Years ago in the process of some power amplifier design work we were able to identify low level distortion in a 3AG 5 amp quick blow speaker fuse as used in a Hafler DH-220 amplifier.
We had designed new audio circuit boards for this amplifier and were evaluating one "feature" of the original DH-220 output wiring layout that was unusual. Hafler had placed the speaker and output circuit protection fuses inside the overall amplifier feedback loop.
We could evaluate that design idea by driving the amplifier to near full power with 20Hz sine waves. We placed our scope test probes not at the audio outputs, but at the output of the voltage amplifier section driving the current amplifier section. This way we could observe the correction signal the feedback was applying to the output signal. Sure enough the correction signal showed a tiny 20 Hz sine wave correction, which was of course caused by the change in thermal resistance of the fuse as is alternately heated and cooled depending upon the power from the input signal across it.
Inasmuch as our general design goal is to make the feedback loop do as little work as possible, we then redid the output wiring to take the fuse out of the loop, thus eliminating the correction signal it demanded. This ended up with a bit cleaner and more refined deep bass response in the amplifier. The tiny bit of thermal distortion in the fuse was less harmful then trying to make the feedback loop correct for it unnecessarily.
Perhaps a ceramic output fuse would have lower thermal distortion, but since they are not available in the quick blow fast breaking type necessary to protect the speakers and amplifier output circuits, we do not use them for this design purpose.
Frank Van Alstine