CREE LEDs not all working

[Calvinorc]

Hey folks,

 

I just got finished putting my semi-DIY LED setup together. I used:

• NanoTuners BioCube 14 heat sink, drilled and tapped for CREE CR-E LEDs
  
• two 1050mA Thomas Research Products drivers
  
• 9 CREE XR-E Royal Blue LEDs
  
• 9 CREE XR-E Cool White LEDs
  

 

Bought replacement 16 gauge power cords from Lowes, along with 16 gauge stranded wire, 4-40 3/8" screws, nylon washers... I got some straight line crimping connecters for testing, and "waterproof" wing-nuts for the real deal. I'm using the Reefkeeper Lite, with the ACL module for timing and dimming control.

 

I got done soldering, let it cool for a day or two while I waited for the ACL to get delivered. Wired everything up just now, unplugged the current actinic and white PC lighting from the RKL, turned them both off, and plugged the LED power cords in. Tried the white, and all but two lit up. Tried the blue, and 4 didn't light.

 

I haven't soldered anything in literally decades, so I'm thinking that my solder joints might be the culprit. I checked really thoroughly for possible shorts... but didn't see anything obvious. No wire strands sticking out, no melted wires touching, etc. I'm thinking it must be the solder joints.

 

I guess my question is, now that I know which LEDs aren't lighting, and have them marked.... what do I do?

 

I have an El Cheapo volt/ohm meter, but have no clue how I could test for this. Should I just assume it's my sloppy soldering, and re-do each LED's connections? Is there some way to tell if one side of the connection is good, and the other is lousy? (aka, + is good, - is bad, re-do the -)

 

Here's some shots of the overall job, and zoom-ins for detail, along with a shot of it all (well, almost) lit up.

 

[smai-tai]

You should have one wire coming off of a positive (+) LED and one coming off the negative LED (-). One set for the royal blue and one set for the cool white. I can not tell how you have it wired. I am not sure, but mine is wired for the first LED in the run and the last LED in the run to make a + & - feed to the designated driver.

[Ahanix]

+1, I wired one of my led's backwards and this is what happened, I would double check all of your stars

[smai-tai]

+1, I wired one of my led's backwards and this is what happened, I would double check all of your stars

 

Yes I would double check all of your connections. Even check to make sure the wires are soldered onto the pad on the LED as well.

[Calvinorc]

I checked and re-checked as I was doing it, and I got the order right as far as + and -... + - + - + - all the way through on both strands. I have the black wires on the royal blue leds +- and the white wires on the cool whites +-. I'll take a better look at the solder tomorrow, and see if I can find anything wrong.

 

Is it OK to hit the soldering with the iron, and remove the wire, then re-do it? Will that screw up the stars?

[mpc]

If in a continuous string some LEDs won't light up there is one good explanation - short. They are shorted out through heatsink. I see you have soldered in a way that metal ends of the wires overlap the laminated part of the star PCB, it is not a good sign and means that with high probability they touch the side of the star board and that is enough for the short. You won't see it with eyes. Use a Beeper on your multimeter and check for shorts. One lead on the heatsink and with other one probe EACH soldering pad on EVERY LED.

That would be the first thing to do.

[Calvinorc]

Ok. Thats's what I'll do first thing tomorrow. I can have one of the in the office hold the wires for me with needlenose pliers so I can make sure the connection is just in the pads... Thanks for the advice guys!!!

[Calvinorc]

I just checked, and found a bunch of shorts. I removed the solder and wires, and soldered new wires to the un-used pads. One problem tho... one of the LEDs is shorting out even after I removed everything. I'm thinking I have a flux bridge... like some of it got under the actual LED pad, between the LED itself and the star. Will spraying it with flux removed, or alcohol, do the trick? Or do I have to do something more involved??

[mpc]

Ahhh - tomorrow came (after 5 days)...

 

You didn't have to resolder on new pads - simple more careful re-soldering would have been fine.

What goes for organic solvents - I would be very careful - tho you have XR-E with quartz lenses those lenses are bonded to the case with some I don't know silcony-rubbery substance. Be careful not to damage it - I don't know if it is solvent resistant or not - just so you know (Don't drop the LED in the bowl of acetone).

Sometimes your short can be associated with screw connection - rarely tho. I don't think that anything can be between star board and the copper trace on the top of the laminated layer - they are glued onto the star - there is simply no way. Check around the screws. And check for excessive solder - it might be touching a side of the star board or screw. If not then carefully check the perimeter of the star. Solution will be simple once you find the cause.