About strange noise from my amp!

[Otherwise]

As someone might remember, I have built this class A amp. Some pics in this thread.

I have happily been listening it and now I noticed that:

When I turn the right ch. volume knob, a strange buzz sound (quite loud) appears in two different places. First at 2 o'clock and then at 5 o'clock. The pot goes from 9 to 6 clockwise. I first thought it just could be some crap got in the pot, but when I turned the left ch.pot, the sound disappeared It appears again but at different spot when I turn the right ch. pot.

Is my amp oscillating or does it think that it's a receiver I have tried to disconnect the power led/fan and the water pump from the main switch, but the buzzing sound is still there..

 

[bocka]

Seems to me that the amp oscillates. Try to connect the 220pF cap directly to the output of the opamp. Use a rf filter at the input, about 2k2 and 470pF.

 

[peranders]

I think you have a small problem with speed. You could try feedback directly from the output of the opamp and let the output stage go without feedback at higher frequencies. You can do it in two ways:

1 Remove the 220 pF, place a capacitor from the output of the opamp to the inverting input. Value of the capacitors must be calculated but 220 pF doesn't seem to wrong.

2 Remove the 220 pF, place a capacitor plus a resistor in series from the output of the opamp to the inverting input. Value of the capacitors must be calculated but 220 pF doesn't seem to wrong. The resistort value shall be half or little less of 15 kohms (in your case)

You must check the frequency response of the output stage, much easier to give the right medicine. 1 komhs gate resistors seems much, try to decrease the but check for oscillations.

One tip is also to choose MOSFET's for the application. The fet's should be NOT too powerfull. Smaller fet's have smaller capacitances = better.

 

[Otherwise]

Ok, thanks. Did a little experiment and took the 220p cap away and connected it between opamps pins 1 and 2. Sound disappeared

I don't have an oscilloscope (I really should if I'm going to do these amps more..) but could something be seen just by measuring ac/dc voltages somewhere..? Almost only thing measured is the current the amp takes, which is quite precisely 2 amps. (schematic says 1A for mono).

I know what it means the amp is oscillating, but what does "speed problem" mean? And what can it cause if the feedback cap is a little too big or small? Does it affect freq response or distortion much?

 

[bocka]

2A current drawn is too much. It should be about 1A to 1.2A. Maybe the current source is oscillating. Try a 1nF cap between base and collector of the BC547.

BTW, you are not able to measure rf oscillation only with a simple dc or ac measurement. Use at least an analog scope. ebay is not a bad place to get a one...

 

[peranders]

I know what it means the amp is oscillating, but what does "speed problem" mean? And what can it cause if the feedback cap is a little too big or small? Does it affect freq response or distortion much?

With speed I mean that you can't connect several stages and get it stable. One stage must be significant slower, the opamp in your case. Other stages (your output stage) must be much faster than the opamp, otherwise you have an oscillation!

Your output stage must have at least 500 kHz to 1 MHz bandwidth (not calculated the value, just an estimation) in order to be safe. 500 kHz from this stage needs care in designing.

One trick to use a slow output stage is to exclude it from feedback at higher frequencies, this cap thing we mentioned do that.

 

[Otherwise]

2A current drawn is too much. It should be about 1A to 1.2A. Maybe the current source is oscillating. Try a 1nF cap between base and collector of the BC547.

I could try it.. But I somehow thought that if it takes 1A at mono, stereo would take 2A??

Anyway, this is quite "experimental" project, so if the amp isn't working so well it could, it doesn't matter much. I chose the schema for its simplicity and easy-to-get components. Also I wanted to try p2p soldering.