Simple - I used to design and test semiconductor material -
I started in the mid 70's and went pro in 1980.

I remember some of the first high speed CMOS HMOS and NMOS
that ran really fast but shorted power to ground in the output
transistors in the process. It seemed ok in theory - but after
millions of cycles the part melted the epoxy off the IC !

All semiconductor heats up. It is conducting current. Current flow
causes vibration heating.

Martin

On 11/19/2016 8:14 PM, DerbyDad03 wrote:
On Saturday, November 19, 2016 at 5:56:49 PM UTC-5, krw wrote:
On Sat, 19 Nov 2016 09:47:06 -0600, Leon lcb11211@swbelldotnet
wrote:

On 11/19/2016 9:33 AM, Brewster wrote:
On 11/13/16 8:43 PM, krw wrote:


It's not about the tube getting hot. It's about the semiconductor
junction heating. Without a *really* good heatsink, they will
overheat. Most are *way* under designed.



An old rule for electronics is 10 sq. inches of heat sink surface
(exposed to room temp) for every Watt of power. Many of the "flood"
style bulbs which don't have the room for this get by with thermal mass.
They don't saturate the heat sink for 1-2 hours, giving you time to get
your lighting task complete before damage begins. This is the bane of
the Edison style bulb for LEDs. Tubes have the surface area going for
them, but the cheaper tubes find a way to skimp on the aluminum to
mitigate any advantages 8^)

-BR



There seems to be a lot of confusion concerning LED's and their heat out
put. I have LED lamps that replaced incandescence bulbs and the light
is cool but the part that converts to the proper voltage generates heat.

I also have probably 20' of high density ribbon LED lights that run on
12 volt and they do not even get warm.

You can bet that the semiconductor junctions are getting warm. The
fact that the surface isn't may also mean that the semiconductors
inside are getting downright toasty. My thermos mug doesn't get warm
on the outside, even when filled with very hot coffee.

How do it know?