Problems mounting leds....

[scott011422]

Hey all, I'm having a hell of a time figuring out why I keep blowing leds...

I'm using these leds:

http://www.ledsupply.com/creexpe-blu-1.php

Mounted like this:

They work great after I solder on the wires. They stop working after I mount them, Never to work again. I have Arctic Silver between the copper and the star.

I have 3 assemblies that work great. I had one more that worked untill I took it off the copper and remounted it. What am i doing wrong??

 

[CKOD]

Are you sure those screws arent causing problems? A lot of times even though there is solder mask, the top copper is still spread out that far, and if your screws are breaking though the solder mask and shorting out the + and the - of the LED (possibly even though the thin insulating layer and shorting out the MCPCB to itself)
Try either nylon or fiber washers for insulation/padding, or if your project lends itself to it, use some arctic silver thermal adhesive and glue down the star, and then remove the screws and washer after it cures 24 hours. Ive done some mcpcbs onto heatsinks with ASTA just lining it up, and putting the MCPCB/heatsink/led into a 2-ton manual press. The hole in the ram for the die was prefect to go around the LED and not touch it while it pressed down the star. Solder up the leads afterwards and it was good to go. (80W soldering iron helps with that last bit )

 

[scott011422]

I'm fairly certain its not breaking thru. Once I take it apart, you can't tell it ever had scrwes on it. However, the 3 that work use Socket head cap screws. I ran out and switched to these philips. I took and desoldered one of the dies, It has an internal short. In fact, all 4 I fried have a short. Maybe your right

 

[Antony R]

Interesting idea.... Can you make sure the life time of this led supply for a long time than the usual supply.... Are any features are in?

 

[David_Campen]

I always use insulating washers, I favor teflon washers but as the other poster suggested fiber and nylon would also be suitable. Also I use cap screws with machine threads and tap mating threads into the heat sink. Also I put the screws in the notches of the mcpcb star as was intended.

 

[CKOD]

If you desoldered the LED from the MCPCB and it was measuring short across the leads of the LED (like an actual short, not just checking with the beep mode or diode check etc...) then it sounds like something else is causing problems. What are you driving the LED with? Ive had power supplys that misbehave before going into constant current mode, and the turn on surge would blow out XP-Gs, while XM-Ls and SST-50/90's could take it. But they usually failed open, not short.

 

[scott011422]

I am using this driver:

http://ledsupply.com/flexblock.php

I'm not sure on the driver, all the leds have 3 or 4 min on them after I attached the wires. They all blew instantly once attached to the heatsink.


How about a different plan of attack, What can cause these leds to short out?? Short of connecting the wires backward which was impossable in this case.

 

[moderator007]

The only reason it should would be to much current. If there was a short in your connection to the led it should affect the driver. Not the led. Unless its causing some kind of spike to the led.
It will be interesting to here what your actual problem is once you find.

 

[pavithra_uk]

1. Make sure Heat sink surface is flat. Specially when you screw like that, if there is a bent in surface, MCPCB also bent when you fasten. finally LED crack.

2. May be driver issue (no current regulation)

3. May be top copper trace shorted with Metal core (when you fasten screws)


Solutions:
1. Try to manual supply for LED once mounted on sink (5V supply with series resister)
2. For shorted LEDs, desolder LED from MCPCB and check.



Did you run LED wrong polarity ?

once (even 1 second) run with wrong polarity, Its internal ESD protection diode damage and become short.

 

[scott011422]

The heatsink should be pretty flat. It was machined in a Lathe, Then Sanded with 2000 grit on a granit surface plate.

As far as I know, The drivers supplys 700ma and sends along however much power is needed to keep it that way.

As said above, The led was unsoldered and was shorted.

Also as above, Not possable to reverse polarity. (Led ran on non reversable plug)

 

[sn0wBLiND]

Connecting the driver with the LED in boost mode would result in direct drive and a massive over current if Vin is higher than Vout.
This would be hard to detect visually as you are using blue LEDs.

 

[scott011422]

Really? This may be my problem then. I'm testing on a 12v powersupply. On the first three, I had aligator clips on the led driver. Connected the driver, then turned on the power. However, after that I installed a plug on the driver. So in testing things out I left the driver on, then just touched the bare wires to the plug ends as I went along. So your saying, with no load, and an input of 12v, the driver was putting out too much power. Then When I touched the wires to the already running driver, It dumped that power into the led, buring it out???

 

[LEDite]

I'm fairly certain its not breaking thru. Once I take it apart, you can't tell it ever had scrwes on it. However, the 3 that work use Socket head cap screws. I ran out and switched to these philips. I took and desoldered one of the dies, It has an internal short. In fact, all 4 I fried have a short. Maybe your right

I think the bare metal of the mounting screws may be contacting the bare aluminum of the sides of the star.

Take the screws out completely, and test it with about 3VDC.

LEDite

 

[MikeAusC]

. . . . So in testing things out I left the driver on, then just touched the bare wires to the plug ends as I went along. So your saying, with no load, and an input of 12v, the driver was putting out too much power. Then When I touched the wires to the already running driver, It dumped that power into the led, buring it out???

The output capacitor will charge to the supply voltage if no load is connected - when you connect the LED this energy dumps instantly through the LED.

Also note that for this Driver BOTH sides of the LED must be isolated from the torch body.

 

[scott011422]

I think the bare metal of the mounting screws may be contacting the bare aluminum of the sides of the star.

Take the screws out completely, and test it with about 3VDC.

LEDite

I know they are. Why is this bad? According to the datasheets, the heatsink does not need to be isolated.






I wanted to thank everyone for helping me out. Very new with Hi power leds and drivers. This is my first project. Making wall washers for Christmas. Now I can finish the back of the house and other side without blowing my leds. Thanks again!