WS2811 Technicolor Strings Have High Failure Rate

[Steve Gase]

A conductor file is easy to tweak this way. The whole file is levels so write a script to open the file read each byte and lower it by the % you want the write it back out.

Sent from my Charge by Tapatalk

RJ

I've completed a .seq translation tool.  I'm looking at the .xseq capability now.


The tool allows a range of channels to be specified, and scaling and/or limit (max) to be applied to each channel in that range.

Steve, Buddy...... Christmas is coming.... Did you know it better to give then receive.....

I've never been one to stand in the way of someone doing a give thing so it you feel the need I would be more than happy to receive a copy of that editor.

 

Joe

This is just the first iteration.


It supports Conductor .seq files only for now... (channels 1-16384)
In many cases within xLights, you can convert from one format to .seq -- use this tool -- and then from .seq to another format.


Let me know if it isn't accessible:


You are not allowed to view links. Register or Login




The script uses Perl.  You can use one of many free Perl tools out there.

[Made2Rock]

Steve, I got it and many thanks. I'll play with tomorrow.

I may actually get a chance at putting this show back on the road this weekend. By adding a couple of full wave bridges (You can only use 2 of the diodes) I can drop the voltage down and I'll use your program to cut my levels by 50% I should be ready to attempt to fix up some of these strings.

BTW: I ran the xlights A-B-C-All test and with 8.8 volts into the string I had about 4.4-4.5 volts at the end of a 100 count string during the white part of the test. The LEDs never stopped working and I ran it for about half an hour like that. I will admit I never expected it to run that good with that low a voltage.

Joe

[Steve Gase]

Steve, I got it and many thanks. I'll play with tomorrow.

I may actually get a chance at putting this show back on the road this weekend. By adding a couple of full wave bridges (You can only use 2 of the diodes) I can drop the voltage down and I'll use your program to cut my levels by 50% I should be ready to attempt to fix up some of these strings.

BTW: I ran the xlights A-B-C-All test and with 8.8 volts into the string I had about 4.4-4.5 volts at the end of a 100 count string during the white part of the test. The LEDs never stopped working and I ran it for about half an hour like that. I will admit I never expected it to run that good with that low a voltage.

Joe

I just received my buck converters earlier tonight... I plan to play with them tomorrow.


I would like to get more info on the diode design... where are they placed?  what sizes?


Thanks!

[scharbon]

You know, I wonder if they built these strings to be 5 volt by accident instead of the 12 volt design.  Maybe that is what the problem is.  That is why the LED only uses 5 volts and teh regulators are burning at 12 V and the lights are working great at just under 5 volts.  Just a thought.

Steve

[johno123]

That's actually what was in the back of my mind the whole time - what's the difference between the 12V and 5V versions of the TCs?  Is the 5V versions dropped down to a 3.3V regulator?  Or the same 7805?

[dduck]

Steve, I got it and many thanks. I'll play with tomorrow.

I may actually get a chance at putting this show back on the road this weekend. By adding a couple of full wave bridges (You can only use 2 of the diodes) I can drop the voltage down and I'll use your program to cut my levels by 50% I should be ready to attempt to fix up some of these strings.

BTW: I ran the xlights A-B-C-All test and with 8.8 volts into the string I had about 4.4-4.5 volts at the end of a 100 count string during the white part of the test. The LEDs never stopped working and I ran it for about half an hour like that. I will admit I never expected it to run that good with that low a voltage.

Joe

From what you tested on voltage I became more courageous.   I left on my lights a little
longer at 10.11V at 50% white.   They are not actually blinking but rather at the end of
the strings dimming up and down.   The voltage at the end of the string is also
is rolling up and down from 6.09v to 6.12v  that seems to go along with the dimming.   
Then I went to 100% intensity full white and the lights looked good.   No dimming up and down.

The following is my update for my numbers but at 100% intensity which seems to be in line with yours.

#Color    Voltage drop
1            31.6
2            46.3
3            51.7

Also note this test was with an SSCv4.    For some reason I get slightly less voltage drop with this SSC than
an SSCv3.

[n1ist]

He connected the diodes in series, between the positive of the supply and the positive of the Zeus.   The cathodes on all are pointing to the Zeus.  Each one drops about 0.7V.  Just remember they must be able to handle the full current of the strips and will dissipate 0.7V x the current in power, so they may need a heatsink.  That's why he used the two diodes in series in a bridge rectifier module; it's easy to get high current and to mount to a heatsink.

This is better than just using a series resistor since the drop across a diode is (mostly) constant, while the drop across a resistor depends on the current (Ohm's law).  Of course, the switching buck converter is a better solution, but the diodes are an easy thing to try.

/mike

[arw01]

Ok, what's a switching buck converter? Is that a brand name of a power supply, a type of power supply?  A design of power supply?

Curious minds want to know if this might be a better power solution long term than atx supplies.  Thought next years order to Ray may include some big behind Meanwell units that I can vary the power down a bit.

