Problem with OctoWS2811 Library (Leds basically just white)

[mcsteeze]

Hi,
I have a 4 meter long strip of 60 leds/meter 2812 Leds. I tested the strip with the Adafruit Neopixel Library, a power supply and an Arduino Uno.
Everything worked the way it should so I know the LED strip and power supply work correctly.

I want to switch to using the OctoWS2811 library but I am having problems.
I am using the OctoWS2811 Adaptor and have wired / followed the directions here: https://www.pjrc.com/store/octo28_adaptor.html

However, for some reason when I try the BasicTest example, only about half the LEDs turn on, (sometimes only 8 turn on if I specify ledsPerStrip as the default 120 in the sketch) and when they turn on they are basically all white with a little bit of tweaking out.

Has anyone else experienced this problem?

The only parts I was not sure about in the instructions are the use of the phrase :
- "Top/Bottom RJ-45 Jack". The RJ-45 jacks are next to each other so I am not sure if this means left or right (when looking at the at the end of the board with RJ 45 jacks). Regardless it didn't make a difference. When plugged into the Left jack, nothing happens. In the Right jack at least there is some change in the leds.

- "The LED power supply ground and Teensy signal ground should meet at or near the LED strip signal inputs" - I am not sure exactly what this means either.

My set up is:
I have the V- from the Power supply going to the GND pin hole on the Adapter board.
I have the V+ going to the +5V input pad on the LED strip
I have the Orange wire from CAT6 connected to the DIN pad on the LED strip
I have the Orange and White wire from CAT 6 connected to the GND pad on the LED strip
I have the Teensy plugged in via USB to my computer

All wires are fairly short so the power supply and teensy and led strip are all close together.
Can anyone spot potential problems?

Here is a photo that kind of shows my issue.

(Just imagine there is no movement of any kind except some flickering)

 

[mcsteeze]

As an update, *not* using the adapter board works fine.

These are, as far as I can tell, the WS2812 LEDs (there are only four chips inside) but I am not sure what I would be doing wrong when using the adapter.

I think it might have something to do with the ground of the adapter / Teensy being hooked up wrong (wiggling the wire does change the LEDs a lot) but I am not sure. Does anyone have experience with this?
As helpful as the directions are, this picture does not explain very well what grounds need to be connected where. Can someone provide some insight into this?

Thanks

 

[GremlinWrangler]

Lots of ways for things to go astray with the WS LEDs since they rely on a self clocking interface and the PWM turns them into little transmitters. Most likely root cuase is noise drowning out the LED data, but that can have many sources.

The warning about ground connection is that you want to have big high current connectors feed the strips, and a separate line going off as the ground reference to the Octo board so the digitial data is correct with respect to ground, which sounds like what you have done

If you feed power at the octo board, and then head off down long leads the current loses cause a voltage drop on the ground line, so the LEDs see a different ground potential than Teensy driving them with potential data corruption.

Options for testing:
Try the Adafruit neopixel library, and point it at one of the octo pins at a time- removes a lot of things including current draw and makes things easier to spot.
Stick a multimeter prob to the power supply ground and check:
Strip start gnds,
Strip far Ends gnds
Teensy Board gnd
Octo board gnds

All should be less than 100mv

Then repeat checking with strips off (load blink)
Power supply 5V
Strip 5V at starts
strip 5V at far ends
Octo 5V
Teensy 5V (is this from USB or loca PSU?)
You should see 5.0v-4.9 volts
Then load up demo code (octo or neopixel) and repeat, watching for any drops or bouncing voltages.
If you have a scope the above steps are faster and it's easier to probe the data lines as well.

 

[PaulStoffregen]

Is it possible the orange and orange-white wires are swapped on your CAT5 cable? With the power turned off, try using a multimeter to measure ohms between the orange-white wire and GND on Teensy. It should be very close to zero ohms, since it's connected through the board to GND.