VCA bus mixer using all THAT all the time

Where we discuss new analog design ideas for Pro Audio and modern spins on vintage ones.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

VCA bus mixer using all THAT all the time

Post by JR. »

I first wrote about current source summing back in 1980. I need to get Wayne that old console design article I wrote, if he doesn't already have it.

The principle is pretty simple. Instead of resistors feeding a simple virtual earth inverting amp with the noise gain build up on N+1 for the number of sources combined, I used synthesized current sources in place of input resistors in two different console series and it delivered most of the promised benefit.

I have speculated before about using the current outputs from VCAs directly to feed a bus for similar benefit. Further if I use a VCA on the bus summing amp, I get near unlimited headroom too. I used an old dBx VCA on a L/R bus sum amp on a LOFT console back in the '70s. An extra feature I built into that one was an automated slow fade where you could set the speed of the fade to quiet in dB/second.

What I am thinking about now is a simple combining system for all the digital boys mixing outside the box. It requires two VCA per channel for L and R but this simplifies the pan-fader design as they become control voltages with no losses. Further using 4 VCAs instead of 2 per channel we can treat the signal as L+ L-, R+ and R- which eliminates the need for differential to SE input and little things like input stage overload, etc. (Note: balanced approach also requires 4 bus combining amps instead of two).

The input stage when using 4 VCAs is as simple as 4 cap coupled resistors into the VCA. I could provide a +4dBu/-10dBV option which is just different resistor values. This trades distortion for noise floor. In the old VCAs you could play with the current density of the VCA but the new ones are fixed.

I don't know, but it would be nice if using a VCA on the summing amp also provides a first order cancellation for the distortion of the input stage VCAs (perhaps at unity gain).

It seems like this would give a simple mix bus structure with tons of headroom, negligible ground potential issues, and good performance. For known signal levels one could even scale the input and output VCA operating points for lower noise, or lower distortion.

The output can be passed through as fully balanced differential, but I am philosophically opposed to doing that in something this complex. At the bus level I'd SE it to scrape out CM garbage. This gives us the opportunity to use a THAT balanced output driver... All THAT all the time.. :lol:

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: VCA bus mixer using all THAT all the time

Post by mediatechnology »

Cool JR. Go for it.

I think you sent me part of that REP article. I remember that coming up about year ago in a panning thread and IIRC, in that one, you used current outputs into a TransAmp? BTW Catchy thread title...

One thing that may make that more practical is the 2162 dual VCA which not only gets the package count down but also has 3 pF of output capacitance vs the 15 pF or so of the 2180/2181 series. The 2162 uses smaller geometry transistors in the gain cell.

The panner app note for the 2180/2181 is here. Obviously in your case the output I-V op amps are not needed.

http://www.thatcorp.com/datashts/dn120.pdf The tapering circuitry may be useful.
emrr
Posts: 570
Joined: Sun Mar 02, 2008 10:27 pm
Location: NC, USA
Contact:

Re: VCA bus mixer using all THAT all the time

Post by emrr »

JR, this sounds very interesting. Judging by the sound of the Pico, it would seem that the VCA's day has arrived, and I don't doubt it could sound very good. Keep it up.
Best,

Doug Williams
Electromagnetic Radiation Recorders
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: VCA bus mixer using all THAT all the time

Post by JR. »

This is mostly a mental exercise on my part. I don't think it would be that attractive as a stand alone line level mixer.

I have done some work with a console company that sells a VCA model so it might be worth them considering for one of their large consoles.

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: VCA bus mixer using all THAT all the time

Post by mediatechnology »

JR - Do you have a clean scan of that R/EP article you wrote? I think I have a copy that you sent but am not sure if I have the whole thing.

Wayne
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: VCA bus mixer using all THAT all the time

Post by JR. »

mediatechnology wrote:JR - Do you have a clean scan of that R/EP article you wrote? I think I have a copy that you sent but am not sure if I have the whole thing.

Wayne
http://www.circularscience.com/des_art.pdf I'm still hosting this scan on one of my websites. I'm not as organized as you about old stuff...

I would be glad to help from a design perspective anyone interested in testing this. I am not convinced that a VCA sum bus structure will deliver measurable performance enhancements for combining a modest numbers of channels, but I'm pretty sure it could offer benefits to a large console where mitigating the coherent noise/distortion/phase shift relationship is most useful. The biggest I ever used this on was something like 72 current sources in parallel with about 40 conventional resistor bus feeds to L/R. It was a split console with the whole right side of the board used synthesized current sources. I believe VCAs in place of my old synthesized current sources would work much better.

