[shake]

I'm thinking of building a pimeta-esque amp.  I'm going to have LMH6321 buffered OPA627s for the left and right channels.  However, for the ground driver, I'm thinking of using the TI's suggest pseudo-ground instead of a 3rd channel.  This is an attempt to build the amp using a "star" topology rather than the 3 channel design.  Would this work?
Pseudo-ground driver:

 
The "start" topology.  Credit to jcx:


And would it be better to wrap the buffer in an opamp's feedback loop instead or would there be no benefit? Credit to Tangent:

If I did this method, would I have to use the same opamp (OPA627) and buffer (LMH6321)?

And a non-related question: Would there be any benefit in bypassing the opamps and buffers with ceramic and film capacitors in parallel?

Thanks in advance.
 
EDIT: current schematic:

 

[G.Trenchev]

You can wrap any buffer in the opamp's feedback loop.It will be very beneficial for quality(much lower distortion) and reliability(lower offset).
Personally,I find the pseudo-ground design to be pointless for desktop amps.It is intended to split the output of a single power supply,like a battery for example.It adds resistance to the ground of the schematic,which can degrade performance,and it shifts the ground slightly off 0V.In a desktop amp you can just use a transformer with a dual secondary coil,which outputs 12-GND-12 V,make a decent regulated power supply with LM317/337,and leave the ground to be a normal conductor.This is the best approach.
I'd recommended you to bypass the opamp's and buffer's supply pins(close to the device package) with a 0.1uF capacitor an a 1uF tantal capacitor in parallel.It'll help reduce PS fluctuations and keep the stereo separation.It is very important to keep the layout clean and neat too.Seek to place the components of the two channels symmetrically,and keep the ground wire short.

 

[Avro_Arrow]

I would do as suggested above, but if you do want to try your design, you should at least parallel two BUF634 for the
ground channel as the LMH6321 have 300mA output and BUF634 is only 250mA. In theory, you should be able to
sink both signal channel at the same time (600mA) but in practice we don't usual use maximum power.
 
Putting the BUF634 in the feedback loop for the ground channel (if you use one) keeps the output impedance close
to zero. On it's own, BUF634 has an output impedance of about 10 ohms.
 
If you want my personal opinion, if I were to do a project like you are interested in, I would use the signal channels
from the PPA v2 with a dual rail power supply.

 

[shake]

About the transformer, I've never looked into using more than just a wallwart DC power supply.  What would I need to buy and how would I implement it? One of these transformers?

 

[Avro_Arrow]

There are quite a wide variety of transformers that will work for a dual
rail power supply.
You can use a transformer with dual secondaries, a center tap secondary, or
even two separate transformers.
 
Something like an Amveco 62042 would be a good choice.
At DigiKey.
 
P.S. You digiKey link didn't work.

 

[shake]

 
Quote:

[size=x-small]For 115V operation, connect primaries in parallel by connecting yellow and red lead wires together and black and violet leads wires. To parallel the secondaries, connect green and brown wires and red and blue together.[/size]

 
So Yellow+Red and Black+Violet = AC in and out, Green = +12V, Red+Brown = 0V, Blue = -12V?

 

[Avro_Arrow]

Yes, that would turn it into a center tap transformer.
 
You could also wire the secondaries independently like
in this schematic:
 

 
X1 and X2 would connect to the secondaries of the transformer.
X1-1 = Green
X1-2 = Red
X2-1 = Brown
X2-2 = Blue

 

[Avro_Arrow]

Here is a nice power entry module.
It incorporates the IEC connector, a fuse, power switch and EMI filter.

 

[shake]

Do you have anything simpler?  I had a ~$100 max budget for this...  And I don't really have any case working tools (drill/dremel)

 

[shake]

Quote:

I'd recommended you to bypass the opamp's and buffer's supply pins(close to the device package) with a 0.1uF capacitor an a 1uF tantal capacitor in parallel.It'll help reduce PS fluctuations and keep the stereo separation.It is very important to keep the layout clean and neat too.Seek to place the components of the two channels symmetrically,and keep the ground wire short.

1µF tant cap and a 0.1µF film/ceramic?

 

[Avro_Arrow]

You could just connect the power cord directly to the transformer through a switch...
 
Just make sure no 120 volt AC is exposed.
 
The regulator schematic is just an example.
 
Quote:

Do you have anything simpler?  I had a ~$100 max budget for this...  And I don't really have any case working tools (drill/dremel)

 

[shake]

The transformer puts out DC?

 

[Avro_Arrow]

No, the transformer puts out AC.
 
I mean like this:
 
Power cord > Switch > Transformer > Rectifier/Regulator > Amp

 

[shake]

So from your above schematic, I would need D1-D8 and 2 LM317s?  Is that a basic schematic?

Back to the ground driver though, using the opamp->buffer combo, could I use any opamp or should I use a certain one?  I don't want to have to buy 3 OPA627s if I don't have to, they're expensive.  Does the LMH6321 act as a good buffer in this case as well?

 

[Avro_Arrow]

The regulators in my schematic are LM317 and LM337.
 
You could get by with one bridge rectifier, LM7812 and LM7912 regulators and
a couple of filter caps with a center tap transformer for a minimal power supply.
I can draw you up a schematic tomorrow...
 
You have to remember though, a better power supply will result in better sound.
 
Are you going to do a two channel OPA627/LMH6321 setup?