[alancalverd]

A feedback loop absolutely does exhibit homeostasis, it's why NFB reduces noise, distortion and gain variation in an amplifier, why a voltage regulator regulates, and why a PLL remains locked in the presence of a disturbance.

I can't remember the detail of what Maddox had to say now after 15 years, but he'd given the matter a lot more thought than you seem to have.

So your amplifier has no output? What is the point of it? Mine have loads of NFB so that the output is a function of the input, and my PLLs lock onto the chosen input. If the NFB loop is nonlinear, say an integrator or differentiator circuit, the output depends on the history, rate of change, peak, or whatever function you have chosen, of the input - it is by no means constant.

It is true that the classic operational amplifier analysis depends on maintaining a null at the summing junction, but the system as a whole is not homeostatic.    I thought about little else for some 20 years, and still like to use analog systems for fun. I go back to the days of Ragazzini and Philbrick designs! 

I guess any steamhead will point out that the Watt governor maintains constant speed against a variable load, so is a true homeostatic device, but a voltage stabiliser does so by comparing the output with a constant reference input, so isn't actually a homeostat so much as an opamp with a fixed input!   

[hamdani yusuf]

I guess any steamhead will point out that the Watt governor maintains constant speed against a variable load, so is a true homeostatic device, but a voltage stabiliser does so by comparing the output with a constant reference input, so isn't actually a homeostat so much as an opamp with a fixed input!   

Don't you think that Watt governor also has a constant reference input, which is desired engine speed?
Voltage stabiliser also has to work against variable load, what makes it different?