Doable.? And I'm not laughing.? ;)
That might eclipse the hack of 30++ years ago.? 2n2219 in a block of aluminum with feedthoughs
and the cavity filled with transformer oil.? Got the thing to put out 10W with 30% duty cycle?
at 40mhz with that block anchored to a 6x6x2 inch heatsink.??
Its a matter of getting the heat out.? water works well.? People used to take 3CX100 (2c39)
and pull the air radiator and solder on a cap with hoses, distilled water is a good insulator
and with about 1000V you can push that poor 100W tube to nearly past 500W at UHF!?
For Bitx the real problem is the IRF510 thermal interface.? Once the die is dissipating
about 43W you have maxed out as you can't no matter what get the heat out due to
the TO220 thermal resistance 3.5 degrees C/watt so at 43W its going to be at
150.5C die temp.? ?At 150C it reverts back to sand.? Cooling the epoxy side is
ineffective.? Now if IR put that die on a flange with a heavy bond in(gate) and
out(drain) it has a chance.? ? On the bright side two push-pull would be doing
about 90W out (assuming 50% efficiency, doable at 40M).? My own testing got
to 85 when they expired.? ? However any effort to keep them cool is still a good thing.
RD16HHF is only hair better at 2.2 degrees C/W.? However there are other limitng factors.
If there is a quest for power then devices that can get rid of the heat are a must.
Selected properly and run conservatively they can be near indestructible.? Problem
is they are not cheap.
Allison
