[SI-LIST] Re: Power filtering

  • From: Ray Anderson <raymond.anderson@xxxxxxx>
  • To: si-list <si-list@xxxxxxxxxxxxx>
  • Date: Thu, 27 Jun 2002 12:26:26 -0700

A couple quick considerations to get you started.

Pi filters (and even to some extent L filters) constructed
from either inductors or ferrite beads have some non-obvious
problems you need to consider.

The problem stems from the fact that the filter is
being driven from a very low impedance power source
(in the milliohm to 10's of milliohm range most 
likely). This causes the typical filter to have a
very peaky response instead of a nice low-pass
roll-off. 6 to 10 dB peaking isn't uncommon.
This means that whatever noise is on the input
of the filter may actually be 6 to 10 dB higher
at the output at the frequency of the filter
peaking. Insertion of a small amount of resistance
(say 1-3 ohms) in series with the inductor will
usually tame the effect. (beware of the DC drop
caused by the series resistance). Sometimes your
inductor will have enough intrinsic resistance
to make the filter well behaved. Ferrite beads
sometimes have enough loss to squelch the peaking,
but usually not (they have low DC resistance,
but have lots of AC loss, but usually not at the
frequencies the peaking taks place (in the kHz to 100's
of kHz range).

You need to model the filter using spice to do
numerous what-if cases to decide which way to go.
(this means you need decent spice models of your
inductor and/or ferrite bead including both DC and
AC loss as well as parasitics). Also note that
if you are interested in the high-frequency
bounce back characteristics you also need to
include the parasitic inductance of your capacitors
in the model as well (both cap inductance and 
mounting inductance.)

You could also try to synthesize an unequal termination
Butterworth filter (say Rin = 10 mOhms, Rload = 20 ohms)
but you will probably find you need an inductor of
several hundred mH and capacitors in near the Farad level
(not to practical, but has a nice roll-off when simulated :)
(the ratio of Rin to Rload  as well as the desired low
knee frequency is what kills you)

Enjoy!

-Ray Anderson
Sun Microsystems



Martin Euredjian wrote:
> 
> Not sure if this is OT for this list...
> 
> In isolating an analog (PLL, VCO, Amps, etc.) section from an otherwise
> high-speed digital board I've encountered recommendations to build the usual
> PI filter using an inductor (about 10uH) in one case and a ferrite bead (100
> to 300ohm) in the other.
> 
> What are the merits of each approach?
> 
> Thanks,
> 
> ===============================
>  Martin Euredjian
>   eCinema Systems, Inc.
> ===============================
.
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field

List archives are viewable at:     
                //www.freelists.org/archives/si-list
or at our remote archives:
                http://groups.yahoo.com/group/si-list/messages 
Old (prior to June 6, 2001) list archives are viewable at:
                http://www.qsl.net/wb6tpu
  

Other related posts: