Inductance in PCB Layout: The Good, the Bad, and the Fugly

Unwanted inductance in PCB layout is a fact of life, be prepared so at least it's not unexpected.
Read the article on Hackaday: hackaday.com/?p=325221
Music:
Latin Industries Kevin MacLeod (incompetech.com)
Licensed under Creative Commons: By Attribution 3.0 License
creativecommons.org/licenses/b...

[@arduinoversusevil2025]

He's back! Thanks for sharing Bil.

[@electronash]

"Focus, you..." Oh, nevermind. One of my life goals is having a pint and a tech discussion with Bil.

[@randydireen3566]

AvE are you a god? You exist everywhere on KZbin.

[@cncmechtech]

If I had a plate of that squirrel yakisoba to eat while watching Bil... Magnifique...

[@covishen]

@@randydireen3566 Ave is the God of Canada! With the possible exception of Mike Holmes.

[@xl0xl0xl0]

Please Hackaday, more of this stuff, less blah blah!

[@w2aew]

Yep - I have all three of those books - and each are well used and well worn! Good examples with the PCB image current visualization. Here are a few videos I did on related topics: Power supply decoupling: kzbin.info/www/bejne/b3bEhZeZZ9KlaZY Inductive kickback (relay coils): kzbin.info/www/bejne/mWesaIyZl9treJo Self-resonance of caps: kzbin.info/www/bejne/rJqVZYamgL6MpbM Scope Probe ground lead inductance: kzbin.info/www/bejne/sKDHoXarrdyhldE Homemade high-performance probe socket: kzbin.info/www/bejne/Y2XUaZqbepqYnJo

[@sbybill3271]

I wish I had someone tell me all that in my university days. In fact, my teachers did but not in this amazing and simple way. Excellent, keep up the good work. It's making a huge difference.

[@alexanderquilty5705]

I love Henry Ott’s text on EMC! One of the cool things I learned in Cabling is that there’s 3 separate currents flowing in a Co-Ax cable depending in your termination! Cool stuff

[@remotecoderz]

Amazing analogies Bil, I actually understood what you said! Now if I can keep what I've learned stuck in my gray matter, that'll be a miracle. Thank you!

[@josuelservin2409]

Nice and useful video for sure, thanks Bil. And I would like to see more updates on the bionic prosthesis.

[@nenharma82]

Awesome stuff! Helps a lot with the understanding of these sometimes fairly abstract concepts.

[@GameTechRefuge]

Great video. Has helped me grasp the concept of inductance. For a hobbyist, just starting off like myself. Some concepts and terms can seem scary but the online community in general are great. Thanks again for the videos.

[@KanalFrump]

Thank you Bil! Great explanation of inductance.

[@LupusMechanicus]

Diamond in the rough, thanks Bill.

[@st3ddyman]

That’s an excellent video, I learned a lot. I have that last book but this was far more understandable. More please Bill.

[@jaimemartinezdiaz1004]

These vids are so informative. Thanks for sharing your insight!

[@Stelios.Posantzis]

That's a great, short, practical introduction and demonstration of the principles one needs to have in mind in order to understand circuit track inductance and especially for PCB tracks. It's very succinct and quickly evolves from basic concepts to the particular application at hand in "layman's terms". It certainly provides a good, short explanation to the "why's and how's" of ground planes. Fantastic! 16:10 : I was racking my brains until it eventually dawned on me that "board width" does not refer to the PCB board thickness but to the distance between two tracks. Therefore "track width" is exactly what it says. Sometimes the wording in a particular excerpt can be read the wrong way and it gets stuck in one's head. This is especially common if the excerpt is read in isolation from the rest of the book/chapter/etc. as each author (and reader for that matter) has their own particular "style" as to the choice of words needed to express a recurring concept.

[@berni8k]

Now that is an excellent explanation! Yet so many PCB designers don't realize the important of this. Another side effect of not keeping your return paths tight and planned is the PCB will also radiate a lot of radio interference. If there is sensitive analog circuitry near by or some sort of radio equipment you can have a really bad day.

[@MatthewLong8]

Great advice. If you interrupt the ground plane, have any crossing signals go around, or alternatively perhaps jump the ground plane on the same layer as the signal. Yeah I know the vias add inductance but so does the increased path length. To really know which to choose you probably need one of those fancy modeling softwares as the correct answer almost always is, it depends.

[@garbagebag185]

I have this line filter for a treadmill that was a huge inductor. I connected it to a relay and made a pretty cool shocker. Might have damaged the relay.

