Audio Pcb Layout Techniques

Hi All,
I've designed a few pcb's for myself in the last year or two, and as happy as I've been with the results, I really have no way of measuring success nor a framework to think about improving designs. As such I suspect I'm really fumbling around in the dark and there's improvements to be made on what I do. I'm completely self taught in this, so I've probably developed a few random methodologies that would do the be debunked.

In terms of learing it myself, iterating through 6 versions of the same idea to explore the differences isn't a practical option. So my basic question is

What are the fundamental ideas behind quality pcb design?

I could attach a few pcb layouts I've done as example pieces to discuss if anyone is willing to constructively critique the improvements possible (they are all single sided as I make em at home via toner transfer). So if this is helpful, please say so. I would imagine working through a few examples would benefit anyone wanting to learn this art.

To date, all I try and do are
1. minimize length/proximity of presumed current flows
2. minimize parallel traces/maximise orthogonal crossing of parts
3. er...

So, advice or links are appreciated. Thanks in advance.
 
Last edited:
asprinv.

It's late, and I want to do this topic justice, so I'll mull over it and write a long writeup, a paper so to speak, for you and whoever else is interested.

I am qualified to speak to this matter.

I am an electrical engineer, with a graduate degree in electromagnetics (the discipline most conducive to printed circuit board technology) over two decades of experience, and the top recognized subject matter expert in printed circuit board physics at my company, which is a multinational commercial, industrial, and military electronics giant. I have published multiple technical papers on noise, have developed a novel, proprietary PCB routing technique for integrating ultra-high speed baseband controller circuits into super-sensitive RF transceiver sets, I give lectures on the subject and teach classes on this topic at my company, and hold several US patents on PCB and noise-related technologies.

I am a PCB and circuit noise expert. (also the technical director of our multi-million dollar noise lab) I get "called in" and flown out to design centers around the world to fix PCB noise issues when entire teams cannot (in industry jargon, I am a "cleaner". Think Pulp Fiction. Cleaners clean up the scene of a crime (a broken design that's late for ship), remove the body, clear evidence, and make the crime scene disappear. "It never happened".). It's my bread and butter and I can definitely talk to this.

Give me a little while and I'll write something up that hopefully all can use. You'd be surprised at the little tricks that you can employ to reduce noise and crosstalk on a PCB.

So many experienced and seasoned folks here have helped me tremendously with speaker design concepts and the physics associated with them that I'd be happy, actually enthusiastic to "give back" for a change.

This is my element. :)
 
Last edited:
Frankie Carbone,

(It's very late. So this might not come out right. But I might not have time to get back here, soon-enough, so here goes...)

Wow. We have been hoping for someone like you for a long time. It sounds like it will be a great privelege, and of great benefit to us, to receive even just a few paragraphs, from you.

There are a few things that many people here very-often naturally get wrong, which I hope you will also touch on. One is unwittingly making what I would call magnetic antennas, or maybe they could be called "Faraday loops" (after Faraday's Law), by separating conductor pairs, often both the incoming AC Mains pair and transformer secondary wire pairs AND the input signal and input signal ground pairs. And having quiet and noisy grounds share conductors is another very common oversight, here. RF incursion is another mostly-overlooked area.

I am sort-of worrying that it might be a little difficult for you to translate your knowledge to our rudimentary available construction techniques, since you are probably very-accustomed to always having state-of-the-art construction available. The home-made toner transfer PCBs can almost-as-easily be made two-sided. So you can include that as a possibility. But it would also be extremely nice if you could somehow relate as much of your knowledge as possible to as many DIY construction types as possible, even if just in passing; e.g. no PCB, 1-sided home-made PCB, 2-sided home-made PCB, and, pro-made multi-layer PCB, which is much less common here, especially for one-offs.

(However, I am still also planning to try my "multi-layer" PCB "sandwiching" idea, at home, eventually. e.g. Design four or more 1mm or thinner pcbs (probably mostly 1-sided) that will be stacked tightly together at the end, with large-enough holes where needed in order to solder on inner layers, and etched-away areas for inner pass-throughs, etc etc. The benefits of having two power planes and a ground plane, and of thereby freeing up a fourth layer's layout to have only signal (and probably also ground), should make it well worth it, AND should also make the layout simple-enough that it won't be nearly as complicated as it might at first sound, especially for very simple circuits such as chipamp power amplifiers.).

