WS2811 Technicolor Strings Have High Failure Rate

[rick surette]

Hi Steve,
Does this also apply (lowering the voltage) to the non Technicolor Strings that have the TM1804 IC?
Thanks
Rick

[Steve Gase]

Hi Steve,
Does this also apply (lowering the voltage) to the non Technicolor Strings that have the TM1804 IC?
Thanks
Rick

My questions were wrt TC, and I think that is how ray replied.

[Hauvega]

The regular tm1804 smart strings are .3w per node.  If the technicolor tm1804 pixels are failing, something else is causing the failure.  I would like to see some of the failed 1804 techni nodes.  We are running the regular 1804 nodes at 12v.  What is the difference?

[Steve Gase]

Initially I received one 100ct TM1804 string that I plugged in and it failed at pixel #19.  I also received 2 50ct WS2811 strings that had no problems.  All 3 were part of a much larger order that didn't ship for another month...  I have no data on putting any of these strings outside -- they were only used for bench testing.


I could send the one bad pixel to you... but it might be better to see if I have additional failures after my 8.5v and 50% intensity is applied to the more recent strings.  the string that I received 2-3 months ago may tell us nothing.


i removed that pixel, btw, and the string worked fine after the fix.

[Hauvega]

I say put them outside and test again.  Wait for the weather to get bad.  I ordered some uv dye to see if there is any penetration into the node.  I took some pics using my usb microscope of some of my defective nodes.  Some pics of some of my pixels.  You are not allowed to view links. Register or Login

[RJ]

My testing of the ws2811 was not technicolor nodes and it was just a sample that ray sent me to test on a SSC. They worked fine. But they were never ptu in the weather.

RJ

[joshuashu]

I say put them outside and test again.  Wait for the weather to get bad.  I ordered some uv dye to see if there is any penetration into the node.  I took some pics using my usb microscope of some of my defective nodes.  Some pics of some of my pixels.  You are not allowed to view links. Register or Login

Looks like no coating on board, and they got too hot. 

[CaptainMurdoch]

OK, so I took the plunge.

Just adding a data point for people.

I took my tree out of my garage tonight and set it outside.  ( 46-count 12V WS2811 Technicolor strands folded over.  I had one bad node out of the box which was repaired prior to going live.  I haven't been running them white much but have been running them at full voltage during the 24 hour or so I burned them in a couple weeks ago and again for 4-5 hours last week after mounting them on EMT.

I used Steve's perl script to reduce the technicolor channels in my .xseq files to 60% and also installed buck converters after the SSC using 3-core pigtails.  These are on an Active Hub w/ SSC v4.  I've set the buck converters to 8.0 volts.  I did not do any testing on the voltage at the end of the strand since I sealed the ends up after cutting the 50-count strings down to 46 to fit them on the 10-foot EMT conduit.  I have not seen any color degredation, so I think the voltage drop is low enough on the 20-foot long 46-count strand for me to be OK running at 8V.

I am using the C9 covers and I also put a bead of silicone around the place on each pixel where the wire enters the "top".  I did this to the front and back of the wire so there is a 360 degree bead of silicone around it.  I'm not using a lot of whites in my sequences, mainly red/green/blue, but some fades or mixes.

These were ordered early October and shipped on November 24th.

I'll report back if/when I have any issues.

[dduck]

 I am using lower voltage (10-11v) and always about 50% intensity.   I
avoid white.    Last night I turned back on my lights and a fourth of my strings
had bad pixels.    Last night the weather was cold and windy with rather high humidity.
Other strings started to go bad right in front of my eyes.    I touched the bad nodes
and they were not hot.    If anything the bulb was slightly warm.  So I now have over
half of my strings with bad nodes that I pulled into my garage for further
inspection.   I hope to open some of them up to see what is going on.   

Before going back out after repair I plan to apply silicon or plastic dip as others have tried.

[dduck]

I am getting many many failures.   I fix 1 and another pops up on the same string.   I took apart a
couple.   The first one would light but would not pass data to the following nodes.
The second one would not light and it would pass garbage data.

Please tell me this is moisture damage and no conformal coating is applied!
Ray told me that conformal coating WAS applied to my order.

[Hauvega]

I do not see any coating.  Unless the coating is so thin the hot molten injected plastic removed it.  Production line failure all over.  Have any of the nodes recovered after being inside for a while.  Did You have the c9 covers?  Here is the link for more pics.  You are not allowed to view links. Register or Login

I have a string with 17 out of 50 defects with 2 nodes that are going to be further evaluated.  This string had c9 covers on.

[dduck]

I did not have covers on them.

After bringing them in I turned on the string and all lights turned on.   Then
in about 3 min.  it looked like I had two bad nodes.    This same thing happened
to me outside.   When I first turned on the lights all of my strings lit up.   Then
bad nodes started showing up after those same strings showing bad in a previous
night.

While working on this string other bad nodes were showing up.   So I don't know if
they were failing while I was working on them in the Garage or if perhaps
they are borderline flaky.     

[Hauvega]

Measure the resistance on the data line to either 12v or gnd to see what the resistance is.  It should be in the megaohm range.

[dduck]

I must have to admit that I have not checked resistance values on components much.
So this is what I found depending upon wire lead direction using my multimeter.

GND     DATA      value
+           -          6700
-           +          11000

12V      DATA        value
-           +          1480 at 2000K scale
+          -           shows 1

[arw01]

On the "dead" pixels, any of you bypassed the various pieces on the boards to see if it really was the 7805 or the ldo causing the issue?  I assume you can get at the LED itself and put some voltage on it, and put some 5v on the output leeg of the 7805 to see what happens?