Virtual ground vs regular split rail supply for headphone amp

[enton]

(saw the other thread about this, but it was 12years old and didn't want to grave dig it )

As the title says, I'm making a small headphone amp and it requires split/dual power supply of 2x12V.

Now I've checked out this schematic and these two comments about the sound being better with the virtual ground circuit.

From head-fi:

Although a simple and inexpensive virtual ground solution, some audio designs will sound better when using it. For example, when powering a headphone amplifier with this circuit the bass notes may sound considerably clearer and more life like. The reason for this unusually good performance may be that the voltage regulators create an "unbudgable ground" - holding the ground point in place very firmly compared to other circuits, virtual or not.

From the other thread:

I only have experience from headphone amplifiers, but I've found active virtual ground channel to sound better than passive (relying on capacitor quality). I've never used fancy capacitors though, just Panasonic FM, Sanyo WX and such.

The amp is with TPA6120 chip, and I planned on getting a new transformer with two outputs but I already have the 24V single output one.

So, should I continue building with the current transformer that I have and go for a virtual ground power supply? Or get a new dual output one, and build classic symmetrical power supply?

 

[dreamth]

https://www.diyaudio.com/forums/hea...ough-lesson-audiophile-world.html#post6065151
https://www.diyaudio.com/forums/hea...gh-lesson-audiophile-world-2.html#post6065578

It looks like the simplest form of virtual ground to me...but the output series resistor is included in the feedback loop for higher DF( consequently better base reproduction).

 

[rayma]

This virtual ground relies on capacitors, but a standard ground does not, since it is a direct connection.
There is a 1 ohm resistor between each regulator and its capacitor, which further degrades the quality
of the ground. A dual voltage supply with two isolated secondaries is much better, using identical
regulators with their outputs in series.

 

[dreamth]

With dynamic headphones over 32 ohms that 1 ohm resistor present no problem...Actually the headphones tweaked after the Harman curve(most of them) have too much base...

 

[Mark Tillotson]

This virtual ground relies on capacitors, but a standard ground does not, since it is a direct connection.
There is a 1 ohm resistor between each regulator and its capacitor, which further degrades the quality
of the ground. A dual voltage supply with two isolated secondaries is much better, using identical
regulators with their outputs in series.

Well you can look at this differently - the virtual ground is defined to be perfect(*), but the +/- rails are a bit soggy due to the resistors.

Thus if the amp has good PSRR its not a problem, but with poor PSRR the 1/2 wave signal pulses on the power rails will add to distortion. This is a class B virtual ground circuit if you think about it.

(*) assuming floating supply in.

 

[dreamth]

if TPA6120 is used in balanced configuration, as in the RNHP, there's no problem with any noise at all if only used wth capacitors.Paralleling some 47...100k resistors over the capacitors will also give enough current for input bias the bipolar inputs of any op-amp.Actually RNHP is using the TPA6120 datasheet application as a guide, while most diy projects used it with unbalanced sgnal sources and they "ended" with "higher end" power supply characteristics imposed over TPA6120 or any other op amp used at the front native performance...

 

[rayma]

One node can't be perfect, it has to be a loop. Impedance is not defined using just one node.
Virtual grounds are only of interest in a single supply system where analog is added as an afterthought.
Hopefully that's not the case in diyAudio.

 

[sgrossklass]

That's a neat trick circuit right there. An LM317 referenced to ground will give about +1.25 V, so you can reference the thing to -1.25 V to arrive at 0 V.

Well you can look at this differently - the virtual ground is defined to be perfect(*), but the +/- rails are a bit soggy due to the resistors.

(*) assuming floating supply in.

(Emphasis added by yours truly - boxing, framing & hanging on wall is left to the reader.)

You're taking the words right out of my mouth. The important part is referencing absolutely everything to virtual ground. Then it will be your "real" ground and everything else will be dancing around it so to speak. Any virtual ground output impedance (and regulator noise) will be reflected back into the supplies.

Voltages are so relative that Einstein would be proud.

If you need lower output impedance, a conventional set of regulators can still be added afterwards, I'd think...

Thus if the amp has good PSRR its not a problem, but with poor PSRR the 1/2 wave signal pulses on the power rails will add to distortion. This is a class B virtual ground circuit if you think about it.

Sure, but this can be dealt with in the usual ways - it's not fundamentally different from any other power supply. Keep your V+/V-/PGND loop areas small (usually all routed next to each other), have a more than decent amount of local bypassing (think double-digit µF pus the usual hundreds of nF here), maybe add several hundred µF further off, use only PGND for bypassing, keep SGND separate but close. Plus the usual like keeping input loop area small, keeping an eye on output current path, keeping ground planes away from the inverting input (and/or possibly adding a small feedback cap in the pFs) etc.

 

[PRR]

Remember an active virtual ground with little/no capacitance doubles your dynamic battery drain.

 

[enton]

Well, now I have my answer, I will buy a new transformer and go for a split rail. Thanks everyone

 

[dreamth]

you might not need to buy a new transformer, you can easily build a voltage multiplier...2x like.I did voltage doubling batteries capable if keeping 800ma constant current with ease...