Ferrite Bead Question

[[email protected]]

We're making a pcb that's goal is to take in frequencies and filter out
anything not in the 20-40Khz range. We will have a dc/dc converter,
specifically this one: http://www.calex.com/pdf/3wdsmt.pdf
and we are using decoupling capacitors on all filter opamps Vin and
Vouts; and decoupling caps on other components, including the
+/-Outputs of the dc/dc to get rid of noise.

so my question is: someone suggested we use a ferrite bead to make sure
we eliminate any other noise on the dc/dc.. so do we need a ferrite
bead?

All of the Ferrite beads i'm seeing are for Mhz range.. and I thought
we didn't need one that high, but then the data sheet has a section on
pg 3 that says "Ripple & Noise (20Mhz)". What does that mean? Would
the ferrite bead solve that problem?
If so, what are the specs of a ferrite bead that I should be looking
for?

-Thanks-

 

[martin griffith]

On 15 Apr 2006 09:37:35 -0700, in sci.electronics.design

We're making a pcb that's goal is to take in frequencies and filter out
anything not in the 20-40Khz range. We will have a dc/dc converter,
specifically this one: http://www.calex.com/pdf/3wdsmt.pdf
and we are using decoupling capacitors on all filter opamps Vin and
Vouts; and decoupling caps on other components, including the
+/-Outputs of the dc/dc to get rid of noise.

so my question is: someone suggested we use a ferrite bead to make sure
we eliminate any other noise on the dc/dc.. so do we need a ferrite
bead?

All of the Ferrite beads i'm seeing are for Mhz range.. and I thought
we didn't need one that high, but then the data sheet has a section on
pg 3 that says "Ripple & Noise (20Mhz)". What does that mean? Would
the ferrite bead solve that problem?
If so, what are the specs of a ferrite bead that I should be looking
for?

-Thanks-

100mV noise 20Meg BW....... Thats a lot, I dont think a ferrite bead
will be a lot of help.
Check out the Jim Williams amazing app note, I forget the number, on
the LT1533 at Linear. Also think about a Common Mode filter, from
Murata etc. A CM filter might help a lot. It might be easier to try
for a lower noise converter


martin

 

[Winfield Hill]

martin griffith wrote...

Check out the Jim Williams amazing app note, I forget
the number, on the LT1533 at Linear.

an70.

 

[Joerg]

Hello Martin,

100mV noise 20Meg BW....... Thats a lot, I dont think a ferrite bead
will be a lot of help.
Check out the Jim Williams amazing app note, I forget the number, on
the LT1533 at Linear. Also think about a Common Mode filter, from
Murata etc. A CM filter might help a lot. It might be easier to try
for a lower noise converter

Or roll your own.

<brag_mode>
Mine never have that much noise.
</brag_mode>

Regards, Joerg

 

[John Larkin]

We're making a pcb that's goal is to take in frequencies and filter out
anything not in the 20-40Khz range. We will have a dc/dc converter,
specifically this one: http://www.calex.com/pdf/3wdsmt.pdf
and we are using decoupling capacitors on all filter opamps Vin and
Vouts; and decoupling caps on other components, including the
+/-Outputs of the dc/dc to get rid of noise.

so my question is: someone suggested we use a ferrite bead to make sure
we eliminate any other noise on the dc/dc.. so do we need a ferrite
bead?

All of the Ferrite beads i'm seeing are for Mhz range.. and I thought
we didn't need one that high, but then the data sheet has a section on
pg 3 that says "Ripple & Noise (20Mhz)". What does that mean? Would
the ferrite bead solve that problem?
If so, what are the specs of a ferrite bead that I should be looking
for?

-Thanks-

Sure wouldn't hurt. Some dc/dc converters kick out horrendous
multi-MHz spikes or ringing things, which could easily get into
low-level front-ends, or possibly heterodyne with a signal component.

Go for a high-impedance bead, 600 ohms or so (beads are usually rated
for impedance measured at 100 MHz.) Put one in series with the dc/dc
input, and another in series with the output, with ceramic bypasses on
both sides of the beads. If the input and output grounds aren't the
same, put one in the output return lead, too.

HF spikes can play hell with sensitive front-ends.

John

 

[Glen Walpert]

martin griffith wrote...

an70.

also EDN article by Jim Williams:
<http://www.edn.com/contents/images/6288040.pdf>
and a ferrite EMI suppression presentation from Steward Ferrite can be
downloaded from
<http://www.ewh.ieee.org/r5/denver/rockymountainemc/archive/2003/may13/index.html>