Decoupling Inductors? Power Integrity in PCB Design

If you've looked into Power Integrity in PCB Design, chances are you've come across the idea of decoupling capacitors, but some will show the use of ferrites or even "Decoupling Inductors." But what exactly are they? And how do they affect the PDN impedance? Zach Peterson answers and explores inductance, impedance, and damping.
0:00 Intro
0:36 Decoupling Inductor?
2:00 Typical Power Delivery Network Structure
3:55 Ideal Impedance
4:46 Adding Inductance
6:03 Altering Waveform Damping
8:12 The Problem with These Solutions
9:05 Self Resonance Frequency
10:40 It Only Works on Paper
11:00 The Proven Capacitor Method
11:36 Special Offer
For more Power Integrity videos, click here: • Power Integrity
For more PCB Design for Advanced Users videos, click here: • PCB Design for Advance...
For more High-speed Design videos, click here: • High-speed Design
Check out the Previous Video, "Should You Put Ferrites on a Power Regulator Output?," Here: • Should You Put Ferrite...
The Role of a Decoupling, Inductance, and Resistance in a PDN: resources.altium.com/p/role-d...
Check Out CircuitMaker Here: workspace.circuitmaker.com/Ac...
Design PCBs with a Free Trial of Altium Designer Here: www.altium.com/yt/altium-acad...
Don't forget to follow us on social to stay up-to-date on the latest Altium Academy content.
Follow Altium on Twitter: / altium
Follow Altium on Linkedin: / altium
Follow Altium on Facebook: / altiumofficial
Ready to try the industry's best-in-class design experience yourself? Download it today and get started! www.altium.com/downloads
The Altium Academy is an online experience created to bring modern education to PCB Designers and Engineers all across the world. Here you can access a vast library of free training and educational content covering everything from basic design to advanced principles and step-by-step walkthroughs. Join industry legends as they share their career knowledge, review real-life design projects, or learn how to leverage one of Altium's leading design tools. No matter your level of experience, the Altium Academy can help you become a better Designer and Engineer!
About Altium LLC
Altium LLC (ASX:ALU), a global software company based in San Diego, California, is accelerating the pace of innovation through electronics. From individual inventors to multinational corporations, more PCB designers and engineers choose Altium software to design and realize electronics-based products.
#Altium #PCBdesign #CircuitMaker

[@lolsypussy]

This lecture series is gold. A collated source of info that you won't commonly find on the internet. Can you discuss in the future the difference when using polarized and ceramic caps on inputs and outputs of dc-dc converters? Thank you.

[@Zachariah-Peterson]

Hello, Thanks for watching! We can certainly discuss the use of polarized capacitors and give a good comparison. Great suggestion for an upcoming video!

[@saeidesekhavati1518]

Thanks for sharing these useful information in KZbin!

[@maks886]

This is so useful, it's hard to believe that it's out here for free :D

[@AltiumAcademy]

It really is!

[@myetis1990]

the question almost ten years in my mind is answered :) thank you, Mr.Zach I'm tired to keep the rule of thumbs, so please dive deeply to clean our obsolete knowledge. to make us understand clearly, do you think to tell this impedance problem on a real-life circuit by measurements? because we need practical knowledge along with the theoretical. (I've never been sure I measure correctly to make a decision)

[@Zachariah-Peterson]

Hi Mustafa, we'll keep this in mind for upcoming videos. Thanks for watching!

[@ericschaffer3400]

For any single given measurement, you should be able to take another to corroborate. That rule of thumb doesn't fade - measure twice, cut once. The cut line shows you understand the middle ground.

[@beamray]

what about feritte bids separating power planes near IC? many designers use it just to separate several ICs using single main pwr lane.

[@Zachariah-Peterson]

Hey Egor, this is another great question from the past that I missed. I think you probably saw the ferrites video, that's where the answer to this question lies; it's used to prevent certain noise frequencies originating in one power plane from reaching into the other power plane. Depending on the topology and the switching rate on the isolated power rail, you might find that this creates new noise if the isolated rail starts switching.

[@carlosgalianarubio3809]

Hello. I can not find the vídeo where you say that a ferrite bead must not go after a regulator. Thank you.

[@Zachariah-Peterson]

Hi Carlos, we just recently recorded a pair of videos that looks at ferrite placement in a PDN from a simulation standpoint. The idea is to show how ferrites are sometimes used to isolate different power rails on the same PDN, usually for high pin count ICs. You can watch these here: Part 1: kzbin.info/www/bejne/e5nPo5uon7Zlisk Part 2: kzbin.info/www/bejne/oZ7XhIiwpJ2clcU

[@cvillafane4694]

Hola Carlos, adjunto otro video de Zach kzbin.info/www/bejne/hnLdiWabm7B1qck

[@RuneBroberg]

