Ferrite rings do anything or just marketing frill?

[The little lost angel]

If they didn't help, would this one have been installed -- inside,
where it can't be seen, by a company that doesn't bling up its PSUs?

http://static.flickr.com/66/158482345_d4351fe79c_o.jpg

(300W Delta DPS-300BB)

Ah!

One contributing factor to my skepticism is that so far all those who
made it a point to emphasize the chokes on the VGA output are all erm
bling bling brand names targeting PC "decorators" as one poster puts
it. Hence my immediate skepticism about the usefulness of the choke in
the specific area it was being used in.

How about braiding/twisting the individual wires together the way
network cables have their individual strands braided? Another bling
bling PSU brand claims the same reduction in EMI/improvement in
stability etc for doing so. Is this also the truth or marketing?

 

[John Popelish]

The little lost angel wrote:
(snip)

How about braiding/twisting the individual wires together the way
network cables have their individual strands braided? Another bling
bling PSU brand claims the same reduction in EMI/improvement in
stability etc for doing so. Is this also the truth or marketing?

This approach works well to contain the energy that is intended to
pass along the cable, if the signals are in the form of a balanced
differential pair (one signal going positive as the other goes
negative). This causes their external fields to approach zero when
the cabling is balanced (twisted pairs or shielded twisted pairs).

But it still doesn't solve the problem of the whole cable resonating
as a single conductor and acting as an antenna for any noise on the
signal common that matches its resonant frequency. A cable shield
with nothing inside it will radiate this way.

The ferrite bead around it, absorbs energy that damps that resonance,
as charge bounces back and forth along the shield.

 

[Richard Henry]

Ah!

One contributing factor to my skepticism is that so far all those who
made it a point to emphasize the chokes on the VGA output are all erm
bling bling brand names targeting PC "decorators" as one poster puts
it. Hence my immediate skepticism about the usefulness of the choke in
the specific area it was being used in.

How about braiding/twisting the individual wires together the way
network cables have their individual strands braided? Another bling
bling PSU brand claims the same reduction in EMI/improvement in
stability etc for doing so. Is this also the truth or marketing?

It's marketing truth.

 

[The little lost angel]

This approach works well to contain the energy that is intended to
pass along the cable, if the signals are in the form of a balanced
differential pair (one signal going positive as the other goes
negative). This causes their external fields to approach zero when
the cabling is balanced (twisted pairs or shielded twisted pairs).

So since the wires in a PSU simply carry energy and not data, i.e.
there's no existence of balanced differential pair, doing this would
effectively be pointless right?

The ferrite bead around it, absorbs energy that damps that resonance,
as charge bounces back and forth along the shield.

So the ferrite bead is effective even tho it's only at one end of the
wires? How would it compare to say a metal mesh shield wrapped around
the wires carrying power?

 

[Jan Panteltje]

On a sunny day (Fri, 02 Jun 2006 19:27:27 GMT) it happened
[email protected] (The little lost angel) wrote in
<[email protected]>:

The ferrite bead, look at it as a series inductor
It forms a core around the wire, making its inductance bigger.
(Like if the wire was wound around some iron core).
Higher inductance means higher impedance for higher frequencies.
Z = j.w.L (w is omega, 2 x pi x f

If there is any capacitance at the end of the wire, it wil lform a low
pass filter.

The current in the wire will decrease more for higher frequecies i = U / j.w.L

Think of j as 90 degrees phase shift, impedance and current is a complex number.

Also that bead has some losses, likely also increasing with frequency.

Zank you

 

[John Popelish]

So since the wires in a PSU simply carry energy and not data, i.e.
there's no existence of balanced differential pair, doing this would
effectively be pointless right?

If the supply produced pure DC voltage (no high frequency ripple or
spikes) and the load consumed pure DC current (no logic loads
switching), then yes. Otherwise, there is some unwanted high
frequency content on those lines that might escape as radiation., so a
ferrite bead might be placed to suppress it.

So the ferrite bead is effective even tho it's only at one end of the
wires?

That is the place they usually work best, but the actual best spot may
need to be found by trial and error, making measurements of the
radiated noise spectrum for each trial. Think of the wire as a whip
antenna sticking out of a ground plane. The peak current in the whip
occurs at the ground plane end, with peak voltage at the free end.

