I know ‘flickering’ is a well published topic I’ve read most of them !
I have 100 fcob 12v 720 RGB LED strip with a ESP32 mini, I2C level shifter, 3amp 12v supply, and buck converter for the ESP32. I’ve set WLED to 50 leds and the leds flicker badly I’ve tried another 12v supply and powering the ESP32 from a separate PSU with earth’s 0v connected together,and still flickers I’ve tried without the level shift and things improve but not fixed. I’ve moved the location of the reel etc and still the same.
I wondered if connecting a small value capacitor from the data line to ground to quiet the ‘noise’ if any ?
I’ve changed to FPS to 10 and 5 still the same. I’ve disabled ‘Enable transitions’ that helped a lot but still flickers once in a while
@myozone how long is the cable from the portpin to the first light in your projekt
if its less then 1feet try without a levelshifter but put at least 30-60ohms Resistor in the line
@Jinx I did read that but thought the clock line on I2C could be ‘fast’ therefore these would be suitable.
@Magig_Wled The cable is about a foot, two JST leads long, and maybe twice that length with the level shifter inline.
I have added a 15,000 microfarad 25v capacitor across the 12v line which made a slight difference maybe. I’ll add a ~60 ohms resistor in series with data line and report back.
I’m just testing a 5m reel sat on the floor.
UPDATE:
I’ve stuck a 68 ohms resistor in series with the data line without a level shifter and it’s improved a bit with an odd bright white flash every 1-10 seconds. Moving the brightness up changes the solid colour which sounds to me like a power supply issue ? although the voltage appears stable at around 12.6 volt ish.
Are all LED strips this fussy ? I had in the same Ali order some 8x 5v led sticks for Tasmota which work fine with WLED and Tasmota.
I’m right in thinking the data line with 50 leds is about 50.4 kHz (8bit * 3RGB * 50leds * 42fps)
Further update:
After much messing with series resistors a 360 ohm seems to be the sweet spot, not to high not to low without flicker yet. I did try a linear (transformer) power supply 12v 25A that made no difference. the 12v PSU I have says 3A but I tried it to 5A with no ill effects. The buck (charger XL4015) converter I have seems quite good even regulating a wildly swinging input voltages 6~12v with another overload 12v PSU I have the output stayed at the set 5.00V - No level shifter.
I think a variable pot say 1K would good to find that sweet spot with each setup. I think it all depends on the capacitance / inductance of the data line in different setups and a filter circuit might do better.
Regardless of what resistor you keep throwing at the setup, you are still feeding the LEDs lower voltage data then what they are spec’ed for. Sure they may work, but any little changes will likely alter those results. Including changing wire type/size/length, location with a different noise floor environment. Even changing how those wires are positioned by other conductive things.
Having level shifted signal gives you far more overhead for acceptable signal quality over all factors.
I 100% agree, Proper TTL levels are crucial with high impedance inputs esspecally in high noise environments with switching power supplies and long unscreened, unbalanced data lines and high switching speeds. I have an electronics background so should know better
I may design yet another wled board with proper a level shifter etc.
Problem with those i2c shifters is they’re made for a slow open drain i2c bus. Most won’t even pass an 800 KHz signal, so you’ve basically got a break in your circuit as far as the data is concerned.
I did some testing on a few strips without a level shifter. With perfect signal and no noise, most strips start to glitch if the received square wave was 3.0v or less. If you’re sending 3.3v you’ve got a tiny margin for error and noise.
I put trimmer pots on the output of line drivers, but it’s less useful than you might think. Most lines can be driven by about the same value, and lines that need extreme values tend to have very marginal signal and are probably a poor choice.
