Question on EMI filtering

Question

I have seen people using ferrite beads to suppress EMI from switching power supplies. Recently, I came across the following blog post where the author successfully re-models a switching power supply to reduce EMI noise. At one point, he makes the following comment.

It is common to attempt the use of ferrite beads to suppress RF Interference of this sort, but it's very unlikely that it will help much - particularly at lower frequencies (e.g. lower HF bands such as 160 and 80 meters, not to mention the AM broadcast band) because these devices simply cannot add enough inductance to add a significant amount of impedance: At these frequencies (say, below 10 MHz) it takes multiple turns on a chunk of ferrite to add enough reactance to make even a small dent in the amount of conducted interference!

My question is that whether the author is right or not? I have read that ferrite beads in EMI filtering, acts as a lossy element and turn the RF energy into heat. Did the author forgot to consider this and only considered the high impedance of an inductor at RF? I'm confused as his strategy worked nevertheless

EDIT: But, I have seen another ham successfully build and test an SMPS using ferrite beads.Here is the link and here is the circuit Of course he used quite a handful of ferrite beads.

Answer 1

I've designed a number of switching supplies that had to pass various EMC standards starting at 30MHz and up, and I can say that ferrite beads are excellent at frequencies of 100MHz and above, but not all beads are created equal. For lower frequencies I've found they are not entirely useless, especially the 1206 and larger sizes, esp. with modern low-ESR ceramic capacitors in a pi-filter arrangement (bead in series, cap to GND at either end).

They work surprisingly well for 3 reasons :

1. The bead adds a small series, non-lossy, inductance (and here, choosing the correct bead with a high series inductance helps a lot)
2. They add a small series resistance. This combined impedance can still significantly dampen frequencies of 10MHz.
3. Simply making provision for a bead forces you to route everything through a single, narrower path, this usually helps the layout.

In older designs we used to simply make a pi filter using the same inductor as for the DC2DC converter. This works well for filtering older supplies with lower switching frequencies (and keeps BOM cost down), and for up to a few MHz of noise, but seems to drop off in effectiveness as you approach the SRF of the inductor you chose - but of course you can get quite far with e.g. a 100nF (even if rated at hundreds of volts, if need be) ceramic right at the power entry pins, relying on the wiring inductance to provide the other half of the filter.

LTSpice has excellent models of different Wurth ferrite beads, which allow you to see what attenuation can be achieved with even the tiniest of beads and some low-ESR caps.

Layout is crucial however, and a good DC2DC converter IC will come with a wealth of good suggestions in the layout advice.

Answer 2

Ferrite beads are not even that lossy at such relatively low frequencies, they behave more like a (small, a few uH) inductor, which can be a trap for the unwary.

They're more useful at VHF and higher frequencies. From here

Answer 3

He's right. Most ferrite beads are rated at 100 MHz and up. They have minimal impact in the 1 - 100 MHz range, which is what he's interested in.

And since as a ham, he's operating sensitive receivers in that range, he needs a LOT of attenuation.

Answer 4

Low pass filters need THREE elements to function:

• a SERIES element, either a resistor or inductor or combined R+L in parallel that lets the inductor pass the DC_current with no/low DC headroom cost, and the resistor provides dampening

then into

• a SHUNTING element, the capacitor

with the SERIES and the SHUNT elements built over

• a GROUND plane, needed for separation of Ground_IN and Ground_out wires.

Consider 10 ohms and 10uF. The time_constant is 100 uSec, or a 1,600 Hertz F3dB. this low F3dB will attenuate high frequency Switch Reg trash (perhaps at 1.6MHz) by 60dB ideally. Yes, at 10mA Iddq needed by your analog circuit, your VDD_A now is 10mA * 10 ohms = 0.1 volt lower, thus 3.3 volts becomes 3.2 volts, but most analog circuits will work just fine with 0.1 volt less VDD.

Here is link to an answer that provides Frequency Reponses of filters with and without dampening, with and without DC_series resistance that uses supply headroom. Read all the answers, please.

Bypass capacitor vs low-pass filter

WARNING do not share solder points on the Ground Plane with GROUND_IN, Ground_OUT, and CAPACITOR_GROUND. For good attenuation, have the THREE solder points well separated --- at least 10 diameters of the solder points.

Realize you need to learn about 2_D current flows, in the very low resistance of a sheet of copper foil.

Answer 5

My question is that whether the author is right or not?

The author isn't considering that FBs are predominantly used as a lossy component and so, I think he either misunderstands FBs or he's being lazy in his word choice. He's correct in that FBs don't bring a lot to the party at low frequencies other than a little bit of inductance but, that's not what FBs are all about.

Did the author forgot to consider this and only considered the high impedance of an inductor at RF?

That may be true but his use of the ferrite bead term is not helpful.