1. Home
  2. Electronics Design
  3. Ferrite beads/chips

Ferrite beads/chips

Why do they still make these things? Really? Are they just for suckers who don't read datasheets? Or more precisely, suckers who assume the datasheet tells you everything?

I was perusing TDK's "SEAT" (a whopping 220MB download, written in .NET) the other day. They have DC bias characteristics for some of their ferrite chips. In discrete steps, like 20% of rated current. Okay. So I look at the biggest, beefiest, 5750 package (these are metric sizes BTW),

3A, "120 ohms". It's doing 35 ohms impedance at 1MHz. At merely 0.6A, it's down over 20dB, one or two ohms. A dead short in other words. What's the point?

The highest current rating I can find is a 2012 size, 6A, "30 ohm" part. It's 1 ohm at 1MHz and no bias, 0.2 ohms at 1A. Useless.

Chip components aren't useless. There are tons of multilayer inductors that do real current. Most of them still don't have DC bias characteristics listed. Found one, 1.5uH, 2520 package, 1.4A, that does a consistent "2000 ohms" (peak around 100MHz) at any rated current. Literally orders of magnitude better than the biggest SMT ferrite in the database. And it's only a couple cents more on Digikey.

If you think you need ferrite beads/chips... you're, at best, fooling yourself that you want them. Maybe you can get away with them on data lines. Make sure you pick the highest current you can find. If you're lucky, that "2A" bead or chip will only just begin to saturate when lighting an LED or something. Power lines? Fucking useless. Forget about the "Ferrite Beads and Chips" category on Digikey (and similarly at other places). Shop inductors. That's what they're calling them, I guess.

Oh, and by the same token, be suspicious of multilayer inductors for power or resonant applications. They're lossy, because they're just ferrite beads. Like you'd guess from appearance. Some have reasonable Q, but be sure to check.

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
Website: http://seventransistorlabs.com
Reply to
Tim Williams
Loading thread data ...

Horses for courses. I use pots of beads, mostly the Murata BLM18BB ones, to stabilize microwave transistors in lower-frequency circuits. A

5-ohm bead in the base of a BFP640 makes it completely benign.

Beads have a unique combination of low noise at low frequency and higher, mostly resistive impedance at high frequency. The 5- and 10-ohm ones hold up as far as 3 GHz or more.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs

Signals. Miliamperes. Do you use 3A+ rated beads in those applications?

If not, then why do the damned things exist? The only purposes I can see are: a) to build nonlinear transmission lines, or b) to confuse anyone who doesn't realize they saturate at the drop of a hat.

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
Website: http://seventransistorlabs.com
Reply to
Tim Williams

But the garlic only works as long as you don't burn it. That's why you put it in towards the end of a sautee. As an engineer and cook, you know that.

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
Website: http://seventransistorlabs.com
Reply to
Tim Williams

Certainly. We cooked a big pot of beans today, a mix of red kidney and borlottis, with a little bacon and chicken-apple sausage and a ham hock, not too meaty. Cooked for about four hours. At the end, I shut off the heat and added the chopped garlic. I am *not* a barbarian.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation
Reply to
John Larkin

"John Larkin" wrote in message news: snipped-for-privacy@4ax.com...

Today was leftovers, since I made extra stir-fry and brown rice the other day. Three habaneros....barbarians, what of them?

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
Website: http://seventransistorlabs.com
Reply to
Tim Williams

no DC bias but easier/faster to use than those LaPlace equation models.

The structure is almost identical to eddy current models.

