Design critique

[MWP]

Hi all,

Ive just finished my new design for a DAC + Gainclone Amp i planning to build shortly.
Its one of the bigger more complex designs ive attempted from scratch.

Anyone care to give it a once over?
It would be nice to get some comments/thumbs-up from you guys before i put in the many hours designing and building the PCBs for it.

http://www.overclockers.com.au/~mwp/...p-dac-amp1.pdf

BTW...
Yes, i know there is no volume control besides the two step relay atten. Volume control will be before the DACs.
And, yes, one single ground plane instead of split.

Thanks!

[carlmart]

Originally posted by MWP
Hi all,

Ive just finished my new design for a DAC + Gainclone Amp i planning to build shortly.
Its one of the bigger more complex designs ive attempted from scratch.

Anyone care to give it a once over?
It would be nice to get some comments/thumbs-up from you guys before i put in the many hours designing and building the PCBs for it.

Why don't you feed the bridge directly from the PCM1798? You already have differential outputs for that.

If the reason is to buffer it or to amp it, then it might be better to use a differential in/differential out chip. As it is there might be some CMRR imbalances.


Carlos

[MWP]

Originally posted by carlmart
Why don't you feed the bridge directly from the PCM1798? You already have differential outputs for that.

Yes, but they are current outputs which require IV conversion.
Using the LM3886's to do this would create gain problems and would probably be unstable.

[johnnyx]

Usually the Zobel network goes directly to the amp output, rather than to the output sockets.

[MWP]

Originally posted by johnnyx
Usually the Zobel network goes directly to the amp output, rather than to the output sockets.

Can you explain this further?
I dont understand.

[GregGC]

Originally posted by MWP
Hi all,

Ive just finished my new design for a DAC + Gainclone Amp i planning to build shortly.
Its one of the bigger more complex designs ive attempted from scratch.

Anyone care to give it a once over?
It would be nice to get some comments/thumbs-up from you guys before i put in the many hours designing and building the PCBs for it.

http://www.overclockers.com.au/~mwp/...p-dac-amp1.pdf

BTW...
Yes, i know there is no volume control besides the two step relay atten. Volume control will be before the DACs.
And, yes, one single ground plane instead of split.

Thanks!

Looks good to me. The only thing I find would be that the VCANCEL may have to be different for the Left and Right chanel (or even between the inv. and noninv. of the opamps within the same chanel). So the DC offset could suffer (in your case the IGC has a DC gain of 18), but if you put caps between GC and OPA's (the 10k to gnd from +in of 3886 may have to go to 180k or so)you'd be fine. Hope I didn't confuse you.

/Greg

[MWP]

Originally posted by GregGC

Looks good to me. The only thing I find would be that the VCANCEL may have to be different for the Left and Right chanel (or even between the inv. and noninv. of the opamps within the same chanel). So the DC offset could suffer (in your case the IGC has a DC gain of 18), but if you put caps between GC and OPA's (the 10k to gnd from +in of 3886 may have to go to 180k or so)you'd be fine. Hope I didn't confuse you.
/Greg

Good point.
I prefer to keep DC blocking caps out of the audio line as its quite difficult to find good quality caps here.

I might give a separate DC bias adjust for each of the IV opamps.