I want to clarify this question with my real-world example:

Say you have a device which contains some AC-DC converter (or DC-DC) for rather high power specs (10 kW up to 300 kW, not in the same device tho, just to curb the question parameters) and due to some external constraint know that the device will never need to be as-small-as-possible in two of the three dimensions. Smallest envisioned dimension is ~50x50 cm but can be as large as 1m x 2m, acceptable height is I think determined by the necessary active cooling technology, in the 10kW case this shouldn't surpass 10-20cm though. With one word, the device may be large but as flat as possible.

Is there some (any) advantage in sticking to old semiconductor technologies which dropped out of present day electronics due to their inability to miniaturize or other parameters along that line? I am thinking of thyristors (if this is a correct example - I am just a Software guy) which were all over the place in the 60's. Are there past technologies which lend themselves to AC-DC or DC-DC switching with benefits in

  • price
  • minimzation of losses
  • ease of cooling

which are large enough to consider them against present day technologies if above described size is of secondary concern?

Similarly, is there a price/power advantage if one builds the necessary modern semiconductors along these constraints - i.e. their housing may extend more in 2D than usually required? Is there something to be won?

  • 2
    \$\begingroup\$ I seem to recall learning that until relatively recently (70s to 80s), large rectifiers still used mercury arc tubes. And though it's not a power electronic converter (unless you see it as some kind of high-frequency inverter?) some RF equipment still uses vacuum tubes because nothing else will do; the magnetron in your microwave oven for instance. \$\endgroup\$ – Hearth Jul 6 '20 at 15:51

The main advantage of using IGBTs and MOSFETs over thyristors is not their smaller size (I honestly don't know if they even are smaller for the same power-handling capacity) but their higher efficiency and greater controllability. Thyristors have always had a flaw that they are very difficult to turn off; it requires waiting for the next mains zero crossing or, in the case of a GTO thyristor, shunting very large amounts of current out the thyristor's gate through your control circuitry.

IGBTs and MOSFETs don't have this problem; either type of device can be turned off by simply driving the base with the same type of control circuitry used to turn them on. MOSFETs in particular are also much more efficient than thyristors, since they act almost as resistors when turned on, rather than having a relatively large (though relatively fixed) voltage drop. (IGBTs still have their saturation voltage, though. I don't remember off-hand how this compares to the voltage drop of a thyristor.)

On the other hand, sometimes you do need higher voltage withstanding and current capacity than you can reasonably get with solid-state components. Until relatively recently--into the 70s and 80s, I believe--mercury arc rectifiers, a type of electron tube, were still used in some high-power applications such as radio transmitters. The development of higher-quality solid-state devices stopped that, though, as silicon diodes and thyristors are much cheaper, and don't contain environmentally-hazardous mercury metal.

It's not a power electronic converter, but there are still vacuum tubes in common use today as well. The magnetron in your microwave oven is in fact a vacuum tube, because there is currently no economical way to generate that amount of RF power at the appropriate frequencies with solid-state devices. That may change in the future, but for now, you still have at least one vacuum-tube-based device in your house--as long as you're one of the large proportion of people who have a microwave oven, anyway.

Klystrons and travelling-wave tubes, other types of vacuum tube that are actually fairly similar to the magnetron, are still in common use as well for amplifying RF signals, especially in radio and TV station transmitters, as well as on communications satellites and in particle accelerators, where unusual requirements lead to them being the best choice in some cases.


Satellites in orbit use HeatPipes to move heat from interior to the exterior, to dump heat into the universe. They have wicking lining the pipe's interior wall, for the liquid to return to the interior. The vaporized gas moves thru the center of the HeatPipe, from the hot electronics ("hot" being relative) to the cold of the craft's outer wall.

Other than heat pipes, you are limited by the thermal conductivity of copper, etc.

A standard thickness of copper foil --- 1.4 mils, 35 microns --- is so thin (its a delicate foil) that 1 watt flowing orthogonal to the thin dimension will cause 70 ° C temperature drop across any square of foil.

Thus a stripe of foil, 1cm wide and 5cm long, with heat flowing along the 5cm direction, will have [5cm/1cm == 5 squares] * 70 ° C == 350 ° C drop.

However this is only 1/700 of an inch thick. A 1 inch bar of copper could move that watt for 350/700 = 1/2 ° C.

And 100 watts would be only 50 ° C.

Or you can use blasts of moving air.

If you happen to have CO2, or N2, in liquid form, those gases may be your friend. The opposite of a heat pipe.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.