I am building a high-powered flashlight using four Cree XHP70.2 12V LEDs, but just after the final assembly of the flashlight, they all started malfunctioning in some way. See the images:

One LED is is half bright and half dim along a diagonal:

Broken LED 1

Two LEDs are half lit like this one:

Broken LED 2

And the last LED does not light up at all. These are images of the whole assembly (with reflector removed):

Broken LED 3

Broken LED 4

All of that gray gunk is insulating thermal paste. The blue dots on the corners of each LED are beads of silicone glue (in an effort to prevent short circuits through the exposed corners.) All of the LEDs are wired in series.

I've checked every combination of leads, and there are no short circuits. There is also no conduction between the leads and the copper bottoms. All of these LEDs used to work fine. My best guess is that I somehow damaged them with excessive pressure after mounting the flashlight reflector on top.

Are there any better explanations for why each LED is acting the way it is? How can I fix whatever damage has been done?


I tested each LED on its own, and they all work properly. Somehow, the combination of all of them on the aluminum plate messes things up.

  • \$\begingroup\$ If those are 12v chips, then all 4 segments should be wired in series and they should be all on or all off. Whats the minimum voltage you need to get them to turn on? \$\endgroup\$ Jun 7, 2020 at 3:28
  • \$\begingroup\$ Also, it looks to me like you have them wired in parallel, not series. That or you have a short around the top diode that didn't turn on at all. \$\endgroup\$ Jun 7, 2020 at 3:31
  • \$\begingroup\$ They are definitely wired in series. The same current is going through all of them, even the one that is not lit at all. Even though they are 12V chips, they began to light up at 20V (but they should light up at 48V). There must be a short inside or under the LED that I cannot measure. I'm not sure how that could happen though. \$\endgroup\$ Jun 7, 2020 at 3:37
  • \$\begingroup\$ The 8 lit up segments times 2.5v would give exactly 20V, so the other segments must be shorted out. I'm surprised that flexing the PCB would short that many segments without introducing an open circuit, but I suppose that is possible. \$\endgroup\$ Jun 7, 2020 at 3:48
  • \$\begingroup\$ You tested by removing each one? Or by testing just one still onthe heatsink? \$\endgroup\$
    – Passerby
    Jun 7, 2020 at 5:26

1 Answer 1


I found the problem. After slightly moving each LED around on the aluminum plate, different parts of the LEDs flashed on and off. The layer of copper that composes the circuit in each PCB is exposed at the edges, and was contacting the surrounding pieces of aluminum.

The LED that had a diagonal dimmer part was damaged by putting too much pressure on the dome. I guess I'll just have to rip the dome completely off that one.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.