[balou]

For an unbuffered design, the biasing current would have to be higher than the maximum expected output current.

let's assume a grado headphone at line-level voltage of 2v.
U = R * I
2V = 32 Ohm * I
2/32 = I
2/32 = 0.0625 A
= 62.5 mA

so the biasing current would have to be higher than 62.5mA to assure that the opamp stays class A. But this is completely impractical - thats about double the current draw of a pimeta or a mini³ (of course depending on exact config) - just for the biasing.

 

[Bjornboy81]

^Can you bias an the output stage:???: (I thought it had to be before a buffer)

 

[majkel]

00940, please note that there is an application with current feedback op-amp after the compensation pin and it does make a difference for the primary class A AD744 stage. What people normally do (so did I to check it) is just use comp. pin as output to a buffer or efficient voltage feedback op-amp, which produces bad sound because of differrent kind of comp. pin load.

 

[MisterX]

Something was major wrong with your circuit then because I would never call the AD744 with the output taken from the comp pin thin or bright in a buffered amp (in fact it is pretty much the opposite of that).

 

[NelsonVandal]

Quote:

Originally Posted by villekille /img/forum/go_quote.gif All the amps on tangent's site which have class a biasing are buffered amps. The mini3 doesn't have class a biasing AFAIK but it doesn't have a buffer. Is there anything to gain with class a biasing if using only a high current output opamp without any buffer? The bias current should be higher I guess.

I've tried class A biasing the Mini3 with 9 mA JFETs. Both AD8397 and OPA691. When used with 300 Ohm phones this is quite enough for my listening levels. I can hear no difference at all. I think the "improvement" with class A biasing is placebo all the way. I don't think AD744, AD829 and NE5534 sound better because of class A biasing. I think it's because the output stage is bypassed. I think every part (especially active ones and capacitors) the signal passes colors or degrades the sound.

At first I thought my amps improved with class A biasing, but when I tried this tweak live while playing, I heard absolutely no difference. Expectation is very powerful.