Choosing and using chip ferrite beads for EMI suppression

[j777]

What are the benefits of putting chip ferrite beads in series with both the DC supply line and the ground line where they enter a PCB? I've read that this type of configuration can provide a level of protection against ESD. Is this correct?

 

[berkeman]

I've not seen beads used in the supply lines/traces as part of ESD protection, but I suppose there might be a small benefit.

Ferrite beads in the supply lines are more typically used to help low-pass filter the supply noise, to keep RF noise from coupling into a circuit that is sensitive to it, or into a transceiver circuit that drives lines off-board (where they can more easily act as antennas and radiate the noise). The LPF is formed by the series beads and the parallel decoupling caps for the circuit being protected. Note that the beads must be rated for the Idd current, or their RF impedance will be degraded.

 

[j777]

Is there any other reason (besides a possibly small level of ESD protection) that you'd want to put a ferrite bead in series with a PCB's negative supply line (I called this ground above)? Possibly to prevent the propagation of noise coupled to the negative supply line?

 

[berkeman]

Is there any other reason (besides a possibly small level of ESD protection) that you'd want to put a ferrite bead in series with a PCB's negative supply line (I called this ground above)? Possibly to prevent the propagation of noise coupled to the negative supply line?

It's pretty common to use a bead in series with the + supply to a chip, and to use lots of decoupling caps on the chip side of the bead(s). It's less common to use a bead for the ground side, because you can usually connect straight to the (<<EDIT>>) ground plane, and if you floorplan your PCB well, you will not share any ground impedance between the noisy circuits and the sensitive circuits. This type of floorplan scheme is called a "Star Ground", because the ground paths out to the different functional blocks radiate out from some center of the star ground.

For example, on my PCBs where I'd commonly have a uC and a network transceiver area, I place them apart, with the center ground area between them strongly coupled to my metal device enclosure. This helps to keep the ground quiet for the transceiver, which keeps RF noise off the network cables, and helps me to pass FCC RF emissions testing.

 

[j777]

I was reading through a Murata application guide and found the following:

"Noise is conducted from the DC power supply and GND
lines to the switching power supply and radiated out
using the AC power supply cable as an antenna. To
suppress the noise, the BLM--P series (Chip Ferrite
Bead) and NFM--P series (Chip EMIFIL® for DC
power supply circuits) are installed close to the power
supply connector.
A filter whose rated current is sufficiently larger than
the power supply current should be selected."

 

[berkeman]

Well, for EMI heading out the package on the power lines, it's usually more practical to filter the power supply input connection, IMO. Different strokes for different folks, I guess. Of course, muRata is in the business of selling more beads...

 

[j777]

That's what I plan on doing. I'm going to put two beads right at the input connector. Thanks for your help!

 

[galapogos]

Hi,

I didn't want to post another thread since I have a similar question. Sorry for threadjacking...

What about the usage of chip ferrite beads for signals rather than power? What kind of signals need to be filtered for EMI suppression, and how do I choose what kind of chip beads to use, besides matching the rated current, voltage and DC resistance to what I want? I'm especially confused about impedance rating(is this a case of the higher the better/more RF attenuation?). Would these chip beads negatively affect the actual signal?

Thanks.

 

[berkeman]

What about the usage of chip ferrite beads for signals rather than power? What kind of signals need to be filtered for EMI suppression, and how do I choose what kind of chip beads to use, besides matching the rated current, voltage and DC resistance to what I want? I'm especially confused about impedance rating(is this a case of the higher the better/more RF attenuation?). Would these chip beads negatively affect the actual signal?

Welcome to the PF, galapogos. Here is a good place to start:

http://www.murata.com/emc/index.html

If you click on the "Chip Ferrite Beads" selector tree, that will give you some part numbers for different applications. And the links on the right side of the page are fairly helpful for application information.

EMI ferrite beads are often specified by their impedance at RF, say around 100MHz where many practical EMI problems occur for microcontroller based devices. Think of the impedance plot for a typical inductor, and you'll see the rising impedance with frequency up to the resonant frequency (resonating with the inductor's parasitic capacitance), and the impedance peaks at resonance, and then falls off because of the dominant capacitive term at high frequencies. But the ferrite materials that are used for EMI beads are lossy, so the impedance tops out at some value, and holds pretty well across a wide range of RF frequencies. So when the bead is rated at "100 Ohms at 100MHz, that's what the impedance graph is topping out at and holding at up in that frequency range.

Does that help?