Andy Thanks for the work in the PDF file I will read it fully. There is something you are saying that I see fits within the bandwidth versus gain concepts.
Any active device has a certain gain factor, and so there is the power dissipation limit and the useful power output. E * I = Watts Now knowing there is a limit to the gain, if we have more than one signal the power is thus spread out over the bandwidth of the signals. So the power gain is spread out over the bandwidth. Two strong signals inputs together should in theory be of less power output individually, than just one where all the power was expended on one signal. Now add in a third signal, the sum of the power gain of all of the signals should equal the input power. And so, one signal is thus of less power output than the total device gain. And in theory the power is suppose to be divided up between all the incoming signals being amplified. If we got 20 dB output for every signal across the bandwidth for a device rated at a max 20 dB gain that would mean over unity output. Which is impossible to obtain. The only way you can approach the true gain is via a single signal input or via a very selective pre selector ahead of the stage. Or your device is rated for s single signal gain of 20 dB over a certain bandwidth. Three tones then divide up the power output. Where the third tone might create a louder mixing product is with the agc on. Pulling the signals all down. At least that appears that way and I could be wrong on the agc. And this is just a theory. Dan
