Cleaning up PC PSU 12 v rail?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello all,

I have built pair of little amps built around TDA2003 chips that run nicely on 12 DC single ended supply. I'd like to them mount inside a PC and power them off the PC's power supply.

The PC psu voltage turns out to be very dirty! I can hear windows being resized, hard drive accesses, the mouse moving.

I understand electronics well enough to read technical writing... but it's been years since I built anything. Can anyone point me in the direction of information about passive filtering quite dirty power?

T.I.A.!
 
That circuit is only useful to get rid of regulation noise and ripple coming either from linear regulators, or from the error amplifiers in SMPS. However, note that the inherent noise of the op-amps, comparators and clocks employed for SMPS control gets strongly reduced at high frequencies thanks to the output filter (actually it can be happily modelled as a linear regulator with a LC filter inside the output loop). Also 100Hz/120Hz ripple in SMPS can be attenuated by several orders of magnitude (as in linear regulators or even better) with proper loop control and compensation (something not always found in PC PSUs).

Intuitively, the classic voltage mode PC SMPS can be modelled as an op-amp whose inputs are fed from a 5V reference and from a fraction of the output voltage of the SMPS, and whose output voltage continuously decides which percentage of the rectified mains voltage is being applied to the output inductor (previously scaled by a fixed ratio in the transformer).

However, in switching circuits there are more subtle ripple components at the output, particularly common mode ripple and diode recovery glitches, that require ferrites for filtering. I don't like to call that "noise" because it's obviously not noise, SMPS output ripple is a complex PERIODIC waveform made of many periodic COMPONENTS, each of which comes from a particular switching or resonance phenomena happening synchronously in the circuit. Indeed, detailed analysis of SMPS output ripple waveforms can suggest many ways to improve the circuit.

Remember that the output ripple of an ideal SMPS is either a triangle, a trapezoidal or a sine wave at the clock frequency (depending on topology), which are quite harmless waveforms until they become full of glitches and ringing due to a poor or too compromised design.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.