How Long To Test New Pixels?

[orlandoal]

Hi All,

As a newbie to LED Pixels this year, I have finally managed to get something up and running for the holidays, allbeit not as much as I had wanted to get achieved seeing as I started wayyyyy back in January. I think I might have been somewhat over optimistic.

What I have found is that I am getting some pixels going bad, I did run the pixels on test for several hours when I first received them and once a prop was built I run it for a few days in my garage, with various test patterns from the controller.

Water is/has not been the problem, I live in central Florida and we have had next to zero rain over the last month and my yard irrigation system is turned off.

I have not had to replace loads of pixels probably 10-15 or so since I have been running the show on a daily basis.

Any thoughts on how long I should be testing new pixels for prior to building props?

If it helps, I am running 12v WS2811 pixels, any joints in pixels (more than 50 or where I have had more than a 4" gap between holes) are soldered and heat shrunk.

As usual, any pointers or suggestions greatly appreciated.

 

[CargoLights]

I don't "burn in" my pixels. I find the best indicator of whether they will have problems is to be a bit rough with them and flex the wires around. I seem to catch any bad pixels before I hook them up that way.

That said, I have had exactly 2 pixels fail on me after putting them out, and that's out of about 16,000. 10-15 in year 1 seems crazy. Have you looked closely at what was wrong? I find that most issues are bad soldering (hence why flexing the wires catches it), but I've also had one with a strand bridging data and ground, and one that the red channel just failed on (the last 2 are the ones that failed after installation). I'd be interested to know what kind of failures you're seeing.

 

[orlandoal]

The pixels that have gone bad have not been at a soldered point between pixels, they have been in the middle of a contiguous string of 50 or more, mainly the data line seems to have gone bad, so just soldered in a new pixel. I have had the odd bad joint in a new string of 50 and as you say flexing the wires seems to show that almost immediately.

I don't smack the pixels around, but I don't treat them gently either, maybe I just got unlucky and have some bad chips in the 4k of pixels I purchased.

 

[CargoLights]

Sorry, I wasn't talking about your solder points, I was talking about where they're soldered at the factory. It's quite possible you just got a batch with poor QA and you'll have to work through them at this point. It really is quite strange for that many pixels to work at first, then fail within a month of use.

 

[orlandoal]

LOL, I did not think you were having a pop at my soldering, although after soldering lots of connections on the props, they may well get a bit dodgy.

I cut off the plugs that come with the pixels and solder if I have to extend a string. I also solder on Ray Wu male and female ends to each prop as a matter of course as well.

It was after all the hours of testing and then the failures, I thought maybe there is a set way to burn them in and I had not come across it yet.

 

[merryoncherry]

Like @CargoLights I only had ~1 out of 10k fail in year 1. Year 2 has been similar with most brands and batches, new and seasoned. But I got a bad brand / batch that has lost 1 out of 500 and counting at this point, and an even worse batch that lost 4 out of 500 in a few days that I just ripped and replaced.

I don't know if you can catch it in burn-in, given that it took a week of running for the failures to appear. I try things once after any splicing, then I put them out. Good news that the worst batches were easily reached and repaired / replaced.

 

[orlandoal]

Luckily, the highest failure so far has been at around 10 feet off the ground, reached easily with my trusty ladder and all have been easy to diagnose and repair.

If it was one dead pixel, I could live with that, unfortunately, in pretty much every case, everything after the faulty pixel has been dead, hence me thinking they are mainly chip failures.

 

[merryoncherry]

It can be a chip failure. Or a failure in the soldering of the data line. Or a failure of soldering the chip to the board or the power resistor to the board.

Or if the positive or negative soldering failed, that'd kill the string too. Except that I know it isn't for most of mine... mine hold color after the failure and get power from both ends. A string lit up when it shouldn't be is way worse than one that dies dark....

 

[orlandoal]

It can be a chip failure. Or a failure in the soldering of the data line. Or a failure of soldering the chip to the board or the power resistor to the board.

Or if the positive or negative soldering failed, that'd kill the string too. Except that I know it isn't for most of mine... mine hold color after the failure and get power from both ends. A string lit up when it shouldn't be is way worse than one that dies dark....

"A string lit up when it shouldn't be is way worse than one that dies dark...."

You had to go make it more complicated didn't you

 

[Indigogyre]

Unlike the hooligans that already commented, I TEST my pixels similar to you. I run them for hours on my test controller and once more after I put them into a prop. I do get a little rough with them and toss them about while sorting and up on the table and such. In addition, I start to test to make sure theyw ill pass data out to another string for testing.

