I've recently had two situations where LEDs started blinking. One was my Mantis microscope illuminator, where the current regulator failed and pushed too much current into the LED array. Some got bright, some went dark, and some started blinking independently at several Hz. Then yesterday my key-ring LED started blinking, bright and fast, and later in the day failed.
I wonder what's the blink mechanism.
You could write a poem about that.
Was that before or after the trip to Zeitgeist and how many did you have there? :-)
Usually thermal or overcurrent protection in the driver.
If the source can deliver a good amount of power, the cause is most likely cyclic breaking of the circuit, like a bonding pad wire separating then coo ling down- making contact- heating up and opening again. But the key ring i s something else, that one sounds like some kind of error code transmission before it died for whatever reason.
Please confine your remarks to technical content.
After, but just one.
The Mantis only has a sip resistor section in series with each LED. The key ring light has two (!) 3-volt lithium coin cells in series, direct connected to the LED! The pushbutton bends one leg of the LED to touch one of the batteries.
So it's the LEDs that blink, apparently when overloaded.
cyclic breaking of the circuit, like a bonding pad wire separating then cooling down- making contact- heating up and opening again. But the key ring is something else, that one sounds like some kind of error code transmission before it died for whatever reason.
The key ring isn't a remote, it's just a little cheap light thing. Inside are two lithium batteries and the LED; that's all.
The only mechanisms I can imagine is the one Fred mentioned. A bond contact is failing, the bond wire bends up when hot and opens, then back down after cooling off. Since this connection is very short the thermal effect will be quite fast. Eventually the area where it developed an unplanned "relay contact" erodes off and the LED fails open. If you still have one of the Mantis LEDs you could try to image through sideways.
What these LEDs have in common is that they are white. There is a glob of YAG-loaded epoxy over the die. Maybe the epoxy expands when it gets hot and rips the wire bond off the chip. When it cools, it jams the bond back down.
My gut feel is that it's probably just the bond wire. I can't imagine heating and cooling a whole blob of epoxy being able to create blinking at several Hertz.
ly cyclic breaking of the circuit, like a bonding pad wire separating then cooling down- making contact- heating up and opening again. But the key rin g is something else, that one sounds like some kind of error code transmiss ion before it died for whatever reason.
That's probably just a defective LED then. Isn't separation of the die wire s a major issue with some of the cheaper ones that aren't hermetically seal ed properly.
ly cyclic breaking of the circuit, like a bonding pad wire separating then cooling down- making contact- heating up and opening again. But the key rin g is something else, that one sounds like some kind of error code transmiss ion before it died for whatever reason.
Now that I recall from a SED discussion about 5 years ago, it has something to do with knockoff manufacturers not properly attaching the die to the pa ckage substrate, losing the thermal connection to the exterior leads, which results in eventual overheating and separation of the connections.
Well, are you AC or DC ???
Have something to add? Share your thoughts — no account required.
Ask the community — no account required