active summing with vca's

[enthalpystudios]

Any advantage in doing this, since a vca is current in current out? I'd imagine there would be some advantage, or a complete disadvantage....

not, by the way, just for the hell of it though... i don't see a need to stick a vca in somewhere for no old reason, but if a vca was the route one wanted to take...

I suppose for mono summing, thats a question. but for a circuit with a standard pan, and vca fader, i suppose it would need an opamp to bring that current back to a voltage before being panned and summed, right?

 

[JohnRoberts]

While I already answered this in another thread some of my comments weren't clear so I'll try again.

http://www.groupdiy.com/index.php?topic=21340&postdays=0&postorder=asc&start=15

As with any engineering decision there are pro's and con's.

PRO: The biggest benefit IMO is elimination of N+1 noise gain factor in the L/R summing bus. The noise gain not only multiplies the summing amps noise by the number of channels being combined, but attenuates the negative feedback's ability to reduce the summing amps open loop distortion and phase shift. This may be inconsequential for modest numbers of channels with high performance opamps, but for larger structures this can be a factor. With current sources the summing bus noise contribution is effectively 1x, and noise floor of the individual channels will combine as the SQ root of sum of squares (+3dB for every doubling of channels).

PRO#2: There are no longer headroom issues at post fader gain stage (typically fixed +10dB) and channel attenuation truly attenuates channel noise floor perhaps below the previous fixed gain stage (may vary with specific VCA or gain stage design).

PRO#3: If summing amp, which is no longer a concern for ultra low noise, is also made a VCA, it becomes capable of extremely high headroom performance. The proverbial un-clippable summing amp. If output builds up just trim it down. (Note: there may be a small noise consequence to scaling the VCA current capability for this but since this noise is now a 1x contribution not significant.)

PRO#4: Pan law and fader law are now voltage controlled so more design flexibility for pan/fader law using standard parts and perhaps grouping and/or automation.

PRO#5: Since signals are being sent and combined as currents rather voltages there is no no need to reference or deal with ground potential differences between channel send and master bus receiver.

CON: Cost/complexity (2 VCAs per input channel).

CON#2: does not support multiple assignments simultaneously as you could with voltage sources.

CON#3: Solo with true fader/pan information can not be extracted nondestructively. Perhaps use PFL (pre fader listen) or just live without that capability.

CON#4: VCA noise floor will be higher than simple line input, but << than mic preamp at nominal gains. So may look worse on spec sheet for stand alone line level mixer.

CON#5: Potential bus capacitance build up that may need to be adressed in summing amp stability compensation.


FWIW: Despite some attractive benefits I am not aware of this being done on any of the sundry VCA automation consoles, but that would be worth investigating. I have some experience with current source summing structures and there are a few quirks about dealing with them, but all manageable.

JR

 

[enthalpystudios]

John, thanks for replying... I had actually read that post just perhaps 2 days ago, but neglected to realize that you were answering a question about not just vca pan, but vca pan AND fade, which makes a lot of sense now.

The other thing is that I'm not interested in multiple bus assignments OR solo in place.... PFL and 1 stereo bus is what i was already aiming for.

Also, in your first PRO (#1), the 2nd sentence applies to an active voltage summing topology, right? Just making sure I'm reading that right.

The headroom factor, and the law factor, are both very interesting. As far as sending currents rather than voltages, you mention that there is no need to reference or deal with ground potential differences between channel send and master bus receiver..... so basically, whether vca (already a virtual earth input) or opamp (with non-inverting input grounded), the master bus receivers ground can be tied to just 'a suitable ground,' rather than worrying to great lengths about having common ground noise on that reference and being sure that there is no 'crud' or high resistance paths in that ground?

Hope my regurgitation made enough sense for verification... if I understand correctly, however, that is an awesome physical constraint to remove.

About potential bus capacitance build up, I read last night every data sheet I could find, that2180, 2181, all of their app notes, the dbx 202 data sheet, dbx 2150, the old THAT 2150, and also the analog devices DIP VCA, forget the model number on it.

I can't for the life of me find where I read something about that capacitance build up, as the there is already some capacitance at the vca output? 15pf was it?

I should start printing these things and using a highlighter, as just reading pdfs really loses me when i'm reading tons of information, and referencing books at the same time.

but thanks for answering... a vca fader and pan is a *really* attractive idea for me with what i want to do. very cool that the only cons are cons that i've already made decisions about and will live without (multiple bus assignment and solo in place)

billy

 

[enthalpystudios]

i found this link, I had read it last night, a THAT app note regarding a vca pan circuit. Reading it again, I still don't quite see where the actual pot itself goes. I see "gain control" and 'inverse gain control,' so I'm assuming thats the place, but I figured they'd at least show whats going on.

I'd imagine that a potentiometer, as a fader, could simply divide an incoming 5v with wiper connected to "Control"

THAT VCA Pan Control App Note