[@kd1s]

Yes indeed I remember X sub L and X sub c - from my study of phase angles for my Amateur Radio Advanced class license study.

[@BarriosGroupie]

Don't forget the legendary Linear Technology application books by Jim Williams. Lot's of great stuff on high-speed probing etc.

[@meetyoursathish]

Good... like your efforts to make practical examples 👍

[@nicopistor62]

Great job, thank you!

[@williefleete]

the boards I designed for my 8 bit discrete logic CPU probably arent the best in some places, in terms of ground plane impedance, loop area etc but they are likely a lot better than the prototype boards I used previously which had the power traces interdigitated. might try to keep some of this in mind next time I revise boards or make new ones

[@aliuzel4211]

Yes. This is the great book !!! Howard Johnson.

[@funkyironman69]

High Speed Digital Design is a great book, I've read it. Still need to get a copy of Advanced Black Magic.

[@edgeeffect]

"A handbook of black magic" made me laugh.

[@shinzon0]

THank you for the book tips :D

[@hasanalattar9561]

so what will happen if you change signal to bottom layer and you have two ground planes stitched (bottom and top)?

[@amoghjain]

Thank you soo very much!!! I was just wondering, if ANY signal, going from A to B is EVER one way? All signal have to return back to complete the circuit... . Am i right? for eg, if I send a trig pulse from a microcontroller Pin to let say a ultrasonic sensor, the same signal, MUST return back via the ground plane to complete the signal?

[@JohnDlugosz]

Capacitors don't "store electrons". This is counterproductive to visualizing how they work, so I think it's important, not just a nit. They store _potential_ by shifting electrons a short distance within a material.

[@hasanalattar9561]

and thats E field in the dielectric?

[@JohnDlugosz]

@@hasanalattar9561 Yes, there will be an electric field between the region of excess electrons and the region depleted of some electrons.

[@AlienRelics]

That's true. If they only stored electrons, then the dielectric would not matter.

[@JohnDlugosz]

@@AlienRelics You got me thinking about one-terminal Leyden jars. Faraday's ice pail does indeed accumulate a net charge. Such capacitors would not behave the same as a normal cap in a circuit, with respect to passing AC with a frequency-dependent impedance.

[@AlienRelics]

@@JohnDlugosz The outer covering is part of the capacitor. If you don't have an outer covering, then the capacitor is the internal conductive substance (saltwater or a layer of foil) and the dielectric is the pail and the air between it and the surrounding earthed structures.

[@nickpelov]

That was very useful! However inductance is not always bad. What if we have a regulator that can't keep up with fast switching devices like microcontrolers and other digital stuff? Low resistance/inductance traces for power supply will make the things worse. Switching components will add noise to power rails. If you have longer traces for the positive supply rail combined with decoupling capacitor will reduce the noise that goes back to the regulator and the other devices. Maybe that's not the case with the ground because ground as you said is also used for high speed signals. The ground plain that you showed with the signal trace - I didn't realize it matters. So is it better if the trace goes around the connector instead (so it stays close to the ground plain)?

[@UpcycleElectronics]

Awesome. Thanks for the book refs. I hesitated for a sec before watching as I have 5 designs ready to etch my first prototypes and am just waiting on fresh Riston. I'm always worried I'll watch some facepalm info that scraps weeks of KiCAD maze puzzle games. How about a beginner practical explanation of single sided w/jumpers parasitic issues tips and/or simple double sided home etched basics with typical 8bit uC/10bit ADC designs. I've also been wondering about filtering for uC power pins with ferrite beads. One of the OSHW project designs I'm working on uses a basic AVR's ADC. As a simple hobbyist I don't know where to start with a solid AVcc filter. I just grabbed a known topology for the current design, but would like to know more. That same design has several different power rails. I have no idea about the general rules to follow regarding routing them. That's the one issue I'm most concerned with in my current project(s). Thanks for the upload. -Jake

[@UpcycleElectronics]

@Mdmchannel 0402's all around! Yeah my current project has a ton of places for additional caps as it's a dev board for the Atmega644/1284 version of the AVR Transistor Tester with a main 5v Vcc precision rail (MCP1702 LDO), auxiliary 5v and 3v3 rails, and the raw battery/Vin plus a uC controlled power management rail. I have the whole uC broken out in multiple locations with most power rails distributed accordingly. I'm sure it will have lots of issues with the first iteration(s).

