HELP

[jarris50]

HELP... I can't figure out what's wrong with my wiring. I am trying to wire three Cree Q5's in series. I have tried several times in different order and I can never get more than two to light at once. Even more bizare is the first and third will light but not the middle. I have un-wired it and tested each seperatly and all work fine. For the testing I am using a wall wart set at 3.5v 350ma. Any idea of whats wrong??

Jarris




[img=http://img120.imageshack.us/img120/7623/1262675xj8.th.jpg]

 

[jarris50]

Now I know someone has to have a guess as to what the problem is..

 

[LED Zeppelin]

Try testing with much more voltage.

In series, the Vf of the string is going to be just over 10V @ 350mA, and around 11.2 @ 1A. At 3.5V, the emitters are probably very dim, and small differences in their Vf might not allow all to light simultaneously.

If you hooked these up to a dimming pot, the middle one will probably light last, but will light as you increase the voltage.

If you don't have a wall wart or power supply that can deliver the voltage, briefly try a car battery or rig up some cells.

 

[glockboy]

You did not cut the corner Positive and Negative of the crees, It will short to the ground.

 

[VidPro]

You did not cut the corner Positive and Negative of the crees, It will short to the ground.

that would be my guess too, that somewhere the connection points are making contact with the heat sinc, just enough to drive you up the wall., trying to figure out why each one or 2 even functions, but all of them wont.

 

[jarris50]

Thanks for the help. I'll try boosting the power tonight. I was using a 3.7v 350ma wal wart but I'll break out one of the big guns to give it a shot. I thought about the grounding issue so I unwired all the crees and tested each independently and they all worked out fine. What really sucks is the Cree's are Q5's that I paid 25 bucks each. Guess I should have waited for the ones mounted on stars to get here from China.

Jarris

 

[yellow]

• using stars would mean to add one "frontier" which the heat has to travel- no need to with Crees
• three in series need more voltage (You already know that)
• the one led not giving a glow is the one with the highest forward voltage (but thats not important)
• what kind of focusing do You want to use? The leds in the pic look mounted too tight (or was this just for testing?)
• scrape away the contacts on the back side of the emitter

 

[Gryloc]

Jarris,

They can light up independently if the leads are shorted to the heat sink, but when you power them up, the power could be going from the positive pad of the first Cree and going straight through the heat sink to the negative pad of the last. This can allow one to not light up. Do you have a voltage meter that can test continuity? If so, try seeing if any of the positive or negative pads shorted tout to the heat sink. That can happen easily, even when using AA thermal epoxy. If you do not have a multimeter with that function, get 1 or 2 AA batteries and hook it up to a spare low voltage light bulb (maybe from the spare from a M*g 2D or 2C). Then touch the remaining two open wires where you need to test continuity (one on the lead pad of the Cree and the other on the heat sink). If it lights, then you have a short. This should be a low enough voltage set up where you cannot do any damage to the LEDs.

I would try this before you start applying that much power. If the LEDs are shorted in a way where two are attached in parallel, then you can do damage since two in parallel will run at a max of only 7.8V. If you are only attaching only 3.7V to them all and only two are lighting up, then it is NOT because of not enough voltage! 3.7V can barely light up two white LEDs in series, let alone three (if there is not a short). Even though they do light up at low voltages, this 3.7V should not be enough to light up three, even dimly. If totally unsure and you just want to hook up more voltage, go with something like a battery pack that is around 7 volts, or maybe even use a little 9V battery. If two light up brightly, and not the middle one at all, then you need to check for a short for sure. I would hate to see you damage your $25 emitters, so I would slowly work your way up with testing (continuity, ~7 to 8V, then ~12V when confident).

Well, do not let this get you down. This stuff happens to all of us. Good luck with the testing. Please keep us updated. We are here to help you...

-Tony

 

[jarris50]

Ok, it looks like the problem might be a short. Does anyone have any pointers on how to cut the corners once there mounted to the heatsink?

Cheers,
Jarris

P.S. thanks to everyone for the help.

 

[3rd_shift]

Wire cutter dikes should do the trick.
Or just grind/file off the underside conductive strips completely and carefully remount with arctic alumina adhesive.

Once mounted?
A chisel/sharpened screwdriver and hammer might do it if carefully done.
Go gently at 1st.
Good luck.

 

[VidPro]

a dremmel with a small "cut off wheel" only the fiberglass cutoff wheel, so it doesnt break. and a steady hand, or steadied against something.
cut off wheels are designed to cut metal stuff, and will go through the ceramic like butter.

 

[scottv]

Uhm, try parallel instead of series.

or try 11V instead of 3.7

your choice.

 

[LED Zeppelin]

Maybe try solderwick to remove all your traces of solder, then a tiny drill bit to drill out the conducting vias at the corners. You probably don't need to drill through, just break the circumference of each hole. I think it would be relatively easy with a drill press, but difficult by hand.

Or just dismount them by heating the sink gradually with a torch or jet lighter, and carefully prying them off by twisting the bases with some good pliers. Mount the sink in a vise between some insulating wood if you try this. Don't apply any pressure to the dome or you'll end up with bubbles inside.

The Dremel w/cutoff would work too, but I'd be very careful.

 

[jarris50]

I like the idea of the drill press, seems easier than trying to cut the corners. Unfortunately I have to go out of town in the morning and won't be able to give it a go till next week. I have thought about parallel wiring but the end product will be in a Aelph 3 head running off two 17500 cells and a badboy 500. I'll let everyone know how it works out next week.

Cheers,
Jarris