Ferrite beads (L1 in the photo) filter high frequency power supply noise by converting it into a tiny amount of heat. Power supply noise can cause various problems for many parts, especially in analog audio and display circuits.
Ferrite beads are simple, but choosing one can be confusing because they’re not commonly used by hobbyists. Most designs will still work if you omit the ferrite bead(s), but beads are so cheap there’s no reason to sacrifice the added reliability they provide. We describe how we pick ferrite beads for our projects after the break.
A ferrite bead is rated for current, impedance, and resistance; see this Mouser listing for an example. Unless a datasheet or circuit requests specific bead characteristics, we choose a bead rated for sufficient current, and ignore the impedance and resistance values.
If the bead is for a power supply, we determine the maximum possible current the circuit will use and find a bead rated for double that amount. Last week we calculated the the Bus Pirate’s worst-case current consumption as 525ma, so we looked at beads rated for at least 1000ma. We used this one, which is rated for 1500ma and costs 10 cents.
Sometimes a ferrite bead is used to filter the power supply for one specific part of a circuit. We used a dedicated bead to filter the LCD bias voltage on the DIY digital picture frame, and with the ENC28J60’s ethernet transceiver on the web server on a business card. These parts only consume a few milliamps, so we used a smaller 200ma ferrite bead ($0.11).
Like this post? Check out the parts posts you may have missed. Want to request a part post? Please leave your suggestions in the comments.
I really wouldn’t mind seeing a post on buck-boost converters and the like :)
“we choose a bead rated for sufficient current, and ignore the impedance and resistance values”
And when I select a capacitor I choose it for sufficient voltage and ignore capacitance and ESR values.
And there was that time I bought a car and chose it according to a colour preference, ignoring power, fuel consumption, size, cost, insurance bracket…
Inductance is not a simple concept but not all that difficult to get your head around.
Please don’t purport to be a guide to components if you teach readers to ignore relevant details.
With all the negative comments lately I figured i’d leave a well deserved one. I love posts like these, informative on things not usually covered or easily discovered by the average hobbyist. Thank you!
.dok
Hello Ian,
I was wondering if you have the source code (firmware5) for the project “Record Audio with MSP430” http://www.diylife.com/2008/05/16/record-audio-with-the-msp430-microcontroller/. You created a DIY project that recorded audio into a SD card using the msp430, however the source code in DIYLife9.zip does not contain firmware5.
@ some meng miet
while it is important to understant inductance, for most application, ferrite beads are just extra and not all that important to get a functional design for hobbyist… so i still think the post is appropriate for this site… it is “hackaday” afterall… not “EE-professional-aday”
it is more of an introduction. people should do more research on it before they use it anyway…
I thought ferrite beads were just those little snap-on things that go over cables to block out interfer-oh okay nevermind those are ferrite cores.
Still have to agree with meng miet though, what if you are actually trying to pass a certain frequency through the bead and it ends up attenuating it?
But investigating which bead to use would probably involve complicated bode plots idk D:
Erm, this may seem something of a noobish question, but how do you actually connect these? The linked wikipedia article says to wrap this round a wire, but the component in the photo is obviously not used like this. Love the components posts though, and I second mjs’s request for buck/boost converters.
re: “chrizlax says:
July 6, 2009 at 1:11 pm
Erm, this may seem something of a noobish question, but how do you actually connect these? The linked wikipedia article says to wrap this round a wire,…”
That gave me a laugh, sorry lol
Thanks, I was wondering about why some designs include ferrite beads and some don’t. If the article was too simplistic, well, all I can say is it fit me today.
The part they are showing is actually a small inductor. Inductors typically consists of a core either air or some other material (such as ferrite) with wire wrapped around it. The current flows through the wire and high frequency is attenuated because the poles in the core can’t change as fast as the current is.
If the worst case current consumption of the bus pirate is 525mA SHAME ON YOU!!!! I know it will never draw that much in practice, but if you are going to design and sell something that connects to a USB port it should meet specifications. USB ports at a maximum can supply 500mA (as specified by the USB-IF.
BTW, I’m not a douche. Just a bored computer engineer at work! WOOHOO ITS 5!!!
While it’s good to give a tutorial on a topic, this post didn’t actually contain any information, other than ‘leave a safety factor of two.’ Good practice, but it doesn’t match the title.
I’m an EE, but I’d love to hear some rules of thumb for choosing inductors/ferrite beads in circuits.
@some meng miet
Perhaps you could add something to this article instead of just ragging on the author. I mean if you’re going to waste the time then you might as well contribute something.
