You don't read very well do you, never said ANYTHING about does NOT cost enough...WOW...
I would bet I could mod a Mackie, put an iron core trans on the output and it would smoke the FMR and still be less than HALF per channel...PLUS you get a real power supply and chassis and decent Eq's...
How many times are you going to respond to the same post? Sure, you can "mod" anything to be anything else. So?
By the way the analog section of a DA/AD does NOT run on 5V..More like +-18V...
And the digital is more like 3.2V...
Ahh, but somewhere the analog and digital must meet. Link me a datasheet to an ADC IC that accepts a +/-18V input. 6-7V is far more typical, and the maximum swing for 0dBFS is usually much less than that range.
Such that any box that accepts a +/-18V input signal will merely attenuate its input to fit the range the chip can actually accept.
Here are four examples:
AK5394A : max swing 5.6Vpp
PCM1794A: max swing 6.8Vpp
PCM4202 : max swing 6.6Vpp
CS5381: max swing 7.4Vpp
If you are designing a converter that uses such ICs, you *have* to limit the ADC input to those levels or you risk destroying the IC. So you have a choice, you either assume the incoming levels are always higher and you pad them, or you assume they are usually lower and you clamp them, or you do a combination of padding and clamping (you're going to clamp anyway to make sure nothing explodes). Or you put in a switch so you can let the user decide.
So what does increasing level to 48Vpp accomplish? As you know, increasing gain in a given stage will reduce bandwidth and increase distortion. You have to compensate for that by using faster amplifiers, which tend to be either noisier or less efficient. Thus it's practically impossible to do large amounts of gain, say >66dB, in a single stage, so you force a requirement for two gain stages. Why? Just so you have to attenuate the output?
Another possibility is the desire for transformer saturation. Fair enough, but then why spend the money on Jensens? Those are designed not to saturate. It's much more effective to use a transformer that saturates at lower levels if that is the desired effect. It's also quite a lot cheaper.
Thus, headroom as a goal in and of itself is pointless. Define the actual requirement, exceed it by a few dB, and have a few beers
Anyway, as sexy as +/-24V sounds, it's a whopping 2.5dB more than +/-18V. Hardly seems worth the price of a discrete opamp, more expensive power supply, etc. Also, with an output transformer you could simply invert the output of a 36V IC opamp and feed both to the transformer. That would exceed the headroom of the single-ended 48V supply (although the transformer would probably be crying at that point).
Also, why not build your own opamp out of discrete transistors? I don't get why a part that is about 20 $0.10 transistors and the same number of resistors, and a capacitor or two costs $50+. It's only about $5 in parts.
Like this thing, for example:
http://www.diyfactory.com/data/Great_River_Opamp2.pdf