I’m new to FastLED this year, and have a setup with an ESP8266 driving three strings (60, 100, 100 pixels) of WS2811s on three pins. One of the strings is a bit further away from the chip (17-18’) and is showing symptoms of signal latency (wrong colors, strobing during refresh). I’m assessing my options.
drill a (small) hole in the garage wall to create a shorter route for the wire.
add a null pixel
?are there any software options to tune FastLED to improve my chances, especially since my total pixel count is so low? I’m not doing any fast animations – the display is mostly static with occasional color changes.
What are you using for the longer run? For long runs I tend to use something with twisted cables - like cat-5 – and put data/ground for the leds on a twisted pair.
I’m just using a single conductor, relatively light gauge, alongside 16ga landscape wire for power. So you’re suggesting a twisted pair with the ground connected to the power - on both ends (currently, the GND pin on the ESP is connected to - at the power supply)? A little easier to connect at the PS end right now, as the first LED is about 9 feet off the ground and the current connections are soldered with heat shrink.
I was able to slip the data line through some weather stripping and knock about six feet off the run and it’s still flaky. Heading to the the basement to look for some UTP now…
@Daniel_Garcia how long are you able to go with cat-5 / twisted pair, before you need to insert repeater LEDs? I’ve done some limited tests using 18/4 cable and we see to need repeaters every 10-14 feet. It isn’t twisted pair, and it’s running 5v/GND on the other two wires (I’m using WS2813’s so there are 2 redundant data signals).
With cat-5 cable and making sure each twisted pair gets one ground and one data line I’ve been able to get over 50’ runs (for four strips of leds, even…) - the downside is power has to run separately
I swapped out my single strand for a length of twisted pair, and have the blue & blue/white pair connected to data and ground (and my heavier gauge landscape wire carrying + & - power) and I’m still seeing the same behavior. I swapped the signal pins at the ESP and the problem stays with that string. The good news is that it lands on a combination of red and green about half the time, so it’s not too ugly.
One other possible issue - the power and signal for this string are running alongside the first 18 pixels on one of the other strings; I’m going to disconnect that string and see whether it may be picking up crosstalk from there.
It’s very consistent. It gives the same pattern every time. And the other to strings aren’t affected, it seems unlikely that it would be the wifi server.
try disabling wifi, or simply download any sketch from the library examples. if the problem remains, then the wifi is not to blame. there are several options for fixing these problems (flicker, strobe…)with ESP. One of them is install level shifter like 74AC245.
Problem verified with a demo reel. The pattern runs, but the colors are off. I’m pretty sure I bench tested this string with the 4 colors I’m trying to display, but I’m beginning to wonder if somehow it wants the colors in a different order.
@Tom_Gerhard I was seeing similar issues with my 3 strings of 2801’s. They run at 5V. I was seeing too much voltage drop. You may consider (if not already done) injecting power at each end of the three strands. I ran a 12GA wire alongside my entire 50’ length and all is good unless I try to run everything full brightness 90% leds lit up. If I didn’t run the extra power I was getting glitchy leds near end. Maybe that helps.
@Trey_Coursey thanks for the suggestion. I’m running 12V strings; I measured the voltage at the tail end of one of the strings and it was only about 0.1v drop at the brightness I’m running. I’ll double check this string to make sure it’s the same.
I’m getting convinced that I have the wrong chips in this string. I removed the color correction (SMD5050) and set the brightness to 255, and get very different results. I can address every pixel, but don’t get the expected results. If I send Red, I get Blue. If I send Blue, I get Green. If I send Green, … a very dim blue. If I set brightness to 128, I get Green, dim Red, and Red. White and Black behave as expected. If I change the color order, I get different, unexpected results (and not just the above rotated). I’m going to email the vendor I bought them from to see if there’s another possible chip they may have used. In any case, this is staying up through the season, so I’m going to experiment until I find colors that work and adjust my programming to deal with it. Luckily I’m only using 4 colors (Red, Orange, Blue, Green). I’m going to write some test patterns to see if I can find
