Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Is Bond wire disconnection the LED's biggest failure mechanism?

Status
Not open for further replies.

Flyback

Well-Known Member
In power LEDs for lighting use, is there a fault whereby the LED may just suddenly go open circuit because its internal bond wire comes loose and disconnects?
As you know, the bond wire is extremely thin, and its connection from the LED semiconductor piece to the metal terminal tab of the LED is done by cold-welding, which at best, is an in-effective way to connect such things…but has to be tolerated due to the thin-ness of the bond wire.
Also, the brittle bond wire eventually oxidises and just breaks apart. Operation at high temperature, and the temperature cycling, all act further to decrease the life of the LED’s bond wire.

I am seriously wondering if there is any point at all in making a long-life surge protector for a LED streetlight if its just going to fail from a disconnected bond wire.

Bond wires in LEDs…

https://en.wikipedia.org/wiki/File:...35_mm_RGB_Surface_Mount_LED_EAST1616RGBA2.jpg
 
Maybe they made the bond wire thin because always it goes to the LED chip from the front. It's interesting why they connect bond wires in front of the chip, but not in side.
 
I don't know about the most common, but in my experience it's certainly accounts for a significant number of failures. We also see a smaller number of LEDs which appear to have badly overheated, enough to have discoloured and distorted the plastic lens.

My experience is with LED stage lighting fixtures, so we're working with single-colour, white or 3-colour LEDs - the same seems to be true of all of them. The LEDs are typically soldered to an aluminium-substrate PCB - presumably this is the same construction as with your streetlamps? I've always assumed that the failure was brought on by thermal cycling and dissimilar expansion between the board and the LED. In which case, perhaps some lead forms are more susceptible to this than others?

Certainly I'd think that the reliability that we see from stage fixtures would be inadiquate for street lighting applications. I imagine (hope?) that streetlights must be run with a wider safety margin...
 
Most of the fixtures I see seem to have series chains of LEDs, often with several chains driven from sepperate outputs from the driver. Of course, that's what you'd expect given that you want to control the current through the LEDs.
Needless to say, that means one O/C LED takes out the entire string, making the fault far more obvious.
To be honest I've not worked on very much of this kind of gear lately, but what you've described deffinately feels familiar.
 
I am launching a line of multichannel LED sequencers (which I developed) that use a minimal protocol with time, vector & scalar data rather than vast amounts of raw scalar data for the sequence. They communicate via wifi etc.
Power LEDs like the commercial 5050 SMD types benefit from PWM control as their brightness varies non linearly with it. Thus you can get near full brightness at lower duty cycles. This means that less heating occurs and energy is saved.

I am considering a version 2 product that leverages a photocell to assess the lumens and deliver the optimum PWM vs lumen ratio for 1) Energy savings and 2) LED longevity. being IoT devices this data can be monitored in realtime and reported on.
 
I suppose making a smart IoT LED lamp fixture that self optimizes its energy use and extends its service life would be a good idea. Wanna contract me Flyback?:D
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top