"Marte Schwarz" wrote in
:

Hi,

20 bit signal conversion, so this op-amp will have to be good to
maintain
that and the other specs this unit has.
So did you think about noise?
Yes, several times. This amp has THD+N of 0.00005% and is going into a
device
specificied as 0.002%. Good enough? I think so.

Oh I didn't talk about distortion, I mean noise SNR or ENOB in total
system.


Not my job. I'm adding an op-amp and four resistors per channel. Not much I
can do about the rest of the system, except to note that it claims 20 bits
while having a 98dB dynamic range that barely scrapes past 16 bits worth of
resolution (though this does apply to the DAC too, the whole chain..). All
I'm concerned with is that I do not add to the existing reduction
significantly. George Herold already helped me with that, and I'd already
suspected I could reduce noise to acceptable levels for real 20 bit (120 dB)
dynamic range by lowering the value of the resistors in the passive adder
(which George Herold confirms, but with useful detail I didn't yet have).
Apart from choosing the op-amp there isn't anything else I can do. I'm not
trying to improve the system, just to make effective DC coupling for it
without harming it by adding complex circuitry where simple stuff will do it
better. If I didn't have to handle the bias voltage as I do, I wouldn't be
looking for an op-amp at all.

Some systems are nice, they have no DC bias on either side of the DC blocking
cap, but when they do, I can't think of a way to satisfy all conditions
without an op-amp to mix the bias voltage into its output without it
appearing on the output of the previous stage, so that op-amp must be there
if the capacitor isn't.

If there is a simpler and quieter way to mix two signals without output
inversion than a passive adder based on an op-amp (gain of 2) and 4 resistors
I'd like to know what it is, because in many years I have never found one.