[Amps] Price per Watt Conversation

[Manfred Mornhinweg]

Cathy,

> When we get to 3000V, we can have drop-in replacements for our tube
> amplifier finals! :-)

That would be fun... :-)

But not really practical. The high voltage needed for tubes is a 
nuisance rather than an advantage. And the low voltage needed by most 
transistors is also a nuisance. The golden spot lies in between.

Paul,

 > Yesterday, it was +50V, today it's 65V,
> tomorrow 75V, then 100V.? Where will it end? 

Impossible to predict.

> Where is the point of diminishing returns on operating voltage?

This is an interesting question.

My idea on that subject is that amplifiers should work at such a voltage 
that the circuit impedances are very roughly in the neighborhood of 50 
ohm, or a bit higher, but not much higher nor much lower. Because at 
very low impedances, such as are common with most transistors, stray 
inductance is a big problem, while at very high impedances, such as 
those used by tubes, stray capacitances are an equally big problem. You 
can't make a broadband high power VHF amplifier, neither with tubes nor 
transistors, because of this. You always need tuned networks to absorb 
and "tune out" the stray reactances.

With transistors at least you can make broadband HF high power 
amplifiers, although with some performance penalties relative to tuned 
designs. With tubes you can't even do that! Tubes are really suitable 
for broadband operation only on pretty low frequencies - MF, LF or VLF 
depending on the tube.

1500 watt on 50 ohm is 274 volt RMS. A supply voltage of roughly 400V is 
optimal to run this power in a nice, simple, single-ended amplifier. 
Half that voltage would be OK for push-pull amps, but that requires more 
complex circuitry. Something higher than 400V is OK too, say 500 or even 
600V. Higher than that is too much.

Such amplifiers would most likely need to be tuned, just like tube 
amplifiers. For two reasons: Transistor capacitances will probably be 
too high to allow broadband operation at those voltages, and broadband 
transformers for high voltage get hard to make, as they require a 
proportionally larger product of turns number and ferrite cross-section, 
and this creates phasing problems due to conductor length.

Three basic concepts for tuned solid-state amps come to mind:

1) Bandswitched tuned networkd instead of bandswitched low pass filters;

2) Manually tuned tanks like in most tube amps;

3) Automatically tuned tanks like in some tube amps.

At present there are some high voltage RF power MOSFETs available, but 
not many, and they are all VDMOSFETs rather than LDMOSFETs. That means 
higher capacitances.

Some developments around silicon carbide MOSFETs and gallium nitride 
MOSFETs look interesting.

Manfred

========================
Visit my hobby homepage!
http://ludens.cl
========================