Non inverted, 3 x parallel LM3886 schematics and layout.

Long ago I promised to post schematics and the PCB layout of my 3x parallel LM3886 per channel build, so here they are! The PCB has a couple of small 'improvements', nothing major... just a mute jumper so you can attach a switch, an alternate position for a couple of bypass caps and a few trace / flood fill shape changes.

It has been working well for me for months now and I would say, subjectively of course, sounds nicer than the regulated snubbered single LM3886 I had going before. The bass seems to have more impact - I noticed the door rattling where I never did before - the midrange / treble seems to be smoother, I can't really explain it well, but guitars sound "raw", but at the same time pleasant, whereas before they sounded kinda harsh. But that's my subjective opinion, just to make it clear :p

Board size is 100mm x 100mm per channel and all that's needed is to wire up a transformer. The power supply caps are on the PCB. Should you want to remove the power supply from the board I guess there's nothing stopping you from just not soldering them on, and bridging one of the rectifier AC to + / -, then feeding a DC input at the connectors.

The low level signal grounds are tied together on the PCB in a star fashion, but not directly linked to the power ground on the PCB as I wasn't sure where would be a good place to do this. Maybe it would be OK to join the two ground at the input ground connector, just in front of the star perhaps?

It doesn't seem to require much effort grounding here for hum / buzz free operation anyway. I run the input ground straight to the input ground connection, then a wire back out of that connection to a loop breaker. Then a wire for power ground to the same point. I hear a barely audible amount of hiss from the tweeter at full volume, with my ear less than 1cm away, and no audible hum / buzz from the woofers, again with my ear less than 1cm away. I don't remember hearing any noise without the input ground wire going to the disconnecting network, but relying on 'Rsafe' to link the two grounds though.

I lack the resources to measure the power output and such. All I can say is that I have hooked up an 8 ohm nominal pair of speakers with a 4 ohm nominal pair of speakers in parallel (2 ohms?), on each channel, and it does not seem to struggle.



Anyway, here's the schematic and layout. I have mounted 100nF bypass caps for each chip on the under side of the board right on the pins as I couldn't really see a way to include them in the layout. Semi-inspired by the PA150 over at shine7 (thanks), only I am rubbish with SMD parts, so this is through hole.

[IMGDEAD]http://img217.imageshack.us/img217/9866/pa150sdt0.jpg[/IMGDEAD]

[IMGDEAD]http://img217.imageshack.us/img217/4643/pa150fr7.jpg[/IMGDEAD]

[IMGDEAD]http://img90.imageshack.us/img90/4672/pa150bin4.jpg[/IMGDEAD]



Pictures of my home made PCB version / testing are at http://www.diyaudio.com/forums/showthread.php?postid=1292004#post1292004


[edit]

Now trying to sort out attaching the eagle files, bear with me!
 

Attachments

  • gainclone ds v1.zip
    55.6 KB · Views: 1,059