RE: [SI-LIST] : Ferrites on power leads

About this list Date view Thread view Subject view Author view

From: Chan, Michael (Michael.Chan@compaq.com)
Date: Tue Aug 01 2000 - 12:28:53 PDT


Let clarify where the noise orignate from. If the noise is generated out
from the IC ( through the package inductance ), then will a well desinged
power distribution
system ( assume on the PCB board ) ( planes < 5mil separation, etc ) help
and do we still need a ferrite ? If the package inductance is so low that
there is minimum amount of noise generate out from there, will we need a
ferrite or planes with less than 5 mil separation?
 
 
Regards,
Michael Chan

-----Original Message-----
From: Mayer, Mike [mailto:mikem@artesyncp.com]
Sent: Tuesday, August 01, 2000 1:51 PM
To: 'Vinu Arumugham'; e
Cc: Gaboian, Jerry; Mark Gill; Mayer, Mike; 'si-list@silab.eng.sun.com'
Subject: RE: [SI-LIST] : Ferrites on power leads

Howard Johnson's latest EDN column addresses the keeping noise out of
sensitive inputs aspect of this. It doesn't appear to have hit the EDN web
site yet, but it's in the August 3, 2000 issue.
 
The advice given in a seminar was for the opposite case: that the filtering
was to keep noise in, not keep noise out. Most of the discussion here has
been about the "keep noise out of the chip" case. With the low impedance of
a well designed power distribution system (planes < 10mil separation, etc.),
does a ferrite really help keep noise off of the planes?

============================================================================
=
Mike Mayer Artesyn Communication Products, Inc
Senior Hardware Design Engineer Madison, WI
mikem@artesyncp.com http://www.artesyncp.com
<http://www.artesyncp.com/>
============================================================================
=

-----Original Message-----
From: Vinu Arumugham [mailto:vinu@cisco.com]
Sent: Tuesday, August 01, 2000 1:11 PM
To: e
Cc: Gaboian, Jerry; Mark Gill; 'Mayer, Mike'; 'si-list@silab.eng.sun.com'
Subject: Re: [SI-LIST] : Ferrites on power leads

Ellis,

Noise on the power pin of a device usually perturbs the output signal. Such
a perturbation may be significant, especially if the output is a clock
signal. In this context, if radiated noise impinging on traces is a concern,
my point was that the clock quality is affected more by the noise picked up
by the clock trace , than by the noise picked up by the power trace of the
clock driver.

Thanks,
Vinu

e wrote:

Vinu,

As far as antennae go, no doubt the 5" long output trace is more susceptible
to noise (of a wider frequency spectrum). I believe the discussion is about
noise on power supplies, more specifically how to keep that noise out of the
power supply, and in that sense maybe you can clarify for me how the noise
picked up by the output trace can be a bigger problem to the power supply
pins than the noise on the power leads/traces themselves. I don't really
see your point. Please educate me.

Ellis

Vinu Arumugham wrote:

If a 5 mil wide, 500 mil long power trace of an oscillator picks up enough
radiated noise to disrupt the system, I would expect the 5 in. long clock
trace at the output of the oscillator to be a bigger problem.

Vinu
  

e wrote:

Vinu,

Long narrow traces are also very good antennae(s?) for picking up radiated
noise, which defeats the purpose of the bead.

Ellis

"Gaboian, Jerry" wrote:

Vinu,The long narrow traces are inductive as you would expect. A problem
with this is when you start adding high frequency bypass capacitors on the
trace, the inductance of the trace can cancel the properties of the
capacitor.Regards,Jerry -----Original Message-----
From: Vinu Arumugham [ mailto:vinu@cisco.com <mailto:vinu@cisco.com> ]
Sent: Friday, July 28, 2000 11:17 AM
To: e
Cc: Mark Gill; 'Mayer, Mike'; 'si-list@silab.eng.sun.com'
Subject: Re: [SI-LIST] : Ferrites on power leads
  

Why is a long narrow trace in series with a bead a poor filter? I expect
long narrow traces, being more inductive, to help block high frequency
components.

Vinu

e wrote:

Mike,

In my designs where I use ferrites for power supply filtering, I always
include caps on both sides of the ferrites. Also, the filtered side is
always a generously wide trace, if not a an isolated "island".

I agree with Mark that an improperly executed ferrite filter actually
increases noise. For example, a bead between a plane and a power pin would
probably not provide much benefit if the power pin were connected to the
bead with a long, narrow trace. In such cases, it would probably be better
to connect the power pin directly to the power plane and then use a high
frequency cap next to the pin to help filter noise.

The impedance you are adding is meant to block high frequency noise, not the
desired DC power levels. Generally this technique is useful for
noise-sensitive designs, like the front end of a receiver before A/D
conversion, even if planes are used. Like any other general statements,
there are likely to be exceptions, as Mark noted, specific situations may
require specific actions.

Ellis
  

Mark Gill wrote:

  

Mike -

I have seen both effects, where it has filtered noise from loud power&ground
planes, but also, where it increased the noise at the oscillator due to how
the filtering was designed and implemented. Also, be careful about EMC
recommendations - they are not generic in nature, and can lead to a great
deal of both problems and costs if not applied to the correct situation.
Specific situations lead to specific treatments.

Regards,

Mark Gill, P.E.
EMC/Safety/NEBS Design
Nortel Networks - RTP, NC, USA
  

        -----Original Message-----
From: Mayer, Mike [SMTP:mikem@artesyncp.com]
Sent: Tuesday, July 25, 2000 11:14 AM
To: 'si-list@silab.eng.sun.com'
Subject: [SI-LIST] : Ferrites on power leads

        A couple of people here attended a seminar recently and were given
advice on
a design. Among the suggestions were placing ferrite beads in series with
the power connection of all oscillators and in series with the power pins of

the board connector (it is a daughter card). I'm assuming the reasoning is
that this will "keep noise out of the power distribution system".

        Is this technique valid for designs that use power and ground
planes?

        In every other case we try to minimize the impedance of power
connections.
It seems counter-intuitive to take oscillators and try to add impedance to
their power connections. Are they really that much worse than modern CMOS
devices with millions of transistors switching at multiples of the bus
frequency and hundreds of sub-nanosecond I/O drivers? And does a board with
power and ground planes that plugs onto a board with power and ground planes

need ferrites in series with its power connections?

        This seems like it improperly applying rules that may have applied
to boards
without power and ground planes.

        
============================================================================

=
Mike Mayer Artesyn Communication Products, Inc

Senior Hardware Design Engineer Madison, WI
mikem@artesyncp.com http://www.artesyncp.com
<http://www.artesyncp.com>
============================================================================

=

        **** To unsubscribe from si-list or si-list-digest: send e-mail to
majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE
si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
<http://www.qsl.net/wb6tpu>
****

**** To unsubscribe from si-list or si-list-digest: send e-mail to
majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE
si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
****


About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Wed Nov 22 2000 - 10:50:55 PST