This can occur due to loading effects on the amplifier, and not just due to parasitic feedback as described above.
Or it could be due to an amplifier that was not designed with enough "stability" in the first place.
For any amplifier using feedback (and most do), if there is not enough "margin" built into the design, they can oscillate at extremely high frequencies, often a few megahertz. This high frequency oscillation can create beat components down into the infrasonic (50-100KHz) range that could fry speakers depending on the magnitude and power levels of the oscillation, and depending on the speaker itself.
In some power amplifiers, the use of high capacitance interconnects can trigger unstable oscillation. Likewise, if a speaker system has too much capacitance, it can also be triggered. Note this is capacitance in parallel with the load (or speaker), not coupling capacitors in cross over networks.
The capacitance loading forms another response pole with the output impedance of the amplifier. If this pole (or corner frequency) is within the closed loop bandwidth of the power amp, instability will result and the amp will oscillate.
Note that defective or failed devices in the output stages of the power amp (such as a weak power transistor) can increase the output impedance of the power amp and trigger osillation that was not there before.
Oscillation in the pre-amp is also possible, but it should not be getting through the power amp as most have RF filtering on the input stages that should prevent this from getting to the power stages. Of course, if the power amp is broadband and amplifies the RF, then it will be transfered to the speakers.
Best way to detect this oscillation is with the old fashioned oscilliscope. Find a tech who has one and look at the power amp outputs with the pre-amp disconnected. Then hook up the pre-amp and check again. Then add the speaker cables and keep checking.
If RF oscillation is there at high enough levels to burn tweeters, it should be easy to see on any decent scope.
Or it could be due to an amplifier that was not designed with enough "stability" in the first place.
For any amplifier using feedback (and most do), if there is not enough "margin" built into the design, they can oscillate at extremely high frequencies, often a few megahertz. This high frequency oscillation can create beat components down into the infrasonic (50-100KHz) range that could fry speakers depending on the magnitude and power levels of the oscillation, and depending on the speaker itself.
In some power amplifiers, the use of high capacitance interconnects can trigger unstable oscillation. Likewise, if a speaker system has too much capacitance, it can also be triggered. Note this is capacitance in parallel with the load (or speaker), not coupling capacitors in cross over networks.
The capacitance loading forms another response pole with the output impedance of the amplifier. If this pole (or corner frequency) is within the closed loop bandwidth of the power amp, instability will result and the amp will oscillate.
Note that defective or failed devices in the output stages of the power amp (such as a weak power transistor) can increase the output impedance of the power amp and trigger osillation that was not there before.
Oscillation in the pre-amp is also possible, but it should not be getting through the power amp as most have RF filtering on the input stages that should prevent this from getting to the power stages. Of course, if the power amp is broadband and amplifies the RF, then it will be transfered to the speakers.
Best way to detect this oscillation is with the old fashioned oscilliscope. Find a tech who has one and look at the power amp outputs with the pre-amp disconnected. Then hook up the pre-amp and check again. Then add the speaker cables and keep checking.
If RF oscillation is there at high enough levels to burn tweeters, it should be easy to see on any decent scope.