It's perhaps too much to do a scratch design of 100+ inputs just to test this, but it might be interesting to look at a large VCA based console for possible retro-fit or modification. I believe some combination of newer low capacitance VCAs and using the residual bus capacitance to form one pole of a two pole LPF would work nicely. While the noise of a 100 VCAs in parallel is non-zero it is incoherent so builds as the square root of 100x or 20dB. Not too shabby noise floor for 100 sources at full unity gain. Of course if these channels are operating at levels lower than unity gain the bus noise is likewise attenuated. My gut is that this would be a noticeable real world improvement but I still believe digital combining should be perfect, in theory.

If I was younger and not so busy I might be tempted to make a stand alone combining system that could be patched into inserts or direct outs.Or maybe pick up an old basket case console and mod it... these days I have other projects higher than that on my list.

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: VCA bus mixer using all THAT all the time

Post by mediatechnology »

Thanks JR - I'm printing that pdf now so I can read it on real paper.

I haven't done the mental exercise, but wouldn't there be a modest benefit just in reduction in mix amp noise gain with say, 16 channels?
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: VCA bus mixer using all THAT all the time

Post by JR. »

I believe I even said this in the almost 30 year old article, noise isn't the only or IMO primary benefit of this but OK...

For 16 simple mono inputs, the conventional combining amp will increase bus amp noise 17x. Using VCAs you will see roughly 4x VCA noise + 1x bus amp noise (actually a little more complicated than that, but close enough for this exercise). I can imagine making a uber low noise bus amp that is 12 dB quieter than a VCA or whatever it takes.

Of course if we look at more common stereo combining the pan throws in another loss in conventional approach that VCA can finesse with side chain manipulations. The simple most common pan circuit has on the order of 10 dB loss so now the VCA is back in the ball game. If we look at headroom and other practical considerations the VCA offers other benefits, like headroom and gain structure.

I must be getting old, my engineering side says yes, there's an opportunity for the OTB mixer crowd to give them a kick ass solution that would really work well, but my marketing side tells me they would be turned off by all those nasty transistors inside the VCA and electronics in general. It just doesn't seem un-digital enough.

OK, sure it would probably work for even 16 channels, but i would really focus on the whole performance picture. Phase shift, distortion*, and lastly noise.

JR

PS: **Distortion.. The current source summer will reduce the distortion of the summing amp, but will sum the sundry VCA distortions. I would also investigate the use of a VCA on the sum bus amp to perhaps effect a first order cancellation of VCA distortion if symmetrical... (I can hope can't I?). I actually used a old (202?) DBX vca on a console L/R bus with one of my early LOFT consoles using the current summing, but for a different reason.

PPS: If small geometry VCA is measurably noisier, I might be tempted to use quieter VCAs and mitigate higher bus C with LPF topology. This might be tricky with VCA in bus amp too, but still doable (I think).
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: VCA bus mixer using all THAT all the time

Post by mediatechnology »

I agree noise reduction isn't the only benefit.

In the thread I nuked elsewhere I think I mentioned that it was silly to have so much low-noise amplification ahead of the pan pot and build-out resistor and then burn it off, as resistive "heat," the 10 dB or so in the pan and the rest in the build- out just to produce a current in the mix amp. You're right that puts VCAs back in the game once you consider all the losses.

I do agree that with the market veering towards passive and transformer-based summing approaches log/anti-log transistors don't have the purist market appeal. But as you've said there's a lot of identifiable improvements doing it that way that would provide a solution that is perhaps greater than the sum of the parts. (Pardon the weak pun.)

Your original idea in this thread though would make a great reference design for an 8 or 16 channel mixer. Please keep thinking about it.

BTW I think you'll find the VCA distortions are going to power sum and not correlate. So they'll add at 3 dB/pair rather than 6 dB/pair. In a way another advantage...
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: VCA bus mixer using all THAT all the time

Post by JR. »

The VCA distortion will surely be uncorrelated since the individual audio inputs are all uncorrelated. The distortion should be the same as single VCA. I believe there is still a possibility that the individual input''s distortion might symmetrically cancel if a reciprocal topology VCA is used in the sum amp. In this case the input signal and playback signal could track 1:1.

I suspect many of the pieces to do this are around.. Didn't you scratch out a PAN circuit using VCAs? Another feature to add to the list is perhaps variable pan law...

I may have mentioned this as an attractive marketing vehicle for THAT corp.. A THAT line level input receiver, followed by THAT VCAs, feeding THAT VCAs in sum amp, connected to THAT output driver... Barely any non-THAT ICs inside. :D

JR
Cancel the "cancel culture", do not support mob hatred.
Post Reply