Discrete Opamp Open Design

[gpapag]

Originally Posted by bcarso

Some of us like to also confine currents to loops as local to a given circuit section as possible

This is very important Brad, thank you.
If all amplification stages are connected in series (think board level) to the PSU, local RC decoupling alone, does not.
Parallel connection to the PSU plus local decoupling does.

Originally Posted by bear

hmmm... more tiny ants to herd.

Ah, now I see bear. I overlooked your true nature


George

[bear]

Part of what I was thinking is if there are consequences to designing for high PSRR or not.

Does making the circuit "do work" to cancel PS "stuff" effect the desired operation, or not.

_-_-bear

[bear]

George,

Your diagram shows explicit filter elements or intrinsic resistances and inductances?

A bit of this depends on exactly which, since serial filtering *might* yield vanishingly low noise and hum at the end of the chain.

_-_-bear

[PMA]

Originally Posted by bear

Part of what I was thinking is if there are consequences to designing for high PSRR or not.

Does making the circuit "do work" to cancel PS "stuff" effect the desired operation, or not.

_-_-bear

I think yes. IMO the good audio circuit design should have

- very high speed, both wide bandwidth and very high slew rate
- very high immunity to RFI/HF demodulation and detection
- high High Frequency PSRR
- low distortion with no (or maximally suppressed) higher harmonics
- low noise

[Joachim Gerhard]

Low crosstalk up to high frequencies is on my list too. Even a good potentiometer has enough parasitics to screw that up in the treble.

[jan.didden]

Originally Posted by gpapag

This is very important Brad, thank you.
If all amplification stages are connected in series (think board level) to the PSU, local RC decoupling alone, does not.
Parallel connection to the PSU plus local decoupling does.




Ah, now I see bear. I overlooked your true nature


George

Since you cvan't have all your loads directly at the psu (unless you give each stage its own psu) a good way is to determine the stage with lowest PSRR and put the psu closest to that. And, if you're serious, use remote sensing for the regulator at that stage.

jan

[forr]

Originally Posted by Joachim Gerhard

Low crosstalk up to high frequencies is on my list too. Even a good potentiometer has enough parasitics to screw that up in the treble.

Electronically speaking, everybody shall agree that crosstalk at high frequencies is an undesirable thing as it looks as a parasitic signal.
But for a good stereo rendition, it may not. Who does remember Edeko's articles in Electronics World ?

[Joachim Gerhard]

I know that phono cartridges with -20dB crosstalk can just sound fine. Some bleeding in the other channel may even enhance the perception of spaciousness. Schroeder has shown a circuit like that in the 70th and there have being processors from Carver, Bongiorno and others. I rather prefer that not be be a characteristic of the electronics.

[john curl]

Good stereo separation in the electronics is VERY IMPORTANT. Don't be fooled by phono cartridges Joachim.

[RNMarsh]

In some old timey circuits there was a "blend" control to have you cover that "hole in the middle' with some mono. That helped when speakers were too far apart.
Short version -- If you mix/blend left and right towards mono you will find a lot of info gets cancelled out in the process. That is the effect of reducing channel seperation or isolation -- if you want all the details in tack, you design circuits with as little 'blend' as possible. - RNMarsh