[Amps] IM distortion and such

[Michael Tope]

----- Original Message ----- 
From: "Peter Chadwick" <g3rzp at g3rzp.wanadoo.co.uk>



> Tom said:
>>Even a curved line does not necessarily produce
> IM3 or IM5. (harmful IM3 is 2 * F1 - F2,  or 2*F2-F1, IM5 is
> 3*F1-2*F2, 3*F2-2*F1 and so on through all odd-order
> products).<
>
> If it's a square law curve, for example, there are NO third order products 
> (at least from that stage). Then again, some curvature can lead to 
> cancellation and an improvement in IMD at some power level.
> The data sheet absolute maxima are the ones that really count - exceed 
> those, and you can have problems getting warranties honoured. (Cynicism 
> suggests getting warranties honoured can be difficult whatever you do...)
>

Yes, definitely, Peter. I recall an IEEE talk on Gallium Nitride
amplifier technology that I attended about 6 months ago where the
presenter showed several IMD curves for class B amplifiers that
had high level "sweet spots" where the positive slope of IMD level
as a function of input power actually reversed and formed a local
minimum at a fairly high input power level.

BTW, TWTA linearizers actually exploit this phenomenon to reduce
overall IMD. Sometimes two wrongs do make a right :)

73, Mike W4EF