OctoWS2811 signal too low

kcwarner

Member
Hello,

We've soldered together our Teensy 3.2 board with the OctoWS2811 but weren't seeing any of the lights working when just using the Sync (+ GND, 12V power supply) line on the side. Using an oscilloscope, we see that the data output is only 1.68V max, not the expected ~5V. Any idea why this is happening or what we might be doing wrong?

Thanks in advance!
Katie
 
A photo of the setup would be helpful here.

Going off the words, you shouldn't have 12V being applied, the board is 5V max and the Teensy is most likely dead. Second question would be using sync, since that is intended to be a teensy to Teensy connection. The LEDs should be on the 8 channel outputs.
 
Yikes, 12 volts!

Be really careful if you've got 12 volts anywhere.

Touching 12 volts to any pin on Teensy, even for just a brief instant, will always kill it. There's a possibility your LEDs and maybe even your computer could be damaged, but those are less likely.

If 12V touched that sync pin, I'm sad to say your Teensy is almost certainly dead.
 
Last edited:
Sorry for the confusion. Our lights run off 12V so we connected the power and ground on them to 12V. Teensy & octo have always only been connected via USB to the computer. Blink still works and has always worked so the Teensy isn't dead! Does the octo board need an external power supply though for the signal to reach 5V? That was going to be our next step.
 
IMG_1712.jpg
There's our simple setup. The white just attaches to the computer and the blue is cut so we can connect the lights to each of the appropriate lines.
Is there a chance the board could be fried but blink would still work?
 
We've modified the blink program to run to different pins and our high output is either 331mV or 133mV according to the oscilloscope. Pins 2,5,6,7,8,14,20,21 are all 133mV and the rest are 331mV (including the LED pin 13 which turns on at this reading). We tried this running the Teensy off USB, 3.3V source, and 5V source and all values were read the same. Can anyone explain this or suggest something else we try to get them to all be the same or all be the expected 3.3 or 5V outputs (5V from the Octo board)? Thanks in advance!
 
Does the octo board need an external power supply though for the signal to reach 5V? That was going to be our next step.

Yes, you need needs 5V power. Running from only 3.3V won't work. Teensy can run with only 3.3V power, but the 74HCT245 buffer chip on the Octo board needs 5V.

You can use either 5V from the USB cable, or apply 5V to the VIN and GND pins. Two of the 3 locations on the terminal block connect to VIN and GND specifically to allow you to easily apply external 5V power.

If you haven't cut the VUSB-VIN pads apart on the bottom side of your Teensy, be careful not to use both power sources at the same time. With that link still connected, the 5V power can flow back into your computer. Usually this doesn't harm most PCs, but it's certainly not a good practice.

Is there a chance the board could be fried but blink would still work?

That's unlikely. Usually 12V touching Teensy causes instant death. If you can still upload the LED blink and see the results on the orange LED, your hardware is probably still working.
 
In most instances, we've kept the Teensy connected to USB, we just didn't know if that would supply the required 5V. Thanks for your info!

Using a multimeter, we got our high output readings around 2V off the Teensy pins and 1.7V out of the Cat5e in the Octo. This seems to be too low to communicate with the WS2811 chip in the lights though. Are we missing something else?
 
Looks like your 5V power to the octo board isn't working. Confirm you really do have 5V at the screw terminal?

Also 2V is pretty low on the teensy, either something is drawing way too much current or the measurment wasn't quite right.

Then since it looks like you socketed the Teensy you can try unplugging all the power and then removing the teensy, then with 5V power being provided to the screw terminal:

Looking at
https://www.pjrc.com/store/schematic_octo28.gif
measure from from pin 20 to pin 10, or if that looks to exciting across the small brown capacitiro and confirm there is 5V.
Then using a wire jumper on pin 2 of the socket where the teensy was
Connect to gnd (first pin on teensy row), measure output B1, or socket 1 pin2. Should be 0V
Connect to to 5V (lost pin on the other teensy row), measure output B1 on pin 18/socket 1 pin 2, should be 5V

If this isn't working out then there may be something wrong with the octoboard.
 
Thank you so much for your suggestions and advice. We really appreciate it!!

We can confirm 5V at the screw terminal. With 5V to the screw terminal, we're getting 0V from pin 2 to gnd but only 3.5V from 5V to pin 18/socket 1 pin 2.

With 5V at the screw terminal, we're only reading 1.16V out of the RJ-45 ports for high. We've ordered another octoboard though just in case.
 
