Even small amounts of parasitistic oscillation cause a dramatic increase in distortion, this happens because self-correction due to negative feedback gets impaired during oscillation. Furthermore, transistors no longer work as linear devices during oscillation, they work as switching devices, think of a switching long tailed pair while only one of the devices is turned on at a time...
I suggest debugging the circuit until no oscillation is measurable in any circumstances, except maybe a bit of damped ringing when recovering from clipping. Zero-crossing is a very critical region of operation and should be perfectly clean.
Also note that an oscillating amplifier is a potential disturber for any audio circuits in the nearhood. The clumsy layout and wiring of most linear amplifiers is not optimized for dealing with RF oscillation components at all, sometimes they may produce far more EMI than SMPS or class-D.
I suggest debugging the circuit until no oscillation is measurable in any circumstances, except maybe a bit of damped ringing when recovering from clipping. Zero-crossing is a very critical region of operation and should be perfectly clean.
Also note that an oscillating amplifier is a potential disturber for any audio circuits in the nearhood. The clumsy layout and wiring of most linear amplifiers is not optimized for dealing with RF oscillation components at all, sometimes they may produce far more EMI than SMPS or class-D.
Eva thanks again for your information.
''Even small amounts of parasitistic oscillation cause a dramatic increase in distortion, this happens because self-correction due to negative feedback gets impaired during oscillation. Furthermore, transistors no longer work as linear devices during oscillation, they work as switching devices, think of a switching long tailed pair while only one of the devices is turned on at a time...''
This is a relief to hear, i was concerned with crossover distortion making itself visible at higher frequencies (above 10K).
This switching behaviour you mentioned sounds plausible as a cause for this as there was oscillation appearing on the output waveform.
I hope i can focus on elimintaing the oscillation and the crossover distortion will clear with it
''Even small amounts of parasitistic oscillation cause a dramatic increase in distortion, this happens because self-correction due to negative feedback gets impaired during oscillation. Furthermore, transistors no longer work as linear devices during oscillation, they work as switching devices, think of a switching long tailed pair while only one of the devices is turned on at a time...''
This is a relief to hear, i was concerned with crossover distortion making itself visible at higher frequencies (above 10K).
This switching behaviour you mentioned sounds plausible as a cause for this as there was oscillation appearing on the output waveform.
I hope i can focus on elimintaing the oscillation and the crossover distortion will clear with it
Since you mentioned changing transistors to slower ones helped, I imagine you could also have a 100 or 220 ohm resistor in the Base leads of your drivers, and put a small cap across the B-C leads of the drivers as well.
I used 100 ohm base resistors with 1nf caps across my driver B-C leads and any oscillation was gone after I did that.
Also try a larger zobel cap 0.22 uf
I used 100 ohm base resistors with 1nf caps across my driver B-C leads and any oscillation was gone after I did that.
Also try a larger zobel cap 0.22 uf
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.