Reducing full wave bridge rectifier ripple in preamplifier

Question

It is a fairly simple question with similar answers, but I need help with my project.

I built a single supply 30W power amplifier and a preamplifier with tone control. Both amplifiers are powered from a simple AC-DC converter.

It all works fine when I use a bench power supply, but when I make my own AC-DC converter the problems start.

My simple AC-DC converter contains a bridge rectifier, 3x 3300uF capacitor and 1x 100nF capacitor (the picture is just for example and is not mine):

The problem is that the voltage ripple is transferred through the preamplifier into the output signal and produces noise. The power amplifier does not pick any noise from the power supply voltage ripple and works perfectly (I connected an audio source directly to the power amplifier and it works with no noise.)

I tried putting an LM317 voltage regulator between the AC-DC converter and preamplifier. There is a little improvement but it still contains a lot of ripple voltage noise (100mV.) Adding more capacitance to the AC-DC converter or at the output of the LM317 slightly reduced the noise. If I want to significantly reduce voltage ripple with capacitors, I would need a few 10000 uF capacitors and this is not cost effective for me.

How can I easily reduce the noise on the preamplifier power supply? Why is the TDA7377 power amplifier with a few extra komponents not sensitive to voltage ripple?

Answer 1

My simple AC-DC converter contains bridge rectifier, 3x 3300uF capacitor and 1x 100nF

Total overkill. Your TL072 can under no circumstances even draw enough current to load the 3330 µF significantly. Very likely, using 1 µF with lower ESR and lower leakage would have been better. More isn't always better!

I would need few 10 000 uF capacitors

Not true. Sure, more capacitance reduces the ripple, but I'll leave the calculation of how long a TL072 will need at maximum supply current to discharge a single 3300 µF capacitor charged to 10 V by 100 mV up to you, pointing out that you should compare that time with the duration of half a net period.

I try to put LM317 voltage regulator between AC-DC converter and preamplifier and there is little improvement but still contains a lot of ripple voltage noise (100mV).

Good approach. But: The LM317 is really an ancient piece of regulator. Maybe use something that needs less voltage difference and has a better noise suppression.

(also, capacitors are not optional, and every regulator datasheet will tell you what in- and output capacitors are needed.)

The Supply voltage rejection of the TL072 is >-70 dB typically. I can think of many ways to insert more than a 10 millionth of the supply ripple into your output signal, so make sure you're not doing something like using an unregulated voltage as virtual ground or similar.