Quote Originally Posted by Resident Loser
...not buyin' it...these thingies are connected to the power amp's speaker outputs and while their effect may be audible, I can't see how they would have any effect on the internal feedback loops which are located far upstream (and fairly isolated) from these devices and their connection point.

Add to that, that each amp's design tend to have differing amounts of feedback and these things seem to be a one-size-fits-all affair, not really optimized to reflect such differences.

jimHJJ(...as markw suggested Zobel network anyone?...)
In most designs, the output stage of an amplifier is the most in need of feedback correction so the signal from the last stage is usually included in the feedback mix. Feedback is a broad term, global feedback normally means from the last point on the output to the first point in the input, it is both the best way to apply feedback and worst way because it is easy to introduce unplanned results an thereby get it wrong (read up on slewing induced distrortion and transient induced distortion).

So global feedback is traditionally taken from the output signal as far downstream as possible and then applied to the input stage as far upstream as possible. This means the actual speaker leads in the case of a power amplifier. In the case of a tube amp, picking feedback from the actual speaker leads will include the output transformer inside the feedback loop, in the case of a transistor amp the signal from this point is often used for both feedback and amplifier output stage protection. Putting a complex reactance (capacitor and inductor network?) to ground at this point can change the amplifier feedback or protection signal.

Again, I think the concept is probably a bad idea. Designers may have made any number of decisions about how to shape the feedback signal (that signal is usually shaped by some kind of network) and second guessing the designer seems unwise.

If the design is poorly executed so that these things help, I stand by my original comment - buy an amplifier that is not poorly executed.

As to Zobel networks, all of them have a corner frequency, traditionally that corner is the frequency at which the voice coil inductance starts to significantly increase a given driver's impedance. Zobel networks allow previous crossover stages to work into a fixed load (4 or 8 ohms for example). Putting a Zobel ahead of the crossover accomplishes little,. Plus, is that Zobel designed for the tweeter, the mid range or the woofer? A single compromise network is very unlikely to actually work.