On 5/1/2012 10:59 AM, Paul Christensen wrote:
>> "I doubt that ... the solid state devices would not appreciate the die
> temperatures with vapor phase cooling. However, a small radiator and
> closed water system would work very well. With a suitable pump the
> radiator/reservoir/fan could be located outside or in another room for
> essentially zero noise."
>
> Joe, good point. It's an interesting thought exercise...
>
> I looked up the maximum operating temperature for the Motorola MRF150 and
> Microsemi ARF1500. From the datasheets, maximum operating temperature is
> between 175-200 degrees C which is> 75 degrees C above the H2O boiling
> point. As I recall, water can never exceed the boiling point temperature
> under normal atmospheric pressure. Additional heat does not raise water
> temperature but causes steam and steam too never increases beyond the
> boiling temperature at normal atmospheric pressures. But under pressure,
> it's a completely different ball game. Added pressure raises the boiling
> point and the temperature of steam can get extremely high.
>
> Assuming a set of transistors mounted onto a dense copper block (copper, to
> avoid water contamination) and immersed in distilled water, it seems to me
> that these transistors could operate well under their maximum ratings with a
> lot of room to spare. Worst case is the temperature of the transistor
> substrate could never exceed 100 degrees C,
Unfortunately the "substrate" can exceed the temperature of the medium
in which the device is submerged by quite a bit.
"Substrate" normally refers to the Silicon (or other material) on which
the device is built. The heat has to not only go through the Silicon
substrate, but through the device case and leads as well. Internally
the device is only a fraction the size of the case which at maximum
power leads to a substantial delta T internally. Regardless of the
method of cooling (as long as it's sufficient) the internal delta T (or
thermal resistance) becomes the limiting factor for power. For high
power SS use with water cooling the water is often cooled. Most
computers will shut down well before 100C at the CPU and that is
measured internally.
There are a number of transistors and devices consisting of multiple
transistors with support on board rated for a KW or even more, but when
you read the data sheet that is for pulse. Even for SSB which is usually
figured at a 20% duty cycle these devices are derated substantially and
may only run a couple hundred watts output. In computers they may
actually use refrigerant and in some cases the CPU is submerged in
Liquid Nitrogen. In others the evaporator for the refrigerant is
integrated into the heatsink. For the home user, refrigerant comes with
some risks like condensation (causes arcing and corrosion) and noise so
the cooled devices must be kept in a very low humidity environment while
the mechanics are usually in a separate room and the heat is disposed of
outdoors. Kinda like the Kalifornia KW in a separate room from the
operating position.
They are improving, but operating power transistors any where near their
maximum temperature and/or voltage ratings may severely shorten their
lives. Contrary to popular belief, transistors do age and do not last
forever. As the temperature goes up the doping materials tend to
migrate across junctions as well as into gates. The latest transistors
are no where near as susceptible to this as bipolar devices but they
still age and the higher the temp the faster they age. Still the
degradation is noticeable at temps as low as 100C.
Some where on the net I read a paper on that a while back. A search on
"transistor degradation at high temperatures" will produce many papers
and abstracts. Unfortunately some of the good ones are available only
if you purchase the download.
73
Roger (K8RI)
>
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|