Class D TPA3122D2 Single sided, all thru hole components

[Zaaphod]

I thought I would try out a class D amplifier, but like many on here, I'm not too fond of the idea of surface mount components. I may get into them at some point, but I thought I would at least try a prototype Class D based on the TPA3122D2 20pin DIP. I want to try to keep this simple, so single sided PCB and no surface mount components. I have the chip configured for BTL into 8 Ohm speaker, I will supply differential inputs. The schematic is directly out of the datasheet, except I used 1000uF capacitors instead of 470uF, just because I have a box of them, and the datasheet says they should be 220uF or larger.
I went ahead and put pull up and pull down resistors on Shutdown and Mute, just so I can play around with them, but I will probably just hardwire mute... I may use shutdown.. it would be nice if I had a circuit that would detect the presence of a differential signal and turn on the amplifier, and keep it on for say 3 minutes after the signal is no longer detected. very easy to do with a single ended input with a transistor, I haven't quite figured out how to do it with differential inputs without using an opamp... although I guess an opamp for this purpose would not be all that bad... but then of course the opamp will need to be powered to be able to detect the signal. I'm just trying to keep it simple and small. Anyone have any ideas on how to have automatic power up and shutdown based on the presence of a differential signal? I don't want to UN-balance the differential feed with it, and would like it to be very high impedance.

Here's the schematic and preliminary circuit board. I ordered my parts and will probably breadboard it first. Any advice on this would be appreciated.

[marce]

Start again.
Tracks are way to thin, component plaqcement requires optimising. The high switching currents require proper planes.
In fact follow the layout in this app note:
http://www.ti.com/lit/ug/slou214a/slou214a.pdf

[Zaaphod]

Originally Posted by marce

component plaqcement requires optimising.

Can you please tell me what you mean by optimizing?

I was going off the datasheet which states that all the power supply and bypass capacitors should be as close to the device as possible, so I put them all right up against it, then made everything else fit.

Thanks for the link to the application note

[marce]

A quick note:
Have a look at this thread, the links cover both class D and SMPS's
Switcher EMC design

I would copy the Texas layout positions, onlt replaceing them with your through hole components, so you keep the relative positions, but the size and distance between components will change to facilitate your different size packages.

[Zaaphod]

I don't really want to just copy the eval board, I have some things of my own I want to incorporate into it, and I also don't need all those test points they have everywhere. I've been looking at some other designs, most use surface mount parts, but I think I can do something similar with thru hole parts. One design I really like was by theanonymous1 seen at this link:
http://http://www.diyaudio.com/forums/class-d/124318-my-tpa3122d2n-btl-proto.html

here he used a double sided board but the entire top side was a big ground plane. I think this may be a way to go since with this method it would not be very much harder than a single sided board.
here is my schematic so far.. it has a second set of inputs which will use a remote power on signal to power up the amplifier using a photocoupler. I think I'm still going to need something on the input signal to make it work the way I want it to, but I have to start somewhere. I have 3 ideas for circuit boards, I am not sure what would be best. All these planes and current loops and such have me all confused, so I figured I would just post my ideas and hopefully someone with more experience than I have will be kind enough to offer me some advice. I will post them each in a separate message to make it more clear which description goes with which images. The board I am working on is just the TPA3122D2, the DRV134 will be another board at the opposite end of the cable.

[Zaaphod]

Here is my first possibility... Single sided, I moved everything around and got rid of the wires sneaking between the pins on the IC and also made all the traces a bit larger, kept everything as short as possible. I found some capacitors with 0.1" lead spacing so I could put them right up against the IC pins with no trouble. I also incorporated large ground and VCC planes. I will need a jumper wire to connect the interior gound plane

[Zaaphod]

This one uses the solid ground plane on the top side of the board, bottom has no planes, just traces. I have some feed through holes to make it possible to solder the electrolytic caps to traces on the bottom of the board, yet still connect to the solid ground plane one top.

[Zaaphod]

This one uses a solid ground plane on top, and a huge VCC plane on the bottom. it's hard to see whats going on so I am posting the top, bottom, and combined layouts

[Zaaphod]

Here's a PDF file with larger images, may be easier to see then the small pictures in my posts.

[marce]

Its looking nice. When you compare them to your first design you can know see the difference, with the wider tracks and ground plane etc.
Looking good