  • SIGNAL RF BEAD
  • 0402 1k 100MHz Rdc= 1.1 ohm Idc= 250mA
  • Murata BLM15HG102SN1 .SUBCKT RFBEAD1K 1 2 .PARAM Lfilter=6.8uH .PARAM Rfilter=1450 .PARAM C1=11 .PARAM C2=3 .PARAM C3=2 .PARAM C4=2 .PARAM C5=1 .PARAM CT={C1+C2+C3+C4+C5} .PARAM B1=1 .PARAM B2=7 .PARAM B3=8 .PARAM B4=11 .PARAM B5=15 RDC 1 3 1.1 L1 3 2 {Lfilter/B1} R1 3 4 {C1/CT*Rfilter} L2 4 2 {Lfilter/B2} R2 4 5 {C2/CT*Rfilter} L3 5 2 {Lfilter/B3} R3 5 6 {C3/CT*Rfilter} L4 6 2 {Lfilter/B4} R4 6 7 {C4/CT*Rfilter} L5 7 2 {Lfilter/B5} R5 7 2 {C5/CT*Rfilter} C1 1 9 0.027pF RLC1 9 8 200 LC1 8 2 55nH .ENDS RFBEAD1K
  • PREFERRED POWER LINE FILTER RF BEAD
  • 120 100MHz Rdc= 0.095 I= 1.5A
  • Murata BLM15EG121SN1 .SUBCKT BLM15EG121SN1 1 2 .PARAM Lfilter=960nH .PARAM Rfilter=165 .PARAM C1=1.5 .PARAM C2=1.2 .PARAM C3=1 .PARAM C4=1 .PARAM C5=1 .PARAM CT={C1+C2+C3+C4+C5} .PARAM B1=1 .PARAM B2=1 .PARAM B3=2 .PARAM B4=2.5 .PARAM B5=3 RDC 1 3 0.095 L1 3 2 {Lfilter/B1} R1 3 4 {C1/CT*Rfilter} L2 4 2 {Lfilter/B2} R2 4 5 {C2/CT*Rfilter} L3 5 2 {Lfilter/B3} R3 5 6 {C3/CT*Rfilter} L4 6 2 {Lfilter/B4} R4 6 7 {C4/CT*Rfilter} L5 7 2 {Lfilter/B5} R5 7 2 {C5/CT*Rfilter} C1 1 9 0.12pF RLC1 9 8 171 LC1 8 2 3.nH .ENDS BLM15EG121SN1
*
  • Murata 0603 BLM15EG221SN1
  • 0402 220 100MHz Rdc=0.28 I= 700mA .SUBCKT BLM15EG221SN1 1 2 .PARAM Lfilter=1.8uH .PARAM Rfilter=270 .PARAM C1=5 .PARAM C2=3 .PARAM C3=2 .PARAM C4=2 .PARAM C5=1 .PARAM CT={C1+C2+C3+C4+C5} .PARAM B1=1 .PARAM B2=1.7 .PARAM B3=2 .PARAM B4=2 .PARAM B5=2 RDC 1 3 0.28 L1 3 2 {Lfilter/B1} R1 3 4 {C1/CT*Rfilter} L2 4 2 {Lfilter/B2} R2 4 5 {C2/CT*Rfilter} L3 5 2 {Lfilter/B3} R3 5 6 {C3/CT*Rfilter} L4 6 2 {Lfilter/B4} R4 6 7 {C4/CT*Rfilter} L5 7 2 {Lfilter/B5} R5 7 2 {C5/CT*Rfilter} C1 1 9 0.15pF RLC1 9 8 71 LC1 8 2 4.nH .ENDS BLM15EG221SN1
*
  • POWER LINE FILTER RF BEAD
  • Murata 0603 BLM18EG221SN1
  • 220@100MHz Rdc=0.05 I=2A .SUBCKT RFBEAD220 1 2 .PARAM Lfilter=1.4uH .PARAM Rfilter=270 .PARAM C1=5 .PARAM C2=3 .PARAM C3=2 .PARAM C4=2 .PARAM C5=1 .PARAM CT={C1+C2+C3+C4+C5} .PARAM B1=1 .PARAM B2=1.7 .PARAM B3=2 .PARAM B4=2 .PARAM B5=2 RDC 1 3 0.05 L1 3 2 {Lfilter/B1} R1 3 4 {C1/CT*Rfilter} L2 4 2 {Lfilter/B2} R2 4 5 {C2/CT*Rfilter} L3 5 2 {Lfilter/B3} R3 5 6 {C3/CT*Rfilter} L4 6 2 {Lfilter/B4} R4 6 7 {C4/CT*Rfilter} L5 7 2 {Lfilter/B5} R5 7 2 {C5/CT*Rfilter} C1 1 9 0.15pF RLC1 9 8 71 LC1 8 2 4.nH .ENDS RFBEAD220
* *
  • 0603 60 100MHz Rdc= 0.1 ohm Idc= 500mA
  • Murata BLM18PG600SN1 .SUBCKT RFBEAD060 1 2 .PARAM Lfilter=350nH .PARAM Rfilter=80 .PARAM C1=3 .PARAM C2=3 .PARAM C3=2 .PARAM C4=2 .PARAM C5=1 .PARAM CT={C1+C2+C3+C4+C5} .PARAM B1=1 .PARAM B2=1.3 .PARAM B3=2 .PARAM B4=2 .PARAM B5=2 RDC 1 3 0.1 L1 3 2 {Lfilter/B1} R1 3 4 {C1/CT*Rfilter} L2 4 2 {Lfilter/B2} R2 4 5 {C2/CT*Rfilter} L3 5 2 {Lfilter/B3} R3 5 6 {C3/CT*Rfilter} L4 6 2 {Lfilter/B4} R4 6 7 {C4/CT*Rfilter} L5 7 2 {Lfilter/B5} R5 7 2 {C5/CT*Rfilter} C1 1 9 0.5pF RLC1 9 8 31 LC1 8 2 4.nH .ENDS RFBEAD060
*
Reply to
RobertMacy

Where's the saturation?

Tim

-- Seven Transistor Labs Electrical Engineering Consultation Website:

formatting link

Reply to
Tim Williams

How'd we get from beads to garlic..? I like to put some garlic in with the meat at the beginning too. (like when I'm browning ground beef for meat sauce (Bolognese)) The garlic gets in there and does something to the meat. Then more garlic at the end... and maybe add some in the middle too. What the heck, it's not like garlic is expensive.

George H.

Reply to
George Herold

no dc bias, I thought I said that. it's easy to add in LTspice, though, just had no need when I modeled these.

complaining about a 'free' model. GEESH!

Reply to
RobertMacy

Well, saturation needn't be a DC thing, but...

