Parasitic oscillation

[ACR]

I've done a thread search bit cannot find any clear ideas about the problem below.

I have an output triple emitter follower stage consisting of BF469/470 driving MJ15032/33 pair driving 5 per side MJL21193/94 pairs. In series with each output base I have 3.3 Ohms. However, I still have a parasitic oscillation problem at around 25MHz. Only way to cure it is to put a 100 Ohm resistor in series with the 15032/33 drivers - then the amp is stable, waveforms clean etc.

The 4069/470 are running in class A with 12mA emitter current and the 15032/33 runnung at 70mA. On both the pre-driver and driver stages the emitters are connected together with 1uF speed-up cap across the common emitter resistor (180 Ohm in pre-driver and 15 Ohms in the driver stage. The output device emitter resistors are 0.12 Ohm each.

Output has a 5uH inductor in parallel with a 2.2 Ohm resistor + 8 Ohm 0.1uF Zoebel network connected after the inductor


The amp uses feedforward compensation and I have used a 1nF/270Ohm input low pass filter to ensure clean square wave resonse with minimal overshoot. If I revert to standard compensation I still have the problem - so I do not think it is loop related, but parasitic.

Any Ideas?

[bocka]

First of all I would get rid of that crappy speed up caps. I've never had any success with these ones. They do nothing or make it worse. Your driver stage can deliver enough current without thise caps. Second get rid of the 3.3 Ohms base resistors in the output stage. Again I see no reason for using them. The intrinsic base resistance of the output trannies is about the same.

The BF469/BF470 are RF transistors. Some don't like any capacitive load when connected as emitter follower. Maybe Cbc is high enough that the predriver is oscillating. 100 Ohms in the base might be a little bit high as base stopper in the driver stage, I think 47 Ohms will also do the job. The smaller the better.

Connect the Zobel network before the inductor. This stabilises the output stage in many cases because it loads the EF follower on high frequencies.

Output emitter resistors of 0.12 Ohms seems quite low to me. For stability reasons I would not go lower than 0.22 ohms. Sometimes your amp rums thermically away after some hours or days.

Just my 2ct.