Teensy 3.0 OctoWS2811 Timing issue

Status
Not open for further replies.

jlewis184

Member
Hello,

What can effect the output timing of the OctoWS2811 library with the Teensy3.0?

I'm using the BasicTest example code from the OctoWS2811 library and I'm getting strange results. Only 4 of the 8 strips I've tested are lighting up.

I'm driving 150 pixels, out of pin 2, through a line-driver powered at 5V.

The timing looks to be out of spec. As measured T1H~848nS and T1L~400nS. T0H~280nS and T0L~972nS. The period is right on at 1.25uS though.

Shown below is pin2 of the Teensy on channel 1 and Vdd on channel 2.
 

Attachments

  • photo (4).jpg
    photo (4).jpg
    146.4 KB · Views: 111
The timing doesn't precisely match the spec sheet, but it does match the general behavior of these LEDs. The 0 bit needs to be at least 280ns, and the 1 bit at 750ns. What's surprising is how small the delay between bits can be (as low as 1ns) if the signal is nice and clean. What's more, if you test the signal at the 2nd LED, you'll find the timings are completely altered (each LED seems to generate its own signal for the next). You can also go way beyond 1.25us per bit without issues (as long as you go low before 1.25us, you can wait up to 9000us before the next bit w/o the latch kicking in).

Anyway, I dont think its the timing that's causing your non-functional strips. Got any photos of the wiring? What's the signal at a working strip look like compared to a non-working one? How are things grounded?.
 
If you really want to mess with the timing, there are four #define lines in the code you can edit. But I would not recommend it.

Many people have experienced "strange" problems with these LEDs, especially flickering that seems to happen on only some parts of a display, sometimes triggered only with certain colors or other seemingly strange circumstances. Time and time again, people have gone down this fruitless path of investigating and fiddling with the waveform timing. There are plenty of inconsistent specs out there, even timing in some datasheets that are obviously wrong and don't even add up to 1.25 us, which only confuse the issue and further distract from solving the real problem.

That real problem, almost always, is inadequate wiring, especially on the power lines. The fact that the strips themselves aren't very good conductors is a huge factor. Even just 150 in one long strip often needs strong 5V power supplied to both ends.

It's best to locate your 5V supply close to the LEDs and run large, short wires directly from the supply to the LEDs. Then run signal+ground from the Teensy to the LEDs. The grounds should meet at or near the LEDs.

Don't underestimate the power requirements. 150 LEDs turned on fully (white) needs over 7 amps of current.
 
Well, I tidied up the wiring to snap a pic and sure enough everything worked.

Sorry to waste your time, but thanks for the help.
 
Status
Not open for further replies.
Back
Top