:-p I've got a saturable lossy core model kicking around somewhere, too. Easy to add -- of course. The hard part is finding the parameters though.

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
Website: http://seventransistorlabs.com
Reply to
Tim Williams

I made the cooking analogy... it kinda sorta makes sense, I guess?

Could also go with "don't burn the roux", kind of analogous again I suppose.

Last time I made pot roast, I shoved about a half a head's worth into the thing. Was damn good too.

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
Website: http://seventransistorlabs.com
Reply to
Tim Williams

My wife is half Italian, so she likes those cooked-to-death red sauces. I'm more of a white-sauce sort of guy.

The best white sauce is heavy cream and Regiano parmesan, maybe a little Una Kaas hard aged gouda. Bit of tarragon and pepper, dash of sherry maybe.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation
Reply to
John Larkin

Needs more butter.

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward" 
speff@interlog.com             Info for manufacturers: http://www.trexon.com 
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
Reply to
Spehro Pefhany

People use even the higher ohm versions to above 10GHz. Sure, they are res istive-capacitive up there, but they still have some series impedance remai ning.

I usually use FB's (and most often 0402) in the sub 100 mA region.

--------------- The muRata BLM_P series claim to do a lot of current. (BLM41PG600SN1 claim s 6 A, for example.) But I wouldn't be surprised if they do saturate. Heck , they must by necessity. But how much? The manufacturer claims they are designed for such service. I did not note any info about impedance versus current from muRata. Hmmm... guess I better measure them. At least TDK ha s the info.

muRata catalog statement: BLM--P/S series - For Large Current BLM-P/S series can be used in high current circuits due to its low DC resistance. It can match power lines to a maximum of 6A DC.

Reply to
Simon S Aysdie

I also found out early in my career that a meal heavy on the garlic tends to keep meetings short.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

Have you guys ever tried the Pecorino from Costco? It's hard to stop eating.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

As Phil said, you need to select the correct part for your application. For example, this ain't bad for a 1206-size part:

formatting link

You still have 10ohms at 100MHz at the full 6 amps. Ferrite beads are not all that useful at 1MHz, they are meant to suppress radiated EMI and they do that quite well.

Also, one usually wants low Q in a bead. Except in the ones you buy for the girlfriend :-)

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

10 ohms? For on-board resonances, that's probably about right for damping traces.

Maybe not so much for resonances between bypass caps and such that occur at lower frequencies, but also at lower impedances. But I wouldn't want to use a variable inductor for that anyway.

Probably won't do too much for, say, filtering switching noise out of a sensitive area. The zero bias curve would be fine for that, but since most manufacturers don't publish bias curves...

As for cable resonances, not much out of a, say, 50-200 ohm resonance. Enough to turn sharp spikes into dull peaks, which might get you passing, but it's hardly a good solution.

What if you don't have a way to test full load in the lab -- maybe it was designed for X peripherals that draw 1A, then later on, Y, Z and W are added on and now it's doing 6A? Or...you simply forget to test all the edge cases?

100 ohms (mostly lossy) would be nice to bring peaks down to dull humps. Using the "search inductors instead" strategy, I don't even see any rated for amperes that are that small. But anyway, 10 ohms at 100MHz is all of 16nH -- might as well use a piece of wire with a resistor across it! You can very easily build one of those into the board (though with a much larger footprint).

The smallest value I see is,

formatting link

4040 is a good bit bigger than 3216 (metric). They don't provide Z/F curves unfortunately, but I'd be willing to bet the peak is around 100 ohms at 200MHz or thereabouts. And it'll work for lower frequencies, which means you're still okay on those 10-50' cables (for both conducted and radiated).

There's also these,

formatting link
which are even more expensive, larger in value, closer in size (1212 vs.

1206 -- narrowly beats using two beads in a row), and are spec'd for impedance (well, L and Q, but you can get Z/R/C out of that). Pretty lossy at 100MHz+, though you might want an extra resistor if you need more damping in the 50MHz range. So, it's very close to the same impedance as the ferrite bead under zero bias. Except you keep the same impedance the whole way.

So, in size comparison, both of these are probably better than 2-5 FBs in series, which is still a layout win, assuming you need that much of course. Cost is comparable, though the IHLP one is on the pricey side even considering how many it would replace.

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
Website: http://seventransistorlabs.com
Reply to
Tim Williams

Good ones do.

Sure it works. The trick is to also have a small ceramic cap right where things leave the enclosure. Together with a 10ohms residual RF impedance in the bead that can easily knock down a nasty bunch of peaks by 20dB or more. I do a lot of EMC work.

If push comes to shove I don't have a problem schlepping a fat car battery into the lab. That can do 100A if needed.

Django never forget.

Yeah, but the ferrite is lossy and that's part of the game. Else EMI is like a water bed. You press here, goes down, only to come up at all the other spots.

Those inductors are for power converters, not EMC. And 86 cents is usually prohibitive for EMC measures.

But it's a power inductor, too expensive, and probably not lossy enough. You don't want to just reflect EMI back into the box. Then it would just leak out somewhere else.

How is a >10x higher cost comparable? Ferrite power beads are less than

5c each. 
--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.