Note that our typical home-made PCBs have no plated-through holes, and, any vias must be hand-made by soldering wires through pad holes. Among other things, that means that even with a two-sided PCB, we can't use a component-side connection to a large radial electrolytic, or under a terminal block, et al.

If you get fired up, you could probably become a huge hero simply by showing people here an example of, say, an ideal two-sided layout for an LM3886 chipamp. And there are certain very-widely-used PCB designs on this site for which any suggested changes might be of great benefit to many people.

Sorry if I have asked for too much, or "grabbed the wheel" too hard. I am just excited by the possibilities.

Regards,

Tom Gootee
 
Last edited:

dfuhr7gm

Member
2012-02-11 11:43 am
Hello,

I too have to say, we all can be very lucky to have a person like you in the forum.
The thread's title immediately caught my attention, as I am a student in a sort-of technical high school (mechatronics dept.) and our electronics-lessons seriously lack a good education on pcb design. I started gathering that knowledge by myself, but somehow that's an impossible task without having someone to explain it.

I got a script about EMI scanned, which I received from one of my teachers, but there wasn't much interesting stuff in it. As an example, I couldn't find a single word on how to share quiet and noisy grounds and how to connect them.

If wanted, I'll upload it. Please keep in mind that it's written in german.

Best regards,
Daniel
 
Wow. That sounds exceptional. :)

Frankie Carbone - Thank you.

Sounds like you are indeed the one to speak to this topic. I (and I suspect many others) look forward to your contributions.

I'm very much in this hobby to learn and stretch my understanding and mind, so your proposed write up sounds like some excellent grist for mill! I'm looking forward to the journey.

Cheers,
Aren.
 
I bought that book a while back and it is great. And by the way, marce is also extremely knowledgable about PCB design.

You both should also do some searches for on line material by Henry W Ott as well as Bruce Archambeault, Howard Johnson, and Eric Bogatin.

Regarding sharing of conductors by quiet and noisy grounds: Simply avoid doing that. Use separate conductors for them, to connect to the star ground (asssuming you don't have planes). Otherwise, the voltages induced across the shared conductor's inductance (and resistance), by the noisy currents, or any dynamic (or large) current, will appear at the ground reference points that were supposed to be quiet, and will, for example, arithmetically sum with your amplifier input signal. Remember that even a small-amplitude current through an inductance can cause a relatively-large-amplitude voltage, if it is changing fast-enough.
 
Last edited:
You could also order a small stack of double and single sided FR4 PCB blanks, maybe 4x6 inches each, and experiment as you read. A high-speed rotary tool (like a Dremel) can remove narrow areas/lines of copper, to make planes and traces very quickly. And the toner transfer method can get a PCB pattern from computer to copper in about an hour, or a bit less after you are experienced.

I think that you will find that a single-sided PCB is the MOST-challenging PCB design environment of all, because it makes everything so difficult to get working really well. And routing traces on them is also a pain, just by itself.

I guess that if you want to DIY your PCBs, which is a great way to go for a lot of circuits, it might be best to use THIN two-sided boards, and leave either the entire bottom or top layer for ground plane. (But I still like my idea of stacking several thin boards, to get more than two layers. I need to make some time to try designing a sample layout that way.)

Do you have an oscilloscope (and proper probes with proper high-frequency accessories, e.g. shorter ground leads, and maybe even special two-prong tips with both signal and ground just a few mm apart, for example)? Those can be extremely handy, especially if the scope goes to 100-200 MHz or more. Used oscilloscopes can be extremely cheap. Even brand new ones (e.g. Rigol, from China, starting at about $329) are not bad and are extremely economical, now.

(Personally, I'd also really love to have an HP 3577A network analyzer, which covers from 5 Hz to over 200 MHz (Yes, they go down to 5 Hz.). But unfortunately those are NOT cheap, even used and quite old. I was about to buy one, anyway, WITH the optional s-parameter test set, about a month ago, but then suddenly found a pair of like-new Magnepan MG-3.6/R speakers for under $2000 and bought those instead. :) )
 
Practical progression

Hi Gootee,
I've been etching pcbs as I go on various projects using the toner transfer method. I've mostly completed a small cnc router, with the intent that I can whip up a layout in eagle and then output code to cut it for me. Hopefully once thats working properly I'll have a workflow that allows time efficient experimentation.

