# How to use ferrite beads to attenuate high-freq noise?

Discussion in 'Electronic Basics' started by Dummy, Sep 16, 2003.

1. ### DummyGuest

I have four radios with different transmission frequencies and there's
a power supply and voltage regulator nearby. I put some ferrite beads
on the power supply cable, hoping that it will attenuate the high
frequency signal from radios. Somehow, sometimes I found that the
interference became slightly significant or did not reduced after
in series along the power supply cable? Is the location of the ferrite

2. ### Pieter HoebenGuest

Yes, very important! Place it as close to your radio as possible.

But usually using beads is not enough or even the wrong solution. What
they do can be two things:
- dissipate hf by eddy cirrents, but you need special beads with loss
for this!
- make the hf impedance of the cable part after the bead higher so if
<IF!> your radio has a low impedance there, some attenuation will be
done.
Putting beads in series increases the impedance, so the cut-off
frequency will get lower.

A good radio is not so sentitive to its power lines. So i doubt the

I personally think that your should not only look at the radio's but
go to the source of the problems: the power supply and voltage
regulator. Are they analog or switching?

Also look at the wiring, at loops etc. Are the radio's grounded?

Regards,

Pieter Hoeben
www.hoeben.com

3. ### DummyGuest

Guess my question was a bit vague.

I have four voltage regulators which supply voltages for 4 radios
regulators. My aim is to eliminate the noises that appear at the
output of the voltage regulators. The noise amplitudes were getting
very very high ( up to 13.0V peak to peak! ) when 4 radios were keyed
up simultaneously. I am not sure whether it's the RF frequencies that
emitted by antennas cause non-linearity to the voltage regulators and
distorted the output voltage.

The voltages were supplied to the radios through a battery connector.
But the battery connector supply was not tapped directly from the
output pin of the regulator. Regulator voltage would travel through
and PCB trace, go through a banana jack and power supply cable before
reaching battery connector. In order to eliminated the annoying
noises, I have:

-Put ferrite beads at the battery connector supply line. It didn't
seem to be working fine. The regulator circuits were covered by metal
plates which has the common ground as the circuit. The weird thing was
when the plate was shorted to main supply ground, the noise reduced
significantly, from about 13.0V pkpk to 3.5V pkpk. A factor of about
3.7! Nevertheless, this method had to be used along with the ferrite
bead at the battery connector supply line. If the ferrite bead was
removed, noise amplitudes would stay at the same level.

-Put a big cap (4700uF) at the output of the regulator. Not good
though. By observing the noise through oscilloscope, the noise
frequencies were below 100Hz. Hmmmm...noises only appeared when radios
keyed up.

-Wrapped the regulator with aluminium foil. However, noise amplitude
still wouldn't go down.

-Now I'm planning to put two bypass caps 0.1uF and 680nF as close as
possible to the output regulator on the PCB. I doubt whether that
would work.

I really don't have any ideas why this would happen. Ok, I know by
separating the radios some distance away from the regulators would
help to reduce the noise. Any other suggestions other than this?

4. ### Robert BaerGuest

You have not stated if the regulators are linear or switching.
I assume a single supply feeds the regulators, and there is one
So far, so good - but that means a common lead/wire is used for the
negative line.
That trace shoult "wye" out from the supply to each regulator, and
each regulator output + and - line should go "directly" to a given radio
via a simple wire pair.
If the regulators are 3-terminal linears, use 0.22uF bypass caps right
by each one.
Now as long as the radios themselves are not connected to each other
by *ANY* other means, it would
make no difference if one radio has a - common chassis and another has a
+ common chassis.
Switching regulators should have bypassing according to manufacturer
specs; no more and no less.
Using a large electrolytic cannot help RF bazzfazzz, they aer not
designed for high frequency (say above 1MHz) work.
If the above does not help, you may have something connected
incorrectly; good power supplies have noise in the low millivolt region
at worst.
The fact that connecting a bunch of radios to the power system is a
very good sign of a serious interconnection problem.
You can temporarily add a 50 ohm resistor in series on the + supply
independently connected).
Connect up one radio at a time, then power that radio on, and look at
the noise.
Is the scope grounded? Is the supply grounded? Is it the same ground?
If so, you have created a ground loop which can *create* problems!

5. ### DummyGuest

Main power supply, scope's probe and regulator have common ground.
That means each ground is connected to the same point. I'm using 3
terminal linear regulator. Recommended capacitors have been put at the
input and output of the regulator. Before keying up the four radios,
regulated voltage was smooth with pkpk noise 120mV. Not quite bad
right.

The four regulators are connected in such a way that they are sharing
unregulated voltages from one main power supply. 10V to 12V
unregulated to 8V regulated.

I've added two ceramic caps at the regulator's output on the PCB. It
didn't work though. Since the noise frequency that I've seen in scope
was quite low (100Hz or below). It wasn't RF frequecy. I guess no
point to add those two bypass caps.

Meanwhile, I will try to add the 50ohm res to the regulator's supply
line. By the way, what's the purpose of the 50ohm resistor? Matching
purpose?

6. ### Robert BaerGuest

1) So the scope ground line connects to the ground of the supply. Is
power ground crom the supply power cord connected to the supply output
ground? If so, you caneasily have a ground loop - look for 60Hz and
120Hz "ripple".
Use a handheld (battery powered) DVM in AC mode (or VOM using blocking
capacitor) to make similar checks.
2) The idea for 50 ohm resistors on the output of each regulator is to
partly absorb Rf, and partly to isolate the load from the regulator.
Perhaps the problem will not appear, or will show up with 5 radio
Or maybe the problem only shows with a given radio (other 4 connected,
a specific one as the 5th doing the deed).
It is a means of helping to find where the cause may be....

7. ### Dimitrij KlingbeilGuest

Basically you've done the right thing. The ceramic caps are necessary for
the regulators, so keep them connected. The noise (100Hz) looks to me like
being caused by an insufficient PS transformer voltage. Make sure you
_really_ get 10 to 12V from the PS. Maybe add several 4700 microF caps
before the regulators, however this will not help if the transformer
voltage is too low. Remember that a linear regulator is likely to need
about at least 2V between IN and OUT to operate properly. There will be no
8V at OUT if there are just 8.5V at IN. If there is insufficient IN
voltage, it is very likely to leave the rectifier 100Hz noise pass through.
DO NOT connect the 4700s to the output of the regulators as you have done,
conect them to their INPUT. The strength of the noise looks to me like
there is either no cap at the input at all or there is a very small one or
the transformer cannot deliver the required power because of internal
resistance. Don't use a 50Ohm res. Caps should suffice and if the trafo is
wrong, it won't help either. 100Hz noise looks like a wrong voltage
somewhere, a res won't kill it.

Dimitrij

8. ### Robert BaerGuest

I hope the OP appreciates your comment.
I *should* have thought of that!

9. ### DummyGuest

I don't even know that scope setting would affect the voltage and
noise level reading. Let's say the noise level is 4V pkpk. There are
some BW limit and ch1 50 Ohm setting. Turning ON BW limit option, you
wouldn't see much noise. Something like 5x reduction. Turning ON ch1
50 Ohm option, the voltage would drop from let's say 8V to 2.5V.
However, noise level still remained at 4V pkpk. I'm confused. Who
knows I've been measuring the wrong thing all this while. Maybe the
'real' noise isn't that bad at all.