For testing voltages, remember the WS2811 protocol is very fast waveforms. A DC multimeter will show you the average voltage, not usually a useful measurement. If you have an oscilloscope, you could use it to see the actual waveform. Or without a scope, perhaps program the LED blink and slow the blink time to a few seconds (some multimeters respond faster than others) and change the pin number to whatever pin you're testing. Perhaps disconnect the LEDs if you want to be really careful, but they're usually pretty good about ignoring slow waveforms.
 
Thanks for your suggestions. We slowed down blink and still aren't getting signals like we expect. Could the 74HCT245 buffer be damaged? Vin is 5V and all the inputs are around 3V but the highest output we are reading is 1.6V. This should be 5, correct?
 
That's what should be happening with a slow blink, so something is amiss there, nothing on the outputs is shorting the buffer pins to gnd is it? Or come to that, what happens if you pull them high with a 5k resistor to 5V? wondering if the buffer direction/enable has failed in some way and that's a floating input. Still probably means you need a new board.

If the board is dead it most likely happened during assembly so you do want to look at how this happened to avoid a repeat. Still not quite clear on the 12V.

Is this that you are using the 12V strips where the data pin is still 5V, or is this something else?
 
We are using the 12V strips to power the lights so grounds are common, data is connected to the output of the octoboard, and Vin for the lights is 12V. We've tried plugging Teensy/Octo in via USB and supplying 5V at the screw terminal (never at the same time) and get the same results. It's been suggested these are floating inputs and have ordered a new board but we will be extremely careful about testing and assembling this time.

We'll try pulling the buffer pins high. Thanks so much for your suggestions!
 
This should be 5, correct?

Yes, correct. When Teensy drives the pin with 3.3V, the 74HC245 buffer is supposed to output 5V on the corresponding signal. Only 1.16V is definitely wrong.

We can't see how you're testing, so this is a bit of blind guesswork. But one thing to check is cable and whatever might be connected on the other end. It's possible the 74HC245 is trying to output 5V, but the cable is shorted or has something on the other end that's acting like a heavy load which the poor little 74HC245 can't successfully drive. Of course, if you have no cable plugged in at all and you're measuring right at the RJ45 pins or the 100 ohm resistors on the PCB, then this doesn't apply. A mis-wired RJ45 where all the white wires are where the colored ones are supposed to be and vise-versa could also cause troubles, if the other end it wired. Did I mention how in the dark we are about *how* you're actually doing these tests & measurements?
 
We have nothing connected as far as cables from the Octo board. We've tested more often with nothing in the RJ45 port and just at the 100 ohm resistors, but a few times measuring off the white and colored wires stripped from within the cat5e cable. We've made sure not to switch the RJ45 wires and paid close attention to ground and signal wiring. I've added a picture with the multimeter and cord for the oscilloscope that we've been using. We touch the ground pin/clamp to a ground port on the Teensy or Octo then touch the other pin to the appropriate port or resistor as desired and read the meter or o-scope.IMG_1727.jpg
 
Don't mean to dig up an old thread but I think I am having a similar issue with some new LEDs I am trying. I got these LEDS and I have wired them such that 12V power and GND are connected to the LEDs, the orange wire of a cat6 cable is connected to the data wire of the LEDs, the orange white wire is connected to GND as well (at the LEDs). I am powering the Teensy 3.2 via USB on my computer and I am getting 5V between the 5V and GND pins on the octo adapter. Since right now I only have access to a volt meter when connected to the data lines and GND I get a fluctuating and very small voltage (I'm assuming this is to be expected since I'm getting an average). I am getting 12V (12.5 actually) at the LEDs and along the strip. However I can only get the first 3-6 (i.e. 1-2 groups) to turn on and they are not controllable. They just sometimes pop on random colors (normally blue or green) when I plug and unplug the cat6 from the RJ45 jack.

Anyone have any ideas? Did I wire this wrong for the 12V LEDs? Is there a test someone can suggest to try and narrow down the issue?
 
If you just have a multimeter you can mod blink to drive the octo output you are looking at and bump the delay to 5 seconds and get a very slow square wave that is readable, test with and without the LEDs connected.

You can also try the adafruit neopixel library to fire up just one strand at a time to see both that the strips work, and that your wiring is correct.
 
If at any moment, even for the briefest instant, you happen to touch the data wire to the +12V power, the octo board will very likely be destroyed.
 
Back
Top