I actually really enjoy the circuit layout side of the designing process. It intrigues me like a complex puzzle that actually has no perfect solution but can be tweaked to be 'best' in a number of different ways. I'm really happy putting hours into what I hope is a good design - although explaining why I think it's an enjoyable challenge is harder :) I guess I'm looking for the 'rules' that guide this part of the process to the best possible result given the constraints.

In a perfect world I'd just build multiple iterations of whatever I'm trying to build and compare empirically as to the improvements, but cost and time prevents this being a viable way forward for most projects. Hence me looking to develop a theoretical background on which to base design decisions on.

In terms of testing finished builds, I've a few older oscopes around - the best of is a 50MHz unit which appears to only have one working channel. Standard probes and ARTA on a pc for testing amp distortion into dummy loads. The sort of equipment you're talking about is beyond my needs, and sadly beyond my understanding as yet :)
 
Last edited:
In a perfect world I'd just build multiple iterations of whatever I'm trying to build and compare empirically as to the improvements, but cost and time prevents this being a viable way forward for most projects. Hence me looking to develop a theoretical background on which to base design decisions on
With CAD you can try different iterations on screen, and depending on the software, the type of design can simulate the layout. Mainly for high speed, but with analogue you can use spice and as long as your layout parasitics are not excessive will get a result very similar to your spice sim.
 
Short tracks kept as thick as possible is a good start.

I fell flat on my face with a audio mixer pcb when there was hum everywhere getting into the audio stream.
After some investigation it came from the power supply.
I had simply mixed the audio and power supply grounds willy nilly.
When I completely separated the grounds and joined them once at the edge connector all the hum disappeared.

I have applied this method to all sorts of audio even valve circuits and had some great results. There is nothing worse than listening to hum and noise between tracks.
 

Especially marce but others too. Please keep it simple. :dunno:

I also want to design my first audio PCB but the links overload from all of you is too much for a beginner DIYer . If I have to read all those books and all those papers, understand all of it then perhaps in the end I should open a shop and start doing it professionally :cool:.

There are some confusions even over the basic concepts. I am still unable to understand the ground plan for a DAC. Two separate ground planes (digital & analog) attached only at one point under a specific pin of DAC, one ground plane with segregated portions, ........I have read arguments for both and some other techniques. For example, here Grounding of Mixes Signal Systems Henry Otto favors single segregated ground plane but then I see many experienced members using and advising two ground planes? What is the final consensus?

Can we please have a mutually agreed upon list of 25 Do's and 25 Don'ts (with diagrams, pictures) for audio PCB from experts which if any member follows will make the project at least 75% perfect? :)
 
There are some confusions even over the basic concepts. I am still unable to understand the ground plan for a DAC. Two separate ground planes (digital & analog) attached only at one point under a specific pin of DAC, one ground plane with segregated portions, .......

You'll never reach consensus on that because DAC chips themselves differ. I prefer to use cheap 8-pin DACs in my designs where there's no separate analog and digital ground connection - just to give one example.

Can we please have a mutually agreed upon list of 25 Do's and 25 Don'ts (with diagrams, pictures) for audio PCB from experts which if any member follows will make the project at least 75% perfect? :)
You'll first need to say what 'perfect' means to you. Perfect sound (forever!) or the best possible measurements? Even a simple 'Do' will attract detractors - let's try one for starters : 'Do establish separate signal (quiet) and power (noisy) ground connections' :D
 
Single ground plane with some segregation, and creative splitting, or two ground joined at one point, or three with a power ground.
I am very busy at the moment doing PCB's day and evening, but I am in the process of putting some basic notes together. The stuff I have posted is what I post on a couple of engineering forums for people who ask about PCB design, I consider them the basics.
There are no simple do's and donts for PCB layout, each design is different start your design and I will try and help (and its for free on here:))
(Audio layout is analogue layout, using a DAC means you will have some digital layout.
As to PCB experts, you have me as a regular contributor, and I am a PCB expert.
I will contribute more later...