Your source for industrial grade WS281X RGB LED pixel driving equipment

Made in USA driver boards for use with Falcon Player software on BeagleBone systems

Electrical Design Information and Advice for WS2811 RGB Pixel Systems

The high speed data system used by RGB pixels requires careful engineering considerations when implementing a large installation

In terms of running standard wire for distances of 10 to 40 feet, the WS2811 type RGB LED data signal is EXTREMELY high frequency. It's a wonder that it can work at all. We have gone to great lengths in designing our pixel driver board output circuitry to be as adaptable and robust as possible through immense R&D backed up by vast industrial electrical engineering experience. The WS2811 type data signal is well within the definition of "Radio Frequency". We have found that not too many of the people working with pixels have an RF engineering background so we've put together this page of electrical design information and advice to help get the best and most reliable performance out of your installation. If these procedures are not considered, it will be luck or a miracle if your system works.

Ground path, Ground path, Ground path

The ground path for LED pixels must be designed for both DC performance as well as AC performance. Any length of wire looks like an electrical spring to the data signal, the longer the wire the softer the spring. The data input of the pixel is referenced to it's ground lead. Digital inputs usally work in terms of two "threshold voltages". Below the "low" threshold, the input is considered a "0". Above the "high" threshold and it's considered a "1". The typical "TTL" threshold levels are; anything below 0.8v is a "0" and anything above 2.0v is a "1". The integrated circuit in the pixel is watching this input pin for a certain timing of "0" vs "1" to decode the color data into its memory to light up the LEDs. If it doesn't read timings that its expecting, it will not light up.

So, the data input of the pixel needs to have the "high" and "low" of the data signal (in reference to its ground pin) within a certain window for it to be able to pick up on the timing of the data and begin working. Two main things can be wrong with the ground path (assuming the data wire is fine).
A DC problem with the ground path is caused by the pixels drawing DC power to light up the LEDs. A DC current draw through any wire will create a voltage drop. If too many pixels are on too small of a wire, the voltage drop will become significant enough to skew the data input voltage levels from the pixel's point of view. Remember that the pixels are drawing current between power and ground, so voltage drop on the ground wire is actually "upward". This shifts the apparent data input voltage of the pixel down. Many pixel users are familiar with measuring voltage drop to determine whether larger wire needs to be used or when to implement "power injection". Most are thinking in terms of the supply voltage to the pixels and not about the effect that the "upward drop" on the ground is having on the data input. If in question, you should always use a multimeter to check voltage between a ground on the controller and the ground pin of the first pixel of a string to see how much the ground is "sagging up".
An AC problem with the ground path is caused by the current return path of the high frequency data signal. There are two components of the current on the data output of the controller. (1) Capacitive reactance of the pixel's data input which is referenced against the ground pin on the pixel. and (2) Capacitive reactance of the data wire going to the pixel which is referenced against anything conductive near the wire. This can be mostly directed toward the ground wire running along near the data wire, but also to the earth or anything metal that the wire is running near. The result is that a portion of the return current of the data signal is being drawn across the ground wire like an AC spring. This will result in major issues with the pixel receiving data and can be very hard to measure. The best way to avoid this is by strictly following an ideal grounding topology so that any "ground noise" (meaning fluctuations on the ground wire from data signals) are minimized and predictable. The more channels that are being run on one controller, the more this begins to be a problem. Strictly number of channels and length of wire to the first pixel matters, the amount of lights per string does not affect this.

Star point grounding.

Star point grounding is the accepted topology for having a predictable ground reference at each device in a circuit. The best way to accomplish this is by using a heavy bus system for joining all grounds in one place. The bus system can be constucted with DIN rail terminal blocks such as "Phoenix" blocks. Ideally, the controller PCB is mounted in an electrical enclosure with these terminal blocks. We have found that mounting the controller to the hinged lid of the enclosure and building the array of terminal blocks on the back plate is the most robust way to arrange the parts. On this ground bus, the power supply grounds, pixel grounds, and controller grounds all come back equally to one point.

If used, power injection must be implemented carefully to maintain the star point grounding system. All ground wires that come from the controller and power supply bus bar must rejoin at the first pixel of the string (acting like a single larger wire) before going into the first pixel and daisy chaining to the power injection point. This is extremely critical!

Differential Signal Systems for WS2811 RGB LED Data

Using differential signal pairs for pixel data is the ideal solution when the pixels must be greater than 20 feet from the controller. We have maximized our "standard" outputs to work on wire up to 40-60 feet but it is never recommended to rely on pushing this to the limit if reliability is desired. At that distance, it's strictly a laws of physics problem.
Differential signal pairs allow pixel data to easily be sent 1000 feet on standard wire. Using twisted pair wire makes distance a complete non-issue, though we recommend it for runs longer than 100 feet when possible. Obviously there is no reason to be running pixel data 1000 feet anyway. There are plenty of explanations about why differential signal pairs solve almost all of the problems of sending high frequency signals down wire, so we won't get too deep into that here, but if you think back to the explanations of why AC and DC problems with ground cause pixels to not be able to receive data properly it becomes apparent how determining the "0" or "1" of the data based on comparing the voltages of the differential pair and not ground can eliminate virtually all pixel data problems. The drawback to using differential pairs for pixel data is that you need a differential receiver that converts the signal back to standard pixel data located at the first pixel in the string, or close to it.