PCB / Schematic Advice?

Hi all!

I made a high powered audio amplifier schematic and PCB, but before I go through all the work and money of building it, I would like some opinions and ideas. I have tried posting on diyaudio.com but didn't get anywhere. Does anyone know of a forum that would be suitable?

I would post them on here, but I have gotten yelled at before for posting something "not arduino related", although I do 99% of my projects with Arduinos (and the schematic / PCB includes a Attiny261 with an arduino boot loader).

If anyone would like to see the schematic / PCB, just say so.

Thanks in advance!

Just being curious... What is the task for the controller?

The Attiny261 has an infrared sensor, and is attached to a TDA7418TR. So it is programmed to control bass, mid, treble, volume, etc. etc. from IR control inputs.

It is usually best to post in the forum of the CAD product you created your Schematic and PCB.

Thanks. Good project.

1 Like

Hi, @kgray9

I agree with @JohnRob, you will probably find more experienced advice from those that design for analog/audio situations.
Earth loops and power distribution over a board come to light as possible unforeseen problems going from prototype to final product.
How High is High?

Tom... :grinning: :coffee: :+1: :australia:

1200W@4Ω! Very high power :grinning:. It's based off of 2x TPA3255.

I designed it using EasyEDA, but aren't sure if I should ask questions like that on there.

Yeah. I'm trying to minimize those problems as much as possible, before they become problems in the audio output.


Okay, I hope you have looked at TI's application notes, they have an evaluation board and the app sheet shows the component layout around the IC.
Looks very helpful.
slou441.pdf (2.1 MB)

Tom... :grinning: :+1: :coffee: :australia:

General electronics seems like a better place for a question about, err... general electronics.

Post away! The description for the general electronics section is

Resistors, capacitors, breadboards, soldering, etc.

Seems to me that includes resistors, capacitors, breadboards and soldering things not necessarily connect to an Arduino.

Thanks @TomGeorge

I followed the data sheet's PCB layout and schematic with all recommended parts.

Ok, thanks!

PCB top and bottom layout / wires / etc. :



Even though you didn't use Kicad, I would post there. The forum has a number of professional PCB designers who have always been helpful.

However before you post you should do a little homework on how much current a piece of PCB clad can carry. Note many "hobby" PCB's are only 1 OZ "thick" you may need 2 or even 4 OZ copper. 4 is not very common.
The failure mode is the copper gets hot, at some point it exceeds the "glass transition point" of the PCB at which point the copper is no longer held to the board and will start to life then the copper gets even hotter leading to thermal runaway until the copper opens.

Does you power device have a "recommended" design layout? If so have you followed it? Do they mention the copper thickness.

Copper Thickness explained:

Copper is specified in Oz per square foot. This is really not a thickness but it is what the mfg can control. You can google it and get an estimate of the actual thickness.
Most current capability webpages will deal with copper in Oz.
I'll bet the Metric folks really hate this "Standard"

1 Like

A ground pour on top (component side) is an out of date thing to do; it is not of any significant benefit to do so and can cause problems.

Long traces on the foil side (ground plane) should be avoided.
GND straps over such traces (on the component side) should be incorporated.

Thanks @LarryD for your advice!

So... what I need to do is:

  • Erase the top copper pad
  • Instead add "straps" over the longer traces on the bottom side


Cross-unders should be short too.

You are going to think I'm crazy (and you might be correct). I'm guessing you are new to power electronics and perhaps new to electronics. So I don't to burst your bubble but you are going about this in the wrong way.

Its real easy to put all sorts of parts on a schematic and then lay them out on a board. But making them work is another story.

I'm not trying to be mean but your design won't work. It will not even come close to working. Below is the suggested layout from TI.

Lets look at the currents:
I'll assume you do make 1200w. 600W for the sub and 300W each channel.
Speakers = 4 ohms.

Power = AMPs squared / Resistance

300 * 4 = √Amps
UPDATE: 2/14 calculation should be √(300/4) = 8.7A average = 12.5 A peak.

Amps = 34 amps(RMS) Approximately 15 times this and you can start a car.
Note amps (RMS) is nearly the same as average amps. Peak amps for 34A(RMS) = 48 amps(peak)

Here's where you are going to think I'm crazy. If I were to take this project on I would first build a board with a single TPA3255 with the outputs and a few direct inputs (i.e. no base or treble just a volume pot).

When its done, I would test it and verify it performs without issues. Then after I make the changes I learned by testing this board I would consider the full board.

I've not used the TPA3255 before but I have made a number of high current switching power supplies (basically a class D amplifier but without the audio modulation).

Components are another area that needs a lot of attention. ~~ Looking at your 3D inductors I doubt they will handled 34 amps.~~

I know this is not what you wanted to hear but ...

1 Like

Wow. Thank you @JohnRob !

I really appreciate your advice, even if it means making a prototype board first, and redoing my current PCB layout. :upside_down_face:

I'm new to power electronics, but have made lots of sensor PCBs and small project PCBs.

The inductors are rated at 10 amps. I guess I didn't realize how much power I was talking about :face_in_clouds: .

I'll make a smaller PCB with nothing except for the amplifier and a volume pot, that is identical to the data sheet's layout, with inductors that are capable of 50 amps(?) .

Is 30A for the inductors ok @JohnRob ? The highest I can get from my main supplier for 10µH is 30 amp.

Not if you are expecting to get close to the 300W output. Audio is a funny thing, music will have a short "loud" passage then "normal" volume. If the instantaneous peak demand exceeds the amplifier demand you will get distortion. I've never played with class D amplifiers so I can't describe the audible effects.

Regarding the inductor(s)...
*** Inductor cores can "hold" a magnetic field. They will act as an inductor as long as you are adding magnetic energy. However once the inductor core if full (aka saturated) the inductance goes to 0 and it becomes just another piece of wire.

I know that air core inductors will not saturate.
This is an area you will have to research some more, I can't recommend off hand a component that will positively work.

Google TPA3255 output inductor. There must be many others who have gone down the same road.
And read this thread and all the TI application notes.

*** This explanation is simplified to help someone with no prior experience gain an understanding of inductor core limitations. I don't want to hear some purist telling me my terminology is wrong.

1 Like

Thank you @JohnRob !

Dang, I thought I just got the traces right for one IC. Well, how do the traces look on this?

Some notes:

  • The traces from the inductors are 170 mil wide
  • The traces from the IC are as wide as I could make them before expanding them to 170 mil
  • The vias are 24 mil wide with a 12 mil hole
  • The traces to the 4x 0603 capacitors are 16 mil wide
  • Those are the 30A inductors.

That makes sense. Thanks for the explanation! Maybe I could wind my own air core inductors with 12awg wire? But I'm not sure if from them not being perfect, the audio would be distorted or such. Ill look some inductor stuff up for the TPA3255 and see what I can find.