Ferrites in supply rails?

Hi!

I'm considering to add small smd ferrites (0805)  in series with the supply lines to the opamps stages with a gain greater than 1 like fader amplifiers. The opamps already have 10nf at each supply pin to gnd. Maybe the resistive part of the ferrite can help to filter out ripple  and noise a bit more...

Regarding the psu dc output (LM317/337 linear type), i already have 100uf in parallel with 100nf... but maybe adding a ferrite in between the output caps or before the 100uf may help reducing output ripple  a bit more??

http://cds.linear.com/docs/en/application-note/an101f.pdf 

Thanks for your advise!

Jay X

My understanding is that ferrite beads help suppressing/filtering out radio frequency interference, nothing to do with power supply ripple...

Hi!

Ok... so, another possibility is to make an Rc filter at the output of the psu. 10R + 220uf = 72,4hz

There would be a voltage drop: for 150ma load current:  1.5v.  But with variable voltage regulators this is manageable.

JAY X said:

Hi!

Ok... so, another possibility is to make an Rc filter at the output of the psu. 10R + 220uf = 72,4hz

There would be a voltage drop: for 150ma load current:  1.5v.  But with variable voltage regulators this is manageable.

Click to expand...

You really should not need this. If your dc supply comes from a linear regulator, the ripple will already be very small. Most op amps reject power supply ripple by a further 100dB. So, as I say, there really should be no need for this. If you are having noise, hum problems then the more likely cause is poor layout or grounding.

Cheers

Ian

> http://cds.linear.com/docs/en/application-note/an101f.pdf 

Fig 2 - RIPPLE: TYPICALLY 100kHz to 3MHz
SWITCHING SPIKES: HARMONIC CONTENT APPROACHING 100MHz

Are you troubled with MHz ripple??

Or power-line harmonics?

Define your problem before throwing solutions at it.

Though, really, for the few cents a "ferrite" might cost, you could just try it, not even ask.

Might help clean up HF noise from switching regulator, but generally not necessary .

JR

JAY X said:

Hi!

Ok... so, another possibility is to make an Rc filter at the output of the psu. 10R + 220uf = 72,4hz

There would be a voltage drop: for 150ma load current:  1.5v.  But with variable voltage regulators this is manageable.

Click to expand...

Many mixers use such a principle: the PSU delivers about +/-18V, which are distributed to all the modules, and there are resistors (typically 10 - 47ohms) that distribute them to the active stages.
The main reason is NOT to filter noise, mainly for rail protection in case of short-circuit and fault-tracing.
I've used this principle in many designs, never had to regret it.