quasi balanced summing

[tv]

Question: if I understand correctly, this circuit can also output buffered unbalanced signal (of "the mix") - if all is dc coupled, one could also add a "standard" balanced input circuit afterwards and add a dc-killer servo to it (so that would be 4 opamps total) ?

Is there any reason such circuit wouldn't work with the likes of ne5532 etc?

 

[TedF]

I'm sure 5532s would work fine, there's nothing critical in the circuit.
My own tests on the bench don't show any advantage in using the 5532. I prefer the 33078 because it has lower quiescent current, and its output stage is assymetric, so any distortion tends to be 2nd harmonic. (although the nasty high frequency distortion that opamps suffer from when loaded, is always 3rd).

I'm not quite sure what you mean, tv, when you talk about buffering....

Regarding DC servos, I must admit that I find them more trouble than they are worth! :roll:

 

[tv]

by buffered I meant that both opamps in the circuit do this as a side-effect, too - so the output is effectivelly buffered as well as summed?

 

[cuelist]

[quote author="TedF"] Regarding DC servos, I must admit that I find them more trouble than they are worth! :roll:[/quote]

Indeed... what's wrong with a good ol' capacitor of the right kind :wink:

 

[NewYorkDave]

Regarding DC servos, I must admit that I find them more trouble than they are worth! :roll:

Amen. In many cases, they're solidly in the "designer showing off" category.

 

[tv]

Maybe ... probably - if you use 2nd opamp in a dip, it can be a fraction of cost otherwise spent for capacitors.

 

[Moby]

[quote author="TedF"]Maybe I'm missing something, but the method of mixing using dual virtual earth amplifiers has been covered here before...... :guinness:
The 'superbal' amplifier deals symetrically with a balanced bus, and as long as the reference ground connection is good, it's impervious to ground current interference, it's got 6dB more headroom (as good as 60V rail!), It's quiet and well behaved over very long runs.... I wouldn't design a mixer any other way.
As an example of its flexibility, I once designed a talkback/monitor system for a major London studio using unscreened pairs operating in balanced 'virtual earth'. with hundreds of metres of wire; it was clean, quiet and trouble free.
[/quote]
Yes, I agree about balanced summing. But I asked about quasi balanced. There is no channel diff amp, so just "single phase" source is present. The other "phase" should be noise in the cabling... Same for signal and ground network. So, there is no +6db, but have to be quiet as balanced. That's why I'm asking about that ground network. I'm not sure how that have to be terminated or whatever :?

 

[tv]

I suppose the given circuit would also work in a quasi balanced situation - only you'd have to route the "grounds" to "-" bus.

Maybe I'm wrong, but to me this circuit looks like it's already self-terminating so you just have to connect the busses.

 

[Matt Syson]

Hi
The Superbal (Ted F ) circuit has an advantage in that it is truly impedance balanced and can attain better noise rejection than say the 'Calrec' design (variants used by many others). The Calrec has an advantage of simplicity and possibly slightly quieter although this depends on exactly how it is set up. The + leg of the Calrec actually goes to BSE (back stop earth, nothing to do with cows) which is the ground for all routing switches. This system was pretty good at rejecting hum but it was wiring placement sensitive. Hum rejection of 30dB or so is usually enough to get it below the noise and is better than unbalanced.
A previous post missed something in explanation I think, the quasi balanced part collects the assorted hum and induced rubbish and attempts to cancel it at the mix amp. Having resistors of the same value as the normal mix resistors helps the rubbish to appear at the same level and phase as rubbish on the mix bus. Because the Superbal is balanced rather than the Calrec which is only differential the Superbal is more...Super! I hope some of this makes sense, my brain hurts tonight.
Matt S

 

[Moby]

Thanks Matt, now everything sounds clear :wink: Point is to took ground from same place as mixing source and lead signals and "ground" through identical cabling. Also, resistor network have to be same for signal and ground. :green:

 

[Mlewis]

i noticed that this version has a +ve and -ve polarity output. could the circuit be used as a ballanced in - ballanced out line amp using the values described in the superbal line pre thread?

http://www.groupdiy.com/index.php?topic=12586&highlight=superbal?

would there be any problems with the -ve polarity output when varying the gain or anything or is it fairly bulletproof?

thanks.

 

[TedF]

It's really quite bullet-proof.... just observe the usual care about interfacing opamps with the outside world.... buffer them with a bit of R and try not to subject them to high voltages!
FWIW, I've found that most 'differences' in sound are caused through variations in interfaces. :?

 

[Mlewis]

thanks for the quick reply and a big thanks for sharing your info!

not quite sure what u meant by variations in interfaces though. are you talking about common mode differences between the +ve and -ve signals/impedences that you are interfacing to? kinda like what Bill Whitlock was trying to correct in the paper linked bellow?

http://www.jensen-transformers.com/an/ingenaes.pdf

or differences in the various output topologies that you are interfacing to?

thanks.

 

[Gulistan]

and what about this ?