[LAU] turning a consumer soundcard into "prosumer" w/ quasi-balanced outs

[Niels Mayer]

On Wed, Jun 9, 2010 at 3:05 PM,  <fons at kokkinizita.net> wrote:
>> But a wide-bandwidth op-amp is going to be
>> intrinsically noisier.... and a quiet op-amp is going to be
>> intrinsically slower,
>
> No, there's no such simple relation.

Actually there is a relationship, and it is certainly not simple.
However the generality holds, mostly because when you get down to it a
fast device achieves speed through either size, or high-current.
High-current induces one kind of noise, and small device geometries
induce another; field-effect, it's own, alongside material device
physics, temperatures, etc... For example one might prefer FET or
MosFet for linearity and gain-bandwidth linearity, but they're noisy.
Ultimately it comes down to gain-bandwidth, which is nonlinear in
low-noise Bipolar Junction-based opamps like you'd normally use in
audio: http://waltjung.org/PDFs/WTnT_Op_Amp_Audio_3.pdf

http://focus.ti.com.cn/cn/lit/an/sloa051a/sloa051a.pdf
"Selecting High-Speed Operational Amplifiers Made Easy"
...........
Although no op amp is ideal, modern processing techniques yield
devices that come close, at
least in some parameters. This is by design. In fact, different op
amps are optimized to be close
to ideal for some parameters, while other parameters for the same op
amp may be quite
ordinary. It is a trade-off. Some parameters can be improved, but only
at the expense of others.
It is the designer’s function to select the op amp that is closest to
ideal in ways that matter to the
application, and to know which parameters can be discounted or ignored.
[...]
Op amp design is a series of trade-offs and an improvement in one
parameter is always
accompanied by degradation of others. This is certainly the case with
current-feedback
amplifiers where, to a degree, dc performance has been sacrificed to gain speed.
.............

-- Niels
http://nielsmayer.com