Neopixel compatible chain probably fried?

Hey
So I've been using two 50pcs chains like this one they are controllable via the standard Adafruit Neopixel lib for Arduino.
They've been working reliably the whole time but recently I wanted to rewire them. Stupid me did so while everything was running.
My quetion now is: Is it possible to fry them by shorting the data line with ground? Because that's what I suspect happened.
I am powering them directly through a beefy 5v power supply. Data and Power delivery share ground. I have safety resistors between the data lines and the arduino and a capacitor to flat out spikes in power draw.
Now after the rewiring one of the chains doesn't react to data input anymore. I tried both data lines with the other chain and it accepts them both. I measure 5v across the 5v+ line and gnd on both chains. data is also definitely put thtough. The LEDS are still working I know that cause they all flash briefly when power is connected. I suspect that one or more chips might be fried.
To make sure I removed isolation from the data line between the first and second LED of the broken chain and connected data there with no success. But while I'm writing this I'm thinking should I maybe also have skipped the ground line of the first one? I was under the impression only the data line is sequential and everything else is parallel.

I hope I was able to describe everything with enough detail. Of course no one can know what's wrong from afar but maybe you guys have some suggestions as to what I could test from this point.

Stupid me did so while everything was running.

Well I did something similar but not quite the same. I had the LED strip on three pins of a header and I connected them up to three header sockets, at the same time in the right way round while pulses were being produced. Dispite having a resistor in the data line I managed to blow some of the LEDs in the middle of the strip. Only a short one of 16 LEDs. But I had to remove those LEDs and replace them with good ones.

So it is not necessarily a problem with the wiring being wrong, just a short time of not having the ground connected before the signal can blow them. In other words they are not hot pluggable, due to the type of socket I had. Or maybe the damage occurred when the strip was disconnected.

In my defense I was working on a Raspberry Pi where shutting down, powering down, plugging in a socket and powering up again and getting into the python development system can take about two minutes.

The LEDS are still working I know that cause they all flash briefly when power is connected.

Well that doesn't mean much, it tells you about the controller is broken, which is part of the WS2812 chip, so no mending that. At least those LEDs are easier to replace than the surface mount strips I was dealing with.

knorxo:
My quetion now is: Is it possible to fry them by shorting the data line with ground? Because that’s what I suspect happened.

No,but you may fry the controller.

Thanks guys I was quite certain it was the controller I broke.
My question now is: Could it be any controller in the chain that's broken now or will it always be the first one connected.
And also. IF I want to test whether the controllers behind the broken one work, do I have to connect ground BEHIND that controller as well or can I assume the ground line to be shared?

The "controller" to which HKJ-lygte refers is the mystery Arduino which you are using to drive the LEDs. :grinning:

However you are suggesting that it is appropriately driving some working chains. You mention "the data lines", these LEDs have two control lines (making them more convenient to drive as they are not timing sensitive) so if you are only driving one chain at any one time (because you could choose to have alternate output pairs for different chains) then if those lines successfully drive a working chain, that proves the controller to be working.

Given that you do not confuse the wiring, an interesting test would be to connect both a working chain and a non-working chain in parallel to your data wires so both see the same data. Unless you ahve long intervening cables, the Arduino should be perfectly capable of driving both, so if neither works that indicates a short circuit but if the working one still works and the non-working one still does not, then you do need to chase up the fault there.

For the non-working chain, I would advise simply cutting off the first ten LEDs and re-testing.

The controller I was talking about was the controller inside the LED.

Could it be any controller in the chain that's broken now

Yes, the chain would work up to the broken one.
There are two types of fault you could have.

  1. the controller not receiving or acting correctly on the data.
  2. the controller working and lighting the led but not passing on the data to the next one.

So the fault led could be the first led that does not light up, or the last one that does.

these LEDs have two control lines

You sure about that Paul? From the link I can only see four wires connecting the LEDs together, which would be power, ground, data in and data out.

Grumpy_Mike:
You sure about that Paul? From the link I can only see four wires connecting the LEDs together, which would be power, ground, data in and data out.

Yes, they have four wires connecting the LEDs together, from one module to the next.

Muse on that for a moment. :grinning:

I have had a chain of these hanging over the curtain rail here for some six years or so. Sometimes I connect them to the old Pro Mini and play with them. :roll_eyes:

Those could of course be WS2815s, but they look like the older data and clock ones.

Hummm, so if there are several chains, and one of the chains is broken, and when the broken chain is swapped with a working chain, what happens? Did the problem move or did the problem stay in the same place?

We wait with bated breath! :grinning:

There are two kinds of addressable LEDs that have two data lines.

The APA102, with a data and clock, and the WS2813, which has 2 redundant data lines, meant to prevent the sort of failures you experienced - one of them is a backup “passthrough” that will bypass a damaged LED. Clearly, if it is the latter, then it didn’t do it’s job…