[jcx]

of course dual mono lets you really do the right thing with star gnd and just 2 amp boards and Will beat 3-channel active gnd in measurable performance - you'd need 6 pins (+ a overall chassis shield?) in the power connector to do it right with 2x dual supplies (at least dual CT/4 independent secondaries if not 2 xmfrs)

and by using "balanced" 4-pin headphone connection even though driving SE you really could't do better without re-laying out the Beta22 pcb

 

[Steve Eddy]

Quote:

Originally Posted by jcx /img/forum/go_quote.gif of course dual mono lets you really do the right thing with star gnd and Will beat 3-channel active gnd in measurable performance - you'd need 6 pins (+ a overall chassis shield?) in the power connector to do it right with 2x dual supplies

I'm not so sure there'd be any significant benefit to going with dual mono supplies.

Providing all three channels are operating in class A, this should result in a constant current draw from the power supply.

Also, he'll be using a o22 regulated power supply, and the B22 is bristling with cascodes so I really don't see much in the way of crosstalk through the supply rails.

Finally, given that it's a three channel amp, I assume he'll be using TRS plugs/jacks and it seems that whatever benefits there might be due to dual mono supplies would be pretty well swamped by the crosstalk from the common contact resistance in the TRS connection.

se

 

[jcx]

granted with 3-pin TRS headphone connection the benefit isn't as great and Class A operation makes most of the concern moot anyway

but the excuse used for 3-channel is "ground contamination" - dual mono ps lets the star gnd be at the output connector without any gnd contamination if the R/L supplies float (at least without any more “gnd contamination” than is inherent in the Beta22 pcb gnd layout)

the extra supply appeals to me much more than building a 3rd Beta22 board – active gnd may make sense with op amps and single polarity supplies but how did this silly “3-channel” idea get higher audiophile “fashion” currency than dual mono supplies if you’re really trying for over-engineering one-upmanship?

 

[Steve Eddy]

Quote:

Originally Posted by jcx /img/forum/go_quote.gif granted with 3-pin TRS headphone connection the benefit isn't as great and Class A operation makes most of the concern moot anyway but the excuse used for 3-channel is "ground contamination" - dual mono ps lets the star gnd be at the output connector without any gnd contamination if the R/L supplies float (at least without any more “gnd contamination” than is inherent in the Beta22 pcb gnd layout)

But the whole "ground contamination" thing was a myth based on a lack of understanding on the part of those promoting it as such.

Quote:

the extra supply appeals to me much more than building a 3rd Beta22 board – active gnd may make sense with op amps and single polarity supplies but how did this silly “3-channel” idea get higher audiophile “fashion” currency than dual mono supplies if you’re really trying for over-engineering one-upmanship?

But what the third channel DOES do is maintain a constant current draw from the power supply (again, provided all three channels are operating in class A which they should be in the B22 with most headphones).

If the current draw from the power supply is constant, then there can't be any interchannel crosstalk via the supply rails.

And if you pay proper attention to grounding, there shouldn't be any problems there either, leaving the only significant source of crosstalk the common contact resistance in the TRS plug/jack.

se

 

[jcx]

Constant power draw could be a supply regulation benefit with battery supply – line regulation with rectified AC is a bigger problem to overcome in wall supply than output regulation – a good enough active regulator for rectifier/cap ripple rejection will likely have very low load side output impedance

There wouldn’t be any interchannel crosstalk with dual mono ps – each amp&ps only sees its own linear Class A load current modulation times the regulator output impedance – ie a (very) small linear signal

The Beta22 circuit should have pretty extreme ps rejection so the trade is whether the constant (and higher) current draw on a single supply is worth the 3rd-channel added noise and distortion in the output gnd that can be avoided with (cheaper to build) dual mono ps – also note that since the 3rd channel has to sink both R/L current it will leave Class A sooner if R/L have common large amplitude signal (like any bass frequency with sound wavelength much larger than you ear spacing) unless you bias the gnd channel 2x the R/L for a total of 4x supply drain

The place where ps rail signal can more easily couple to the circuit is on each individual R/L channel Beta22 pcb – where the ps current draw isn’t constant – the 3rd channel only helps the view from the external power supply


anyone chasing the implied level of performance would probably use 2 of the same power supplies the 3-channel builders are using for double the supply C and effectively ½ the transformer winding impedance, cutting ripple at each dual mono regulator's inputs by ~3-4x

 

[Steve Eddy]

Ok! Ok! I give! I give!

UNCLLLLLLLLLE!

I was just trying to point out that the third channel isn't entirely without some merit.

se

 

[TimJo]

Quote:

Originally Posted by jcx /img/forum/go_quote.gif Constant power draw could be a supply regulation benefit with battery supply – line regulation with rectified AC is a bigger problem to overcome in wall supply than output regulation – a good enough active regulator for rectifier/cap ripple rejection will likely have very low load side output impedance There wouldn’t be any interchannel crosstalk with dual mono ps – each amp&ps only sees its own linear Class A load current modulation times the regulator output impedance – ie a (very) small linear signal The Beta22 circuit should have pretty extreme ps rejection so the trade is whether the constant (and higher) current draw on a single supply is worth the 3rd-channel added noise and distortion in the output gnd that can be avoided with (cheaper to build) dual mono ps – also note that since the 3rd channel has to sink both R/L current it will leave Class A sooner if R/L have common large amplitude signal (like any bass frequency with sound wavelength much larger than you ear spacing) unless you bias the gnd channel 2x the R/L for a total of 4x supply drain The place where ps rail signal can more easily couple to the circuit is on each individual R/L channel Beta22 pcb – where the ps current draw isn’t constant – the 3rd channel only helps the view from the external power supply anyone chasing the implied level of performance would probably use 2 of the same power supplies the 3-channel builders are using for double the supply C and effectively ½ the transformer winding impedance, cutting ripple at each dual mono regulator's inputs by ~3-4x

Interesting discussion...

 

[IPodPJ]

Quote:

Originally Posted by TimJo /img/forum/go_quote.gif One solution to this is wire it in the style fishski13 did, with the power cable hardwired to the ps. This reduces the cost by only requiring one set of connectors, and there will never be any confusion. Take a look at the last photo in the post. http://www.head-fi.org/forums/6116283-post5916.html

That's fine except for when you get a different rack and you need more space between gear.