[Amps] 12V IM3

[Manfred Mornhinweg]

Repeatedly some of you have mentioned the supposed advantage of running a device 
at very much lower output power than its saturation level, to improve linearity. 
This is true only to a point, not all the way down to zero! To get good 
linearity, an amplifier should of course not be driven into hard saturation, 
that much should be pretty obvious. But running it at a very low level, such as 
1/4 the maximum non-sat output, brings the distortion up again, due to crossover 
distortion, unless it's class A.

Another very important point is efficiency. The dissipation of an amplifier does 
NOT vary in direct proportion to the output level it's run at! An amplifier gets 
more efficient, the higher it is driven. A decent class AB amplifier might be 
60% efficient when driven just to the brink of clipping. A theoretically perfect 
class B amplifier would be 78% efficient at this level. But if the same 
amplifiers are driven to only one quarter that power level, the perfect class B 
one becomes only 39% efficient, and the practical class AB one just 30%. A 
consequence is that a practical 1500W output amplifier, designed to operate 
close to saturation, will dissipate 1000W (plus filaments, fans, bleeders, etc), 
while an amplifier designed to saturate at 6000W and run at 1500W will disipate 
3500W, plus the four times larger heater, bleeder, fan etc! That is if it uses a 
relatively low quiescent current. If instead that current is higher, the 
situation is even worse!

This makes clear that improving the IMD in the brute force way, by using an 
oversized amplifier, results in an extremely large, heavy and expensive 
amplifier, that's also expensive to operate. It's a very poor method for 
improving the IMD!

With more clever circuit design, such as negative feedback, bias modulation, and 
in extreme cases predistortion, a lot more IMD improvement can be obtained at 
far lower cost, and no added weight nor power consumption.

Keep in mind that most tube type ham linear amps aren't using any negative 
feedback at all, let alone any other linearizing techniques, and that even those 
solid state amplifiers that have a large gain excess rarely use it to provide 
more linearity by negative feedback, but instead burn off the excess drive power 
in an attenuator! Now that's a real waste. Something as simple as increasing the 
negative feedback to consume the excess gain of modern MOSFETs goes a long way 
in reducing IMD.

Manfred

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