How would it compare to say a metal mesh shield wrapped around
the wires carrying power?

A metal shield around the wires contains the electric fields but the
magnetic fields just move to the outside of the shield. Ferrite beads
do nothing for electric fields, except to choke off the charge
movement that feeds them.

 

[Rich Grise]

So since the wires in a PSU simply carry energy and not data, i.e.
there's no existence of balanced differential pair, doing this would
effectively be pointless right?

No, you're not understanding the point. For one thing, the outputs of
the PSU are balanced to the extent that all of the current that flows
out of it flows back in. Going through the ring, these magnetic fields
cancel. But if there's some electrical noise, the conductors act like
an antenna. Switching noise on the lines is probably single ended, too;
these are the things that the ferrite ring filters out.

So the ferrite bead is effective even tho it's only at one end of the
wires? How would it compare to say a metal mesh shield wrapped around
the wires carrying power?

If it is grounded, it will protect against capacitive feedback, but for
the kind of coupling that a ferrite ring is used for, a shield without
the ring would just make it worse, since it's a single conductor that
can pick up induced fields and reradiate them.

Hope This Helps!
Rich

 

[Franc Zabkar]

I've been noticing in some ATX PSU, they have been putting ferrite
rings (that's the round thing I see on the end of video cables right?)
claiming that it reduces EMI and improves quality of the video output
as well as system stability.

My intuition says that's bullshit because on those leads, the PSU is
delivering power at 12V level, unlikely to be affected significantly
by EMI. Then there are all those caps on the cards and motherboards
that would probably deal with any ripple/noise etc already.

Aluminium electrolytic caps don't filter high frequencies very well.
That's why a lot of designs add smaller caps (eg tantalums or
ceramics) in parallel with them.

- Franc Zabkar

 

[Franc Zabkar]

How about braiding/twisting the individual wires together the way
network cables have their individual strands braided?

I have some cables like that. I'm not sure where they came from,
though.

- Franc Zabkar

 

[purple_stars]

i don't know about in that particular case, but definitely they do some
good in other cases. i have an icom 706-mkiig radio that kept shutting
itself off and it turns out that it was picking up RF from the antenna
tuner on the cable between the radio and the little external
display/control that you mount on the dash. a little ferrite snap on
bead from radio shack at either end of the cable fixed it right up.

 

[Captain Dondo]

John Larkin wrote:




Those figure 8s are handy. I used them on a product's intenal ac mains leads
together with a couple of Y caps to get it inside the conducted limits. It was
previously just a couple of dB over. A heck of a lot cheaper thana full blown
line filter.

Graham

Would they help with induction voltages? We run very long cables -
2000' - along 480 VAC power lines and the induction voltages can put 70
V on a cable pair that's not connected to anything (nutted on one end,
voltmeter on the other)....

I wonder if something so simple would work.

(And no, I don't have an EE degree either. )

--Yan

 

[John Larkin]

Would they help with induction voltages? We run very long cables -
2000' - along 480 VAC power lines and the induction voltages can put 70
V on a cable pair that's not connected to anything (nutted on one end,
voltmeter on the other)....

I wonder if something so simple would work.

(And no, I don't have an EE degree either. )

If you mean 60 Hz voltages magnetically induced from power lines into
your signal cables, no, the ferrites won't help. They have useful
impedance only at high, megahertz, frequencies. I have seen larger
cores, toroids and laminations, used to add common-mode impedance at
line frequencies, as in audio and video systems suffering from induced
hum.

John

 

[Del Cecchi]

If you mean 60 Hz voltages magnetically induced from power lines into
your signal cables, no, the ferrites won't help. They have useful
impedance only at high, megahertz, frequencies. I have seen larger
cores, toroids and laminations, used to add common-mode impedance at
line frequencies, as in audio and video systems suffering from induced
hum.

John

The normal thing to do is use a small transformer to get rid of the
common mode, or alternatively, blocking caps. What is the cable
carrying? Ethernet uses transformers I believe. So there is no real
reason to worry about the 70 volts of AC common mode, except it will give
you a nasty jolt. Shielded cable?

del cecchi