WS2811 Technicolor Strings Have High Failure Rate

[CaptainMurdoch]

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!

I understand the voltage regulator issue and that the WS2811 is constant current, but I've been wondering if the effect could be reduced by lowering overall current as well.  For instance, is the failure rate higher if you run with white or mixed colors versus just using the primary red, green, or blue individually since the total current through the voltage regulator would potentially be only 1/3 of an all white draw.

I'm crossing my fingers, but I've burned mine in via an Active Hub and SSC v4 for approximately 24 hours and haven't had a single pixel die on me during testing.  This is using an xLights test sequence, mainly the A-B-C red-green-blue color cycle, but I also let them run 6-8 hours in 'mixed colors' mode the first day.

How much time have people had on their pixels before they started seeing issues?

I'm also curious if anyone is reporting success without any issues.  Normally people don't speak up if things are running fine.

Since I've heard some people talk about turning their voltage down to 10V, I've also considered the possibility of picking up an adjustable power supply and doing my own 10V power injection right after the Active Hub.  I believe the SSC would handle the 10V power input just fine since it has it's own 7805 chip for the logic on board and the input voltage is just passed through to the pixels.  For my small set of 8 strings and the fact that I have the SSC's sitting in the same battery box as the active hub (since I don't have my zeus yet), I think this would be easy enough for me to do.  Any thoughts?

[CaptainMurdoch]

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?

I think it's more of a heat dissipation issue which is why people are talking about turning down the voltage so there is less power/heat to dissipate in the voltage regulator.  That's also why I'm curious about using just Red or Green or Blue to keep the current down to reduce power/heat dissipation in the regulator.  To me, using just R, G, or B could have more of an effect (on required power dissipation, not your show) than lowering the voltage by ~16% (10v vs 12v). 

[keitha43]

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

It seems rdebolt has had bad problems with the TM1804 version of the Technicolor nodes(you should see his video). So much so that Ray is having to replace all his nodes with WS2811 versions. The WS2811 version is undergoing a design change to hopefully improve the problems they have been having with that version. It seems many people are trying to lower their voltage to around 8 volts to see if it will help with voltage regulators overheating on the WS2811 versions as a work around.

[Steve Gase]

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

It seems rdebolt has had bad problems with the TM1804 version of the Technicolor nodes(you should see his video). So much so that Ray is having to replace all his nodes with WS2811 versions. The WS2811 version is undergoing a design change to hopefully improve the problems they have been having with that version. It seems many people are trying to lower their voltage to around 8 volts to see if it will help with voltage regulators overheating on the WS2811 versions as a work around.

rdebolt, can you post the link to that video?

[keitha43]

You are not allowed to view links. Register or Login

[Steve Gase]

makes me want to cry...

[keitha43]

If I remember right his burn in went fine. It started after putting it outside and got worse each day. It appears Ray's "design engineer" made his own changes to Zack's original design.

[Made2Rock]

You are not allowed to view links. Register or Login

That's about what my megatree looks like. The house on the other hand works one minute then is broken the next. Problems come and go with the house but it be good then go bad and turn around in another minute and be bad.

On a side note I built a new dmx dongle and used a passive 4 port to create a new programming station. I no longer have the problem of the first few nodes not working. Now number is actually number 1 (what a novel idea).

Joe

[Made2Rock]

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!

My lights are running about the same as yours. Can you give me a link to where you found this out. I can tell you in my case what is bad will turn around in the next minute and be working but There is always plenty of things wrong in my display.

Joe

[keitha43]

zwiller mentioned the other site in his post. You just have to spell out the initials.

[dduck]

I was running my 100 count 2811 TC at 50% intensity.   My burn in was just R, G or B.
No problems.   Also ran my show for the past two days with no problem.
Tonight I noticed 10 lights at the end of one string are out.    I think I will follow
the advice to just do R, G, or B in my show and at 50%.   I was previously trying just to
avoid all white.

The other site mentions also possible issues with moisture.   I have been getting lots of fog the last few nights.

[zwiller]

So the circuit/design is different between Ray's 2811 "regular" nodes and the TCs and the issue is only with TCs? 

[keitha43]

Moisture doesn't appear to be the main problem however there is a process(can't remember what it is called) where the circuit board is sealed in a substance as a last defense against moisture before the black material mold is formed. And apparently the production people were not doing that process and Ray was pretty upset about it from what I read.

[zwiller]

Conformal coating IIRC.  Starting to realize why my lights never shipped, Ray must be in damage control mode. 

[keitha43]

So the circuit/design is different between Ray's 2811 "regular" nodes and the TCs and the issue is only with TCs?

You might read up on the other site if anyone has had trouble with the regular node shape. I only jumped over there to read up on the Technicolor strings after someone here mentioned a lot of people were having failures and were posting their findings over there. It must suck to design something and then the manufacturer make changes on their own and the inventor gets blamed for it. I still think the Technicolor design was very creative and once the circuit board problems get fixed it will be the node shape to have.