Set the number of leds in the strip.
#define NUM_LEDS  64
#define DATA_PIN  13
#define BRIGHTNESS  255
#define LED_TYPE  WS2812B
#define COLOR_ORDER GRB
CRGB leds[NUM_LEDS];

void setup(){
  Serial.begin(115200);
  pinMode(10, OUTPUT);
  pinMode(11, OUTPUT);
  //sanity check delay - allows reprogramming if accidently blowing power w/leds
  delay(2000);
  FastLED.addLeds<LED_TYPE, DATA_PIN, COLOR_ORDER>(leds, NUM_LEDS);
  FastLED.setBrightness(BRIGHTNESS);

  setupLEDs();
  loopLeds();
  clearLeds();
  //blinkLeds();
}

int readByte(){
  while (Serial.available() == 0){ /* wait for data */ }
  return Serial.read();
}

void loopLeds(){
for(int j = 0; j < 3; j++) {
  for(int i = 0 ; i < NUM_LEDS; i++ ) {
  memset(leds, 0, NUM_LEDS * 3);
  switch(j) {
  case 0: leds[i].r = 255; break;
  case 1: leds[i].g = 255; break;
  case 2: leds[i].b = 255; break;
  }
  FastLED.show();
  delay(5);
  }
  }
}


void clearLeds(){
  for(int dot = 0; dot < NUM_LEDS; dot++){
  leds[dot] = CRGB::Black;
  };
  FastLED.show();
}

// shows colored LEDs for 500ms then blanks out ready to go
void setupLEDs()
{
  memset(leds,150, sizeof(leds));
  FastLED.show();
  FastLED.delay(500);
  memset(leds,0, sizeof(leds));
  FastLED.show();
}

void loop() {

  if(readByte() == 0xFF){
  if(readByte() == 0x00){
  if(readByte() == 0x00){
  int channels = readByte();
  for(int dot = 0; dot < channels; dot++){
  leds[dot].r = readByte();
  leds[dot].g = readByte();
  leds[dot].b = readByte();
  }
  FastLED.show();

  digitalWrite(10, HIGH);
  delay(300);
  digitalWrite(10, LOW);


  }
  }
  }
}

This works fine, the RGB led (WS2812B) gets data once in a while and after that it has to wait 300ms to be able to receive a new signal. The failure is less likely to occur, but still happens after a while. The Arduino still gets serial data, but as it seems, never the "right" data. The RGB led and the normal LED never change/flash anymore.


EDIT:
What?! I just placed the led code in front of the FastLED code and the problem does not happen anymore!
Like this:

void loop() {
   if(readByte() == 0xFF){
  if(readByte() == 0x00){
  if(readByte() == 0x00){
  
  digitalWrite(10, HIGH);
  delay(150);
  digitalWrite(10, LOW);

  int channels = readByte();
  for(int dot = 0; dot < channels; dot++){
  leds[dot].r = readByte();
  leds[dot].g = readByte();
  leds[dot].b = readByte();
  }
  FastLED.show();
    }
  }
  }
}

Maybe it's because of the delay?

Edit:
I tested it, the lowest I could get is 6ms delay. I can put it anywhere before the FastLED.show(), but the output is not okay. The leds don't lock up, but it's not the right solution because the output to the leds is not okay. I get different colors or flashes when playing with the static color sliders. Color changer mode is okay.

I have no idea what's going on.

 

[Lightning303]

You should not use delay as it halts the whole arduino for that time. In the meantime data comes in via the serial port and you cant use it as the arduino is just doing nothing. The buffers overfow and you loose all the infos.
So better to use millis().

http://arduino.cc/en/Tutorial/BlinkWithoutDelay

 

[Dragy]

Yeah I know, but adding the delay does prevent the leds from locking up as it seems.. But will try that method too of course.

 

[Dragy]

Well, I just ended up using Adafruit's NeoPixel library. It works fine! Feel free to use my AtmoDuino code! I think it's just some bug in FastLED or something.

#include <Adafruit_NeoPixel.h>

#define NUM_LEDS  64
#define DATA_PIN  3
#define COLOR_ORDER NEO_GRB
#define STRIP_RATE  NEO_KHZ800

Adafruit_NeoPixel strip = Adafruit_NeoPixel(NUM_LEDS, DATA_PIN, COLOR_ORDER + STRIP_RATE);

void setup(){
  Serial.begin(115200);

  strip.begin();
  strip.show(); // Initialize all pixels to 'off'
  loopLeds;
  strip.clear();
  strip.show();
}

int readByte(){
  while (Serial.available() == 0){ /* wait for data */ }
  return Serial.read();
}

void loopLeds(){
for(int j = 0; j < 3; j++) {
  for(int i = 0 ; i < NUM_LEDS; i++ ) {
  switch(j) {
  case 0: strip.setPixelColor(i, strip.Color(255, 0, 0)); break;
  case 1: strip.setPixelColor(i, strip.Color(0, 255, 0)); break;
  case 2: strip.setPixelColor(i, strip.Color(0, 0, 255)); break;
  }
  strip.show();
  delay(5);
  }
  }
}


void loop() {
  if(readByte() == 0xFF){
  if(readByte() == 0x00){
  if(readByte() == 0x00){
 
  int channels = readByte();
 
  for(int dot = 0; dot < channels; dot++){
  int r = readByte();
  int g = readByte();
  int b = readByte();
 
  strip.setPixelColor(dot, strip.Color(r, g, b));
  }
 
  strip.show();
  }
  }
  }
}