@chrizlax
connect in series, say between the bulk capacitor of a switching power supply and the rest of the circuit.
@cp
Some don’t have ferrites because they’re mostly or all analog with no fast switching components. Some don’t because the designer just didn’t give a ____ what the fcc thinks.
@ Trey, are you sure? Because the current specs are MINIMUM guidelines, not absolute maximums. Low usage and unregistered usb devices should be able to source atleast 1 unit of power, at 100ma. High usage devices should be able to source atleast 5 units of power, or 500ma. The Specs do no list a maximum current.
There seems to be some confusion, here, between ferrite beads and inductors. While ferrite beads do have some inductance, they are commonly used as “frequency-dependent resistors”. Most data sheets show a plot of their resistance versus frequency, and they usually have a rating like “100 Ohms at 100 Mhz” to show where the lowpass cutoff begins. Ferries pass DC (0 Hz) components with very little resistance or inductance, and attenuate high frequencies well. Therefore, they’re really good as filters in power supplies where you’re interested in DC. They’re also superb at preventing high-frequency noise from getting out of digital chips and on to the power plane. Just place a ferrite between the power plane and the IC’s power leg.
I love how when someone posts a negative comment, every one rags on them, like they have some reason to tell a person off now. Not all comments are going to be positive–get over it. And by submitting a comment that does nothing but berate another person, how are *you* contributing?
All in all, I think this was an informational article and I hope to see more like it.
@yaff: Thanks for explaining the difference between inductors and ferrite beads. I was not aware of this and was about to jump upon the claim made in the article about converting high frequency noise to heat.
Inductors can also be used to cancel noise but (ideally) do not create any heat. They store energy in the magnetic field during half of an AC cycle and give it back on the other half of the cycle. In fact they function the same as capacitors but with “voltage” and “current” switched and with “magnetic field” instead of “electric field”. So they do not dissipate energy as heat just as capacitors don’t dissipate any energy. In fact you can plug a inductor or capacitor into the wall socket (but don’t do it unless you are sure the part is rated for this) and it will draw no power even though you can measure a current flowing through it. Your power company will not charge you any money for this because the power flows into the load for half of the cycle and back out to the power company for the other half of the cycle. There will be small resistive losses though due to the fact that the wires are not perfect conductors.
I never knew that ferrite beads operated any differently than inductors. Wikipedia has some good starting information. It seems that ferrites are better for high frequencies both because of their resistance and because of the lack of parasitic capacitance. From a quick look it seems that with the right material they can filter down to 1 MHz. Too bad (for me) that they can’t filter audio-frequency noise out of the power supply.
The physics behind ferrite beads, inductors and all parts with magnetic flux are the same.
“While ferrite beads do have some inductance, they are commonly used as frequency-dependent resistors”
please bear in mind that the rising resistance of the ferrite bead is exactly because it has inductance. The (complex) resistance of any inductor is Z=1/(j*w*L) with w=2*pi*f. So any inductor will show a lower resistance to DC than to HF-Noise or Signals.
I think/hope you (yaff) meant the following:
any conductor has a (parasitic) inductance, but ferrite beads are special inductors, because they operate up to high frequencies where the complex resistance is dominated not by plain resistance but by inductance and parasitic capacity so it is easier just to stat the complex resistance.
Its in the material, you cannot use a hunk of metal to make a transformer at 1000kHz, Ferrite takes you to much higher Frequencies because the conductivity of ferrite is much lower resulting in no eddy currents. Look at the skin effect and you know why.
To filter AF-Noise you could throw a LC-Filter at your problem, RL or RC could be too weak
I hope this helps someone
While tweaking the least significant decimals of any design parameter “can” be done, it’s sadly either over or under applied. The reality is that ferrites can be both magic silver bullets of evil slaying power-or just another useful component. The false perceptions have taken on a life of their own. There’s a “Placebo Effect” Aura -like the niche marketed magic cables etc where the perception of what ferrites do has become an unquestioned reality. Here lies a path well worth education’s best efforts. As properly applied ferrite devices literally make some designs workable or unworkable. Ferrites have some performance factors unmatched pound for pound compared to any other method of solving a parameter need. That’s why an iron core transformer power supply weighs tens of pounds to the tens of ounces for switch mode ones for one example.