Hello Zach! I am but a mere amateur and hobbyist, and I have learned several new things from your videos that I will try to put to good use in the next designs I am planning! However, I do have a question on the subject of decoupling inductors - and particularly in relation to the term "decoupling". My understanding of decoupling has been that it both serves the purpose of providing current for the component it's placed near, as well as providing isolation - decoupling - of the component's effects from power network as a whole! Placing inductance on a (linear) regulator output makes no sense, but what about the input? Not the immediate input, perhaps, but as a part of an LC filter with the bulk and decoupling capacitance often present immediately prior to the regulator? On occasion, I have even used resistors at a pinch when prototyping to get rid of switching noise - noise that didn't cause a problem for the circuit creating it, but *did* cause problems if allowed to travel further up the power supply, and possibly through other regulators. I hope you will find time to address when ferrites, inductors in general and LC filters *are* the right answer to filtering ;-) Regarding split ground planes, a video I watched just before this, I was wondering about separation of ground in cases of having (somewhat) significant mostly-DC currents as well as small AC or DC signals, and whether restricting the obvious path for such *DC* currents could be a reason to create splits or voids in a plane? A current being switched at maybe 1 Hz might not cause EMI issues if switched slowly, but it could cause a voltage drop for *other* signals along the same ground plane?

[@LightningHelix101]

“Decoupling inductors” are usually part of the LC at the input you mention. They are calling EMI filters, and the goal is to make that power supply look like a constant load to the source it’s getting it’s energy from

[@LightningHelix101]

The split planes idea come from this idea that ground is made out copper with certain impedance (not a perfect short) Impedance is lowest at DC however, so splitting the plane would have the least effect on how DC power/signals effect other power or signal lines.

[@RuneBroberg]

@@LightningHelix101 DC signals spread out the most across the plane, tho - and given significant enough current, they will cause a voltage drop between points on the plane, even outside the direct linear path between the source and sink. Such a "changing DC offset" could be an issue for low level AC signals, was my idea, and I wondered if this was a (very limited) use case for splitting a plane, causing a restriction that would make less DC current flow in that particular section of the ground plane? I guess not routing high current DC power/signals through the ground plane, but in separate traces instead, is the more obvious solution ;-)

[@maks886]

Watch Rick Hartley's presentation about splitting ground planes here on altiums channel, it's very good

[@Zachariah-Peterson]

​@@LightningHelix101 Yes they are used in EMI filters at the input to a power converter, also sometimes in a pi filter on the output from a bridge rectifier, you'll see them in some of the better reference designs.

[@timthompson2779]

I've seen C-L-C filters used to isolate analog from digital power supplies. This appears to be done instead of using separate LDO's when cost is paramount and frequencies were probably not that high. Is this not recommended?

[@Zachariah-Peterson]

It depends on what is using the digital power supply and what the analog frequency is running at. The danger is that the digital side creates voltage ripple that is seen on the analog side because it appears at high frequency and will be insufficiently filtered. Also with the CLC filter is that the ESR values on those components create a pole in the filter transfer function. But maybe this is at such high frequency and high enough load that the digital side does not excite a strong transient... Like I said it all depends. If you're using the analog side at or near DC then probably not a problem.

[@saumitradwivedi2784]

What should be the ideal placement of ferrite bread on the PCB?

[@Zachariah-Peterson]

Nowhere. Just about every expert on PCB design agrees that you should avoid the use of ferrites unless you can prove you need them. They cause more problems for EMI, SI, and PI than they solve.

[@Dukel192]

What if you need a point on the power rail to measure the current. Would you recommend placing a ferrite or a 0R?

[@Zachariah-Peterson]

Maybe place a current sense resistor on the regulator output pin and measure the voltage? That would be fine on regulators that are not operating at super-fast speeds to support advanced processors with high I/O counts. In those more advanced systems the current sense resistor could have greater resistance than the PDN!

[@Dukel192]

@@Zachariah-Peterson Thanks for you reply. In my case it is indeed a high-speeds sensor is being powered so a current resistor does not seem to be the right choice. I found several designs powering an image sensor that included an bead on the power line, and even split one output in two power lines with two beads. Would you suggest using a 0R in this case, or keep the ferrite bead?

[@sigfreed11]

I must be confused, switching power supplies always have an inductor in series in the output, are you saying that isn’t a good idea?

[@Zachariah-Peterson]

No that's not what I mean, that particular inductor targets low-frequency switching noise, which is appropriate. I'm talking about using an inductor to try and remove mid-to-high frequency noise as part of an LC circuit with high frequency rolloff in a PDN that delivers current to high speed digital devices. The problem is that the inductor size paired with low capacitance used at moderate frequencies (hundreds of MHz) adds too much impedance to the PDN, and it actually increases transient noise seen at the power pins rather than reducing noise. A lot of app notes for high speed components like FPGAs recommend using a ferrite in this way, but this should not be done because it creates a larger transient voltage at the input pins of the device when I/Os start switching. Better to use more capacitors. If you're just supplying power to a static component, meaning the current drawn by the component does not change or it changes very slowly, then additional inductance on the PDN is just fine as any transient events will be very weak.