[Made2Rock]

He connected the diodes in series, between the positive of the supply and the positive of the Zeus.   The cathodes on all are pointing to the Zeus.  Each one drops about 0.7V.  Just remember they must be able to handle the full current of the strips and will dissipate 0.7V x the current in power, so they may need a heatsink.  That's why he used the two diodes in series in a bridge rectifier module; it's easy to get high current and to mount to a heatsink.

This is better than just using a series resistor since the drop across a diode is (mostly) constant, while the drop across a resistor depends on the current (Ohm's law).  Of course, the switching buck converter is a better solution, but the diodes are an easy thing to try.

/mike

Bingo, give the man a cigar he is completely right.

A SILCON diode will drop about .7 volts in the forward direction it is called the forward voltage drop. Next problem is how to do this.

1) single diodes are actually a problem on where to mount them and cost almost what a full wave bridge does. A full wave bridge comes in nice little package that can be mounted on the side of the power supply. You CAN NOT USE ALL 4 DIODES IN THE BRIDGE. You can only use 2 of the diodes so to get a reasonable voltage drop you can put 4 diodes in series and get a little over 3 volts dropped.

So I decided to BUTCHER (you will have to butcher you cables to do this) my power supplies and put these diodes in. I've attached a pic.

Now for power dissipation I have something like 10 amps average going thru these and it'll be 1.4 volt drop across each bridge. So each bridge is going to have to dissipate about 14 watts and a 2nd one is right next door to it so in that little area It'll be 28 watts. I might get away with it BUT I REALLY WOULDN'T RECOMMEND DOING THIS.

I believe Mike talked about a heat sink and I believe one should be used but right now I'm just trying to get every thing working.

My connections are as follows for the bridge.

I bring in the 12 volts to one of the AC sides of the bridge, Then I jumper the + side to the - side and get my output from the other AC side. I then BUTCHER my cables and ran the 12 volts into this mess and pulled the 9 volts out from it.

NOTE: The big 24 pin connector also has 2 12 volt lines going into it so you have to cut those and cut the other 12 volt lines and run your 9 volt lines into them. You also have watch the amount of current you are running on these wires.

MY SETUP IS FOR TEST!!!!!!! AND ONLY RUNS 1 STRING OF LIGHTS. So do not think that little wire will work for 20 amps ( you will melt the wire ).

I got my bridge at radio shack simply because I needed them for testing. They are rated at 25 amps and have a forward voltage drop of 1.7 volts. I am seeing about 3.2 volts across the pair of bridges. NOTE: You only use 2 diodes at any time in a full wave bridge so its forward voltage drop is what you can expect so when it states a 1.7 volt drop that is for both diodes and like I saw 3.2 volts across 2 bridges gives you 1.6 volts across each one. So you can expect to see what the forward voltage drop is.

NOW FOR THE PUREST

A diode does have resistance and this resistance will cause a change in voltage drop with a change in current. I've seem a couple tens of volt change in my testing with only one string. My option is there is a regulator on the chip I'll let it deal with these variation, besides my lights are junk any fluctuation in brightness is a huge improvement over not working. So I DON'T CARE and I don't think it will matter.

Hope this helps and I wouldn't recommend it unless you have some tech background to understand the variables not addressed above.

Joe


 

[tbone321]

A buck converter is a DC to DC converter which is in a way a fancy regulator.  The primary advantage to them is they are able to drop voltage without wasting energy or creating excess heat.  The down side is they tend to get big when they are designed to work with high current demands.

[Steve Gase]

The "Dimmer" tool for xseq (xLights) raw files is now complete:
 
SeqXlater-xseq.pl
 
 
The conductor file version was also updated -- it had a problem where the channel number comparison was off by one.
 
SeqXlater-seq.pl
 
 
Both are found here:
 
You are not allowed to view links. Register or Login

 
Note: I found it necessary to do the xLights file format because it can deal with more than 4x4096 universes.  Conductor format is limited to 16384 channels because Conductor uses multiple units to handle every new group of 4 universes.

[dduck]

Status update:   I went ahead and changed my mega-tree to 10.1V and kept my show at
50% intensity with no whites.   Everything was going okay.   Then I started to get a misty
light rain.    A bad node showed up and lights after it where white.   So I wanted to see if
by powering down the power supplies then powering them up after a short waiting period
if they would turn the all white nodes to off.    Nope...   At least 3 more bad nodes showed up
on other strings and some were near the start of the string   

So I'm thinking it had to do with the rain.  I'm not powering these up again till it is sunny.

[keitha43]

I am beginning to think in quality control testing that instead of soaking them in a bucket of water, they should test them with a wind machine and fire hose.

[Hauvega]

I my r&d,  I found no conformal coating on the pcb and wire.  Measure the resistance on the data and gnd wires and see what the value is.  The normal value should be the high megaohm.  I plan on testing some of my modded nodes to see how they hold up.  I would like to get hold of some of the final prototypes for analysis.

[keitha43]

I am curious as to how the water is getting inside in the first place. Maybe the black material is not bonding to the wire jacket or the led and needs some other material?