[@UpcycleElectronics]

@Mdmchannel Thanks. The general Vcc reg is powering the whole AVR, the auxiliary 5V is mostly intended for a zener test circuit that needs noise isolation but it is also routed away from all of the analogue/tester circuit. I am concerned about running several rails in parallel, but they are all around 7cm away from the important stuff. The uC is in the middle of a 5cm × 7cm standard protoboard with everything relevant broken out on the 2 × 24hole sides. (I've turned a 5x7 protoboard into a shield that plugs into a single sided etched board). I also broke out each uC port on individual 3×7 protoboards. The power rails are all routed to the opposite side of these 3×7 boards. Assuming everything goes correctly I will post it on the official mikrocontroller project thread and the eevblog one too, I might even write it up on hackaday now that I think about it. This is probably the worst AVR TT design ever. I just wanted something to play around with all of the project options. Even the power supply is on a protoboard breakout. So technically this thing has no supply just terminals for rails. -Jake

[@johnnywalker2947]

I have a few unpopulated military pcb's. I could never find out what went in these two places (symmetrical places). The backside is 1 millimeter thick copper. The footprint is not easy to find. I could send you a pic in about a month. I've always have been curious what went there. I have some hermeticaly sealed voltage regulators which are absolutely gorgeous and yes you can connect them in parallel for more current handling . I wish I could sell these. I also have at least 200 Tantalum capacitors. 35 maximum voltage I also would like to sell.

[@10e999]

I would love a book review . What about "rigth the first time""?

[@arleygamer12]

Great video. While working on my graduation project I found "RF Circuit Design" by Richard Chi Hsi Li, it studies layout design (among other stuff) and gives some recommendations depending of frequency for PCB design. If you are interested in RF or Mixed Signal PCB design you should definitely check it out.

[@josephemmons6749]

Well I'm glad I stuck with it because there's a solid amount of information at the end. It is presentation quality but kind of slow Still worth it.

[@sevasbharath]

On your 1 inch impedance table, the fatest rise time is 1ns and not 318ns that you stated.

[@Mrav79]

I've got a C128 with your name on it! I mean literally, it's printed on the PCB.

[@leppie]

10:41 I did this (9V over a step down transformer in reverse) in high school (around 89) when I discovered it accidentally. I literally shocked my classmates :D

[@Noone10011]

Reminds of an instructor I had. The first thing the instructor did was ask if anyone is a perfectionist. He then said get over it. Nothing is perfect in electronics.

[@funkyironman69]

Just bought a copy of Noise Reduction Techniques in Electronic Systems for $20 US, inc shipping to Australia :)

[@lextr3110]

any prototype board you would recommend?

[@berni8k]

For RF nothing beats a copper clad board with dead bug style construction on top. i.stack.imgur.com/7Hr90.jpg For everything else i am a fan of perfboard where you only get a grid of unconnected round pads and you are free to wire it however you want. Its still not ideal, but it lets you plan your ground. For really high speed you likely have SMD chips anyway so better to get a cheap PCB from china for $5.

[@lextr3110]

@@berni8k oh thx a lot..

[@criznach]

5 minutes in and this is brilliant! Thanks Bil. See you all at Supercon?

[@restcure]

It's OK - I'm sure the neon bulb and coils were put someplace where you'd be sure to remember where they are ... that's where _my_ stuff is.

[@m8uwut]

Would it be more correct to say caps and inductors *impede* change and not resists? That would seem to be more consistent with other literature.

[@robertw1871]

Mark Ivey Impedance and resistance are equivalent, the word impedance is more commonly used for AC signals and resistance is more common for DC. There are no complex mathematical components at steady state DC, so it’s more simplified, for AC the phase lead or lag between voltage and current makes it more complex, but at the end of the day ohms law still applies, 10 ohms is 10 ohms regardless of what is causing it. Reactance, reluctantance, impedance, resistance are all the same things just different words for different contexts... the only thing to keep in mind is the context, AC will have two components, a real and an imaginary part, this is just due to the mathematics used.

[@yoramstein]

Right to say is that Inductors and capaciots react on changes (frequency). Inductors develop more reactance when frequency goes hogher Xl= 2 pi fL and capacitors the opposite Xc= 1/ 2 pi f C when f goes higher Xc goes smaller.

[@timon781]

Heh, someone discovered Andrew Kramer :)

[@robertw1871]

Just head down to the physics store and pick up some inductance free wire, problem solved.