WS2811 Technicolor Strings Have High Failure Rate

[Made2Rock]

The problem --

  I will replace "bad" nodes and get the house working only to have more nodes go "bad" in a few hours. The nodes were carefully handled and are secured on the house by command strips so there is little to no stress on the wires. The lights are being driven by the SSCv4 controller hooked up to an active hub. I'll use x-lights to test the strings.

  I saw one string yesterday go from a "bad node" to the string started to work. So I'm not sure these are truly bad or if I am doing something work. NOTE: This problem has shown up anywhere in the string.

Any ideas??

Joe

[tbone321]

Did yu burn them in prior to mounting them on the house.

[Made2Rock]

Yep, I check and recheck things just to make sure. The one interesting thing I have noticed but brushed it off for now is The first node I can control is actually the 4th node in the string. So if I address the SSC to start address 100 is I turn on 100 it is the 4th LED that turns on. I'm not really sure they are related but it is another thing I noticed.

[tbone321]

One thing that people are getting away from is that the SSC was NOT set up to wor with the WS2811 and IIRC, there are no plans to change that.  I am not saying that this is your issue but it may be a part of it. 

[RJ]

The protocol used on the SSC works on the Ws2811 I do not think it is killing his nodes. the only issue is the 2811 must have the right clock speed set. 

RJ

[zwiller]

Just wanted to add that there is thread over DIYC that the technicolor nodes are not holding up well to water penetration...  I would get them inside and dry and test again.  Maybe try some Corrosion X? 

[Made2Rock]

They have been dry up until today so my guess is that is not causing the initial problem.

I will say I'm surprised they are not holding up to water. I cut one of the bad nodes apart to verify they were being run in high speed mode (which they are) and I can say the entire block that holds the LED is solid plastic. The leads going into the chip are about 1/2" of solid plastic. So they very enclosed.

Of course with my luck they add a new problem to the mix.

Thanks
Joe

[tbone321]

The protocol used on the SSC works on the Ws2811 I do not think it is killing his nodes. the only issue is the 2811 must have the right clock speed set. 

RJ

I didn't mean that the protocol was hurting the nodes, just that it may be causing some of his operational issues.

[Made2Rock]

Just wanted to add that there is thread over DIYC that the technicolor nodes are not holding up well to water penetration...  I would get them inside and dry and test again.  Maybe try some Corrosion X?

zwiller, I came home today and fired it all up to see how it was working and after a day of some rain the house that had problems run perfectly, half an hour later one string is acting up. The megatree on the other hand is a mess.

I'm thinking RJ is wondering what I am wondering. I've seen nodes be "bad" then they are good. A bad node can not turn around and become good. On the other hand if the clocking is borderline it would account to working one minute and not the next. The part that doesn't make sense to me is the wiki states the SSC uses a 800Khz clock and the high speed setting on a ws2811 is 800Khz. I tried to cut one of the nodes apart and from the remains it looked like pin 7 of the chip wasn't connected to anything. From what I have been able to find out this puts the WS2811 at 800Khz.

Joe

[zwiller]

Glad to read it does not appear to be a water penetration issue.  If I have learned anything from ordering lights from China they are constantly changing things.  (including a new higher current led)

I wonder if you have the new higher current led and are under powering?  Running 100ct is too much for the SSCv4, I guess around 80 is the new max.  Maybe test the current/try power injection.  Also what is distance from hub to SSC?  Maybe you are too far.  Try testing a shorter run?  Some more info about your setup might help.

[Made2Rock]

I was having trouble with a Zeus (but that is another story) So I pulled it and replaced it with a Active hub The very first string had trouble about 5-10 nodes up the string so it is not a power issue.

There was another issue that I didn't think was part of this but maybe I need to look it. When I program a SSC to an address that address shows up as the 5th node in the string. Something is weird about that but maybe this is something to look at.

Joe

[tbone321]

I would run the config utility on the chip again and make sure that your null node count is set to zero.

[rdebolt]

The TC nodes are having an extreme failure period. The 2811s have a voltage regulator that is burning up because of high voltage. You will not be able to run Technicolor nodes with an ATX power supply! Actually probably won't be able to run them with any power supply for this year!

[Steve Gase]

The TC nodes are having an extreme failure period. The 2811s have a voltage regulator that is burning up because of high voltage. You will not be able to run Technicolor nodes with an ATX power supply! Actually probably won't be able to run them with any power supply for this year!

So, the advice is to not use them to avoid further damage -- until a better power solution is found?


What about the Technicolor TM1804, does anyone know if they suffer the same problem? 

[arw01]

Now I thought an ATX power supply was one of the better regulated supplies available for non lab grade type work?  Are they spiking or not handling the transient load changes gracefully or something else to cause over voltage?