More Unwanted Rainbows

[rm357]

If you follow the link on the wiki to the rigid strips in Ray's store, he has the string protocol data on that web page. The chip in the rigid strip does 3 RGB LEDs, but the protocol is basically the same for the nodes, with 1 chip on the rigid strip equivalent to 3 nodes in a string. You should be able to see the data with scope, but picking out individual words of data will take some work.

I don't think the policy is intended as a tech-speak gag order, but if the conversation starts to get real technical, it's best to take it off-line from the forums...

Good luck.
RM

[rdebolt]

I think you need to go to the closest 7-11 and buy some lottery tickets if you can guess that good.  That is incredible luck if that was in fact the bad node. 

Steve

I was not that lucky! Still struggling!

[rdebolt]

The string that I am fighting now I cut at what I thought was the bad node. Then tried both ends seperately and they worked fine so I soldered them back together in reverse and halfway through the second half color mixing starts. It is fairly consitent around a few nodes, but once about every 25-40 cycles one further up the line will not match. The problem then is it is not the same node upstream every time!

[Steve Gase]

Did you resolve your issue with the rainbows?

Out of my order of 20, i had 9 that started showing a rainbow effect about 3-5 cycles into the test mode.
Using the same controller all of the time, I had 11 more that looked good after 5-10 minutes.
I took the 11 that looked good, placed them in the garage, each with their own controllers (all in test mode) and they looked goo for 10 minutes, then I started seeing 5 go into rainbow mode.  After 20 minutes more I saw another 2 go into rainbow.

the rainbows are all smart pixels.  I also have square pixels using test mode and they've been running without issue for about 7 hours now.  the square pixels have a count of 50.  the smart pixels are a count of 128 each.

I'm hesitant to start cutting strings, I've see the pixels having problems be about 75 pixels from the start... then start acting up at 25 pixels from the start.

[Steve Gase]

Maybe the problems are heat-related.
After 10 hours all of my strings are now in rainbow mode.
All are still in bundles as I received them from Ray.

I powered them off, then back on again and most started clear but then started to go rainbow on me within minutes.

next test will be to stretch them all out and let them cool for a while first.

[rdebolt]

Steve,

First you do need to unwind your strings as they do build a lot of heat.

Second I gave up on trying to run 128 nodes. I had very few strings that would work without rainbow nodes. I am now going to run 32 sets of 64 nodes rather than 16 sets of 128 nodes. I have found that running fewer nodes has greatly reduced my number of rainbow (bad) nodes. One thing to remember is when you pin point the first rainbow node- it is almost always the node before that node that is causing the issue. I am not completely done yet, but I am getting closer.

On a finishing note, I should have started this way back last summer, I know; but finances kept me from purchasing my nodes until recently. However I am extremely frustrated with the time that it is taking to get these to work. Don't get me wrong, I am NOT blaming anyone, but I have hours upon hours into just getting the strings to run properly. I just could not get them any earlier than I did. 

Ok I have vented and feel better now, back to work on my lights!!!

[Steve Gase]

I'm surprised that no one else describes this as a universal problem.

[mokeefe]

Oh it's more universal than you think.  I know I've read a number of posts with this exact issue including a few from me.  It seems to be reported more often with recent strings from Ray, but I don't honestly know how many people are working with full 128 node strings from earlier in the year purchases.  The thing that frustrates me the most is that the string behave a 100% correctly with the TH2010 controller from Ray.  RJ thinks that unit is able to drive more current to overcome whatever issue is causing these intermittent rainbow and solid white issues.

Wish I had a solution but so far no luck, other than cutting everything into smaller strings as others are doing.

-Mike

[Steve Gase]

Before I had my SSCs assembled, I was using a black box LT-DMX-1809 DMX Decoder.  That box worked well with my testing of a coroflake with square pixels during assembly but with a quick test of smart string pixels I was having terrible results.  no string worked well no matter how small I made it. 

moving to the SSC I started seeing initial success, and I was impressed that it was much better results than the black box.

of course it soon started degrading to the point none of the 128s were working beyond a few hours.

BUT the good news !!! is that I shortened my strings to 85 each, spread them out, and none of them are having problems so far.  only trying 5 strings so far, but after 6 hours it shows hope of working.

why 85?  ...4096 channels for an active hub, divided by 16 strings, divided by 3 channels of color, is 85 pixels.  If I had the power situation arranged to handle it, 85 allows me to get the highest density for a hub.  (Power for pixels and SSCs is 42 amps - 12v.)

I don't know if the initial success for 5 strings will carry forward to my other 35 strings, but I'll stay hopeful.  My next problem is to figure out how to build another 20 SSCs to use with all of the remainders (cut-offs) from the 128-pixel strings -- now 85.

[jnealand]

Interesting observation.  I was trying to demo a string of 75 pixel nodes using the test hex.  I had already tested them in my workshop using the active hub and test hex with no problems and they were now hanging on a gutter.  I powered them up using a 9v battery (that I usually use to demo a 10 node cut off string) since I do not have the PS and hub outdoors yet.  All the lights lit, but with a definite rainbow effect.  RJs theory of not having enough power may be the exact issue and is also why shorter strings do not seem to have the problem or seldom have the problem.  We need more power for the longer strings.

[JonB256]

Low voltage really affects Blue first and Red last.

Sent from my DROIDX

[Steve Gase]

if my strings were not stretching away from the power source, and instead were circling back, then I'd try connecting the 12v+ and ground back to the source -- so front and back were both connected.  i've  done this with a coroflake and it worked well to keep intensity about the same throughout.

...something to consider if your application allows.