Try constructing for another good example- a “Joule Thief” with no ferrite materials. Then, as a design proof of concept, make a workable joule thief from supressor grade ferrite beads. It’s an object lesson in non-trivial maths to do it by the book. Whacking together some windings and using an old sencore z meter to get close enough for trials is what demarcates hackers from non-hackers. With a rarefied realm of ubergeeken tesla blessed folks who just whip some tefzel wire around a toroid and it ..works… Sadly, I’m not that good:>
From the Wikipedia article http://en.wikipedia.org/wiki/Ferrite_bead it appears that the distinguishing characteristic of ferrite beads is not just that they can operate at high frequencies but that they actually have parasitic resistance at these frequencies. So it is not the case that the imaginary component of the impedance dominates, but rather the real component. This is contrary to what is wanted in a transformer, which I imagine may be made of a different form of ferrite. The link at the bottom of the wikipedia article titled “Understanding Ferrite Bead Inductors” gives a plot showing the impedance being dominated by the resistive component. They claim that a ferrite bead can be thought of as an inductor in series with a frequency-dependant resistor.
And now after reading all comments I think i know just a bit more about Ferrite beads.
Thx HackaDay and al the commenter’s. Positive or negative, I dint knew what those ferrite parts where used for, now I now a bit more I can do some Research myself.
Hackaday is a collective ? Borg? :P
@dan
parasitic resistance is everywhere, every solder joint has some, also every conductor has parasitic capacitance and inductivity. But the real part of complex resistance (impedance) is frequency independent in opposition to its reactance. Some real resistance is needed to dampen resonance and also limits the maximum current.
I cannot agree “the impedance being dominated by the resistive component”, it is shunted by parasitic capacitance and increased by inductance. What you get is a few ohms at dc and 1kOhm at 10MHz for example, the real part of the resistance is constant, so it can only be the complex part.
@Rlyeh_drifter
The complex impedance for an inductor is z = j * w * L and for a capacitor it’s z = 1 / (j * w * C), where j = sqrt(-1) and w = 2 * pi * f.
grrr i meant to write X = j*w*l, z = r + j*w*l. this blog needs an edit button ^^
@noobius
thanks for correcting, typing faster than thinking is a problem :-)
does any one have a suggestion for a reasonable meter that will do L and C in addition to R, V, and T?
When reclaiming parts I’ve always just ignored SMT ferrite beads. This is mostly because of the difficulty in determining enough about them to be able to re-use them. If pretty much all of them are suitable as high frequency filters they suddenly become much more interesting.
Unfortunately current rating seems like it’s the hardest parameter to non-destructively determine.
Are there any rules of thumb, such as minimum power ratings for various sizes?
AFAIK their function is to create eddies inside them that resist/counter high frequencies, they are meant to block unwanted frequencies and therefore need to be selected on that function too, but of course while being compatible with the circuit they’re in.
So yes to say you should only look at the voltage might be going too far even for the must uninterested, even when they can be considered having a broad effect, like gloves can protect your hands but if you go skiing you might still want to not use welders gloves even when they are better than no gloves and do insulate you from the cold too.
@Rlyeh_drifter
Check out the article by Murata, which is linked to by the wikipedia article: http://www.murata.com/emc/knowhow/pdfs/te04ea-1/23to25e.pdf
The graph shown depicts a frequency-dependent resistance. They don’t plot the imaginary part of the impedance but they show the magnitude approaching the real component for high frequency which seems to imply that the imaginary component approaches zero. The real part starts below 1 ohm and rises to over 100 ohms. A second graph contrasts this to an air-core inductor in which the imaginary impedance increases linearly until capacitance comes into play, with the real part of impedance (the resistance) being insignificant throughout.
The wikipedia article and the murata article both make it a point to say that these ferrite beads are different from ordinary inductors, and energy dissipation in the form of heat seems to be a significant feature of their operation.
you said butt beads
@dan
Thank you!
The key is the permeability! I thought it would be somewhat constant, but through frequency-dependance comes a complex permeability, so thats where the rising resistance comes from!
This document served me well:
http://www.steward.com/pdfs/emi/technical/Use%20of%20Ferrites%20in%20EMI.pdf
I’d like to see a post on triacs. I’m working on reverse engineering a project that uses them for switching AC loads without relays.
Thanks.
thanks hackaday
ferrite increases the inductance of a conductor many times, the increase depends on the permeability of the the ferrite . There are ferrites which have values 8000 , this means , just sticking a bead over a leg of say a D connector WILL PREVENT high frequency signals ie noise entering or leaving the enclosure .ON the other hand DC will go through as if there was no resistance. Simply put ,they stop AC but not DC. They are used a lot in RF engineering for all types of filters,rf transformers.