This is my second year and out of my 10k ish pixels I know of only one bad pixel. It stays blue in one of my Halloween talking pumpkins. I've had scares due to poor data or bad extensions but so far I am fairly lucky.

There is also a possibility that they could be damaged while pushing depending on how it is done. If you use pliers there is the chance of crushing something if not careful. I rarely will use pixel pliers.

 

[Clegg]

My 2c as a second year.

After going through some 50+ pixel replacements this year, my biggest regret is not testing the pixels at the start of the year. Next year when I replace them, they will be tested in the backyard for a few hours each night for a couple of weeks giving me time to replace the replacements if they also go .

I'm not doing this to rule out the one or two which are simple to replace during the show - rather identify significantly sooner with time to breathe if there are faults in the strings themselves.

Unfortunately time crept up so I only had a few hours to test some 4600 pixels in the weeks before and not for long enough. By then, there have been no pixels around to replace things.

The more pertinent question to me, to avoid a repeat (and I'm significantly better off than some that binned 10,000+) is how do we guarantee a better quality of pixel in Australia. Wally's gets a good reputation for control on the manufacturing process in USA (at least according to the Facebook groups), others have suggested that we might have got bad batches made a couple of years ago for cultural related reasons. I don't pretend to know the best way forward, but I'd be hell of a lot more reluctant to splash out $1,000s to replace what should have been working properly in the first place if the risk of total string defects remains.

 

[christmasdave]

I ran my show this year (my second) for a full week before December with no problems, then on 1st December had a bad pixel, from there I’ve been chasing pixels every night. With work hours I’m just marking them with tape to fix in the weekend.
Seems to all be solder point issues with mine, a quick hit with some pliers typically sorts them out for anything from a few minutes to a few hours, or if I’m really lucky a whole 2 days!

I don’t feel I’ve enjoyed this year as much as my first, but there was one night everything worked and I just sat out the front until my end time with a big smile.

I did communicate with Paul at lightemupleds who advised this year has seen a higher number of manufacturing issues.

 

[Croydon Lights]

How often are there whole string defects that require replacement at most it would be 3 or 4 per string but there appears to be people who can't or won't just replace the defective one but instead replace the whole string. This is a hobby where you have to do some work and not expect a simple install and set up and have it all working. It all part of the fun

 

[B4IGO]

I'm about 2 weeks in now, and still getting failures from "that batch"... though it does appear like it may be slowing down... I hope like heck I'm through the worst of it. So I'm not really sure how long you'd need to do a burn in to increase your chance of not having to pull props down and re-push them.

I do feel we're in a bit of a "pot luck" situation though at the moment, and maybe COVID has contributed to that significantly. One vendor had issues with a batch of epoxy a couple of years ago and a circuit board cut problem, prior to that they were considered one of "the vendors" to go to. This batch that seems to currently be causing people issues were, again, a vendor with a very good reputation. If this turns out to be a batch component or chip failure, rather than a manufacturing fault, I'm really not sure what the pixel manufacturer could have done about that either... So no manufacturer is particularly immune, it will be how the manufacturers deal with it and assist their suppliers and end customers that will differentiate them.

You hit the nail on the head though @Withdr4wn .... how to purchase in the future. $1k, $2k+++ on a batch you may have to bin with no guarantee of replacement is not very palatable going forward.

 

[B4IGO]

How often are there whole string defects that require replacement at most it would be 3 or 4 per string but there appears to be people who can't or won't just replace the defective one but instead replace the whole string. This is a hobby where you have to do some work and not expect a simple install and set up and have it all working. It all part of the fun

The trouble is, with what I am experiencing, it's not practical. I've had multiple strings go bad... with multiple failures through the string and it's not the first dark pixel, nor the two around it that are the defective ones. When I've eliminated the bad segments, the pixels that were originally dark at the start of the string came back to life.

I've cut one of the bad strings into multiple segments and the same fault (a circuit between +VE and GND) exists multiple places in the string. The only way to tell where the bad pixels are is to keep dividing until you find the good and bad segments, but in some cases the good segments can be just a few pixels... and when will those remaining pixels fail? It's not worth the hour+ it takes to eliminate all the bad bits and rejoin the good segments and have them fail again in the next few days. Its quicker to replace the string, hopefully with one from a batch that is good.