I mentioned in a differnent thread that I had some problems with the newer rebels which are manufactured using the TFFC (Thin Film Flip Chip)process.
http://www.candlepowerforums.com/vb/showthread.php?p=2085634#post2085634
I have had quite a bit of experience with the older 50 lumen per Watt cool white Rebels as well as every other color Rebel made. I have hand soldered them to 1/32 copper sheets to be used as heatsinks with the power terminals hanging off the sheet for easy soldering of wires. I had found this an easy way to use and evaluate these parts.
Well I recently received some 100L Cool White and 70L Neutral White Rebels and immediately damaged the first 3 I worked with. The problem left me quite baffled but I have now figured out that it is a result of thermal damage due to the hand soldering of the thermal pad. The symptoms are strange and for me started to present themselves in the first 15 minutes when being driven at 350ma and even sooner at higher currents. The first thing I noticed is a strange oscillation in the brightness kinda like you have a bad wire or switch somewhere in the circuit. I hooked up an oscilloscope and noticed the Vf of the LED was fluctuating about 20mv. It looked like a bunch of random noise on the AC coupled scope. I hooked up older 50L Rebels and there was absolutely no fluctuation, I also tried several other Lux1 emitters which showed no fluctuation in output or Vf. This eliminated my power source as a culperate. I tried using a Li-ion CR123 with a dropping resistor and an LM317 based current source. The next thing I noticed is there was a minimum current that was required for the parts to emit light. I noticed it at 15ma and after running a damaged part for several hours at 800ma it had progressed to the point that it would not illuminate below 400ma. Mind you it still consumed the power below this point but t would not light. I next noticed dark spots on the LED die and as the parts degraded over the space of a few hours the LED got dimmer and dimmer.
It is my opinion that these newer parts are not capable of being hand soldered without being damaged. My Iron is a Metcal and I am quite good with it. I was using a 700F tip at the time. I have had success using a toaster oven as a reflow oven. I place a K type thermocouple inside and manually control the temperature to try and follow the profile of the reflow profile but I sued Pb solder which melts at a lower temperature so I didn't need to take it up to as high of a temperature as the PbFree profile called for. My reflow never exceeded 200C.
In the picture you can see some of my past experiments. On the bottom right you can see where I hand soldered a 100L to a small piece of copper, I used wire cutters and sniped the bottom of the part off to break the connections to the power pads on the bottom of the board. I then scraped the film off the top traces and soldered wires to the top of the part. I have not tried soldering to a coper sheet using the oven method but I think it would work. That set up would lend it's self to being great for the hotlops. Mag heatsink. The one face up in the middle is epoxied down to the Copper using AA and seems to work rather well.
http://www.candlepowerforums.com/vb/showthread.php?p=2085634#post2085634
I have had quite a bit of experience with the older 50 lumen per Watt cool white Rebels as well as every other color Rebel made. I have hand soldered them to 1/32 copper sheets to be used as heatsinks with the power terminals hanging off the sheet for easy soldering of wires. I had found this an easy way to use and evaluate these parts.
Well I recently received some 100L Cool White and 70L Neutral White Rebels and immediately damaged the first 3 I worked with. The problem left me quite baffled but I have now figured out that it is a result of thermal damage due to the hand soldering of the thermal pad. The symptoms are strange and for me started to present themselves in the first 15 minutes when being driven at 350ma and even sooner at higher currents. The first thing I noticed is a strange oscillation in the brightness kinda like you have a bad wire or switch somewhere in the circuit. I hooked up an oscilloscope and noticed the Vf of the LED was fluctuating about 20mv. It looked like a bunch of random noise on the AC coupled scope. I hooked up older 50L Rebels and there was absolutely no fluctuation, I also tried several other Lux1 emitters which showed no fluctuation in output or Vf. This eliminated my power source as a culperate. I tried using a Li-ion CR123 with a dropping resistor and an LM317 based current source. The next thing I noticed is there was a minimum current that was required for the parts to emit light. I noticed it at 15ma and after running a damaged part for several hours at 800ma it had progressed to the point that it would not illuminate below 400ma. Mind you it still consumed the power below this point but t would not light. I next noticed dark spots on the LED die and as the parts degraded over the space of a few hours the LED got dimmer and dimmer.
It is my opinion that these newer parts are not capable of being hand soldered without being damaged. My Iron is a Metcal and I am quite good with it. I was using a 700F tip at the time. I have had success using a toaster oven as a reflow oven. I place a K type thermocouple inside and manually control the temperature to try and follow the profile of the reflow profile but I sued Pb solder which melts at a lower temperature so I didn't need to take it up to as high of a temperature as the PbFree profile called for. My reflow never exceeded 200C.
In the picture you can see some of my past experiments. On the bottom right you can see where I hand soldered a 100L to a small piece of copper, I used wire cutters and sniped the bottom of the part off to break the connections to the power pads on the bottom of the board. I then scraped the film off the top traces and soldered wires to the top of the part. I have not tried soldering to a coper sheet using the oven method but I think it would work. That set up would lend it's self to being great for the hotlops. Mag heatsink. The one face up in the middle is epoxied down to the Copper using AA and seems to work rather well.
-just cut each + and - contact pad so two would be wired in series -see Figure 1). However, because I was using my soldering iron, the solder would not even melt enough, or for long enough, for the Rebel to actually stick any (both surfaces were tinned). The star dissipated the heat way too well. This was a thicker aluminum MCPCB I got from Cutter when I got my first Cree P3 and not like the ones I got from DX :sick2:.
