[Elecraft] totally OT, except to the extent that it's QRP: LED lamps

[David Gilbert]

Many years ago I managed the optoelectronics portion of a large U.S. 
semiconductor operation.  We weren't very good at what we did and I 
suffered through several quality issues as a result, so I feel I have 
some backside expertise in this area.

The lifetime of an LED is a function of four things ... the size and 
density of crystal lattice defects in the chip itself, the current 
density through the junction, the temperature of the chip, and the 
reliability of the connections to it.  While current density typically 
determines the temperature of the chip, high current density alone will 
propagate crystal defects even if the temperature is held very low in a 
cold liquid.  There are various methods for creating the semiconductor 
junction of an LED and some are more efficient than others in terms of 
light per watt, but lattice defects negatively affect power out 
efficiency in all cases, and it is the propagation of those defects that 
causes the devices to dim over time.

In recent years, process design and control has improved to the point 
where light output efficiency has reached levels I never dreamed were 
possible, which almost for sure means that lattice defect densities (and 
therefore device lifetimes) are much improved.  And as Wayne says, that 
higher efficiency allows the devices to be run cooler ... which also 
would make them live even longer.  It seems to me that the life of 
modern LEDS should mostly be determined by how hard they are being 
driven (several LEDs driven moderately versus a few being abused), the 
effectiveness of the heat sinking for the LED package and whatever it is 
fastened to (I suspect most designs rely on the leads to carry away the 
heat), and the interconnections to the chip.  With reasonable design 
margins and not too much carelessness on the packagings side, I'd expect 
them to last a very long time.

By the way, I don't think a bridge rectifier is really needed.  If I 
were doing the design, I'd probably just use two strings of LEDs 
back-to-back across the AC line ... driven slight harder but with an 
inherent 50% duty cycle.   The distributed heat dissipation of several 
LEDs might be easier to deal with than the dissipation in the single 
bridge rectifier.  It would probably be prudent to include some sort of 
current limiting to protect against line surges, but since LED forward 
I-V curves are not very sharp even that might not be needed in some 
cases.  I would bet that there is more variation in individual LED 
brightness than there is in the forward voltage curves.

In my opinion, there are very good reasons why automobile tail lights 
and traffic signals almost universally use LED lamps now.

73,
Dave   AB7E



wayne burdick wrote:
> Hi Allen,
>
> The new LED bulbs are rated for 50000 hours, but this is a guess, 
> because they haven't burned out yet even in accellerated life tests :)
>
> Your older LED lamps probably failed because they were using a very 
> small number of LEDs, driven to very high current. Newer LED bulbs use 
> dozens of LEDs, each running at low current.
>
> 73,
> Wayne
>
> On Apr 3, 2009, at 2:40 PM, Allen Wisbey wrote:
>
>   
>> I hope you have good luck with longevity.
>> We replaced hundreds of exit light bulbs with LED types at my previous
>> job. They did not last long enough to pay for any energy savings.
>> Granted, that was 10 years ago, maybe technology has improved to a 
>> point
>> that they are a viable option....
>>     
>
>
> ---
>
> http://www.elecraft.com
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:Elecraft at mailman.qth.net
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>
>