Line mixer (active)

[NewYorkDave]

Here's something that could be useful for you guys who wanna mix out of the box using your outboard, etc. It's for when you need capabilities beyond those of a simple passive mixing box, but don't need a full-featured console.

NOTE: THIS IS NOT A READY-TO-BUILD PROJECT. THIS IS JUST THE OUTLINE OF THE CONCEPT. But I think it could be fleshed-out into a nice project.

PDF

As noted, coupling caps are not shown on the schematic but may be needed unless you can achieve and maintain very low offsets. If the mix busses are physically long or otherwise possess significant capacitance, it may be wise (for the sake of stability) to insert a low-value resistor in series with the inverting input of each mixing amp. The PFL/Monitor section could be essentially the same as the output section shown, with the necessary switching added at the input of the mix amp.

The prefade buffer is unity-gain and the post-fader amplifier gives 10dB of reserve gain, since the fader is usually run at 10dB of loss. The output amplifiers are also configured for 10dB of reserve gain. The panpot is my own variation on the usual dual-pot design; I've mentioned it here a number of times. It has performance advantages over the standard configuration, but at the cost of 6dB minimum loss. This loss is made up in the mix amp.

I was tempted to add subgroups and the like, but my intention was to keep the project basic and compact. I think this could lend itself to PCB-per-channel or PCB-per-group-of-four construction, with separate master section PCBs. The builder could configure for as many channels as desired, up to reasonable limits.

 

[Mlewis]

Cheers Dave...

looks good.

 

[radiance]

Thanks a lot for this!!


[quote author="NewYorkDave"]
I was tempted to add subgroups and the like, ........[/quote]

Please do :grin:

 

[enthalpystudios]

dude, it could be *very* sweet indeed.... doa/dip8 footprints for the gain stages ;]

what would be a good way to go as far as a transformer direct out somewhere between the line amp and the pan pot? say edcor perhaps or a 2503.....

man, all too cool.... the standard 22uf/.1uf coupling caps i see used a lot, are those more for ic's, or is the same arrangement useful for a DOA as well?

 

[gang of elk]

Neato. Looks like a great project. :guinness:

 

[Mlewis]

got a quick learner question if thats ok...

would adjusting the Aux level pots and pan pots alter the buss impedence much? My reason for asking was because i was thinking about whether impedance ballancing (like on the link below) could easily be applied to your your circuit or would the impedance change too much when settings were adjusted, to acheive good common mode noise rejection at the summing amp?

http://www.forsselltech.com/schematics/Summing%20Buss2.pdf

overkill for small mixers perhaps?

cheers, Mark.

 

[PRR]

> impedance balancing (like on the link below) could easily be applied to your your circuit

Forssell's technique does not rely on any specific mix impedance, and should not be called "impedance balanced". It sums the ground noise among the many sources and subtracts it from the mix. It does assume that all sources present fairly equal impedances; NYDave's plan does have some 10K-12K sources and a pair of 22K-24K sources, but Forssell's technique will still give significant noise reduction.

 

[Mlewis]

thanks PRR.

 

[enthalpystudios]

jesus, is an opamp buffer amplifier that simple? i sat staring at the schematic thinking ok, this is that, that is this, etc etc, and breaking it down for myself, and thought, well that right there must just be a buffer. and i thought... so no Rf? and thats that. all makes sense. no current at the input. buffer steals no current! blam!

god this has to sound idiotic... simple stuff i'm sure... anyway that clears up the entire reason i found this place (wanted to know how to split a signal without it getting all f0ked up)

anyway, thanks for posting this nyd! it was perfect for me, a step in between block diagram and full on schem... all the extras are confusing when you have no clue about anything at all.

also... quick question about the voltage divider at the input of the voltage amplifier... as the pot's value changes does it change the input impedance to the circuit?

thanks again,

billy

 

[Boswell]

[quote author="enthalpystudios"]quick question about the voltage divider at the input of the voltage amplifier... as the pot's value changes does it change the input impedance to the circuit? [/quote]
For the main channel inputs, as drawn, the answer is slightly. The 100K resistor (R1) will gradually parallel the 10K pot, so the input impedance willl drop to around 9K at full wick. R1 is not necessary, unless it is there to make a small change to the log law of the 10K pot.

For the return inputs, the 10K pot has 22K to virtual gorund, so this will reduce the input impedance to about 6.9K when at full. More importantly, it will upset the log law of the 10K pot, making dB divisions non-linear. This applies throughout the circuit where there are 10K log pots with 22K to summing junctions.

 

[NewYorkDave]

R1 is not necessary

Well-spotted. I've picked up certain habits from designing with tubes and they do show up occasionally when I try to work in other media.

it will upset the log law of the 10K pot, making dB divisions non-linear.

True. If you have dB divisions marked in standard increments on a panel, you'd either want to use, say, a 1K pot for the returns (harder to drive for some sources), or use higher-value input and feedback resistors on the mix amps (more noise). If you went the latter route, you'd have to adjust the values in the panpot circuit as well since proper law there depends on the mixing resistors being equal to the pot value.

 

[therecordingart]

Any updates on this?

 

[enthalpystudios]

i was about to breadboard one with 5534's or a 5532, but then realized that

A) all my dual pots are log, regardless of value
b) it would be smart to build up like 8 and a couple makeup stages to see what its like to mix with it, as one channel wont prove too much.

but i really like the idea of going after a pcb for a guy like this. i already have started a console frame for api size cards, so as long it was less deep than those it would be fine... highly doubt it'll need more horizontal space per channel.

also maybe go for ic's for the buffers, but have DOA footprint option for the main gain stage and makeup stage opamps..... i have some (what i think are) cool ideas for a pcb, but i've never designed one, and its certainly not a good first project for me ;]

hopefully someones into the idea... either way, i'm going to find some quicko dual pots and have a go.

hey dave, this is for a virtual earth active summer right? hence the mix bus going into the inverting input, and all the mix bus resistors being equal to the mix amp feedback resistor?

if it is, based on my (limited) understanding of active summing like this, isn't there a certain way of tying the ground connection to the non-inverting input of the mix amp to reduce noise? i've read it in a few places, and just did a bit of searching, but its not speaking my language yet... isn't the idea to run shielded cable for each output to the mix buss, and tie all those ground together at the non-inverting input? or something.....

i just remember something about using that non-inverting input to your advantage by being selective about the source of that ground connection to cancel out noise on the mix bus, and also to keep all the crap current dumped into ground from other components in the system from causing distortion at the mix amp output.

but i agree, we should keep up on this guy. if i can help much that is.

 

[therecordingart]

Let's keep this going.....BUILD IT!!!!

 

[enthalpystudios]

so concerning pfl, whats a good way to swich the monitor source when any channel has its pfl engaged?

i thought to have a dpdt where one side connects the channel pfl output to the pfl mix buss, and the other side sends a voltage that triggers a relay to switch the monitor source from the main bux mix amp output to the pfl mix bus output.

but how to keep the voltage constant when engaging multiple channels to the pfl bus?

 

[the scum]

This looks a bit like a "jukebox" returns mixer I've been kicking around for a while.

If we can get enough people on board to make it worthwhile, and reach reasonable agreement on basic design and components, I'd be into designing some PCBs.

The mechanical design of the thing becomes pretty important. My first idea would be a card for each channel. Vertically, down the card would be the aux 1, pre/post1, aux2, pre/post 2, the panpot, and then the solo & mute switches. The fader would have a 3-terminal tag on the edge of the card, so you could use either a slide or rotary fader.

The back of the card (the side away from the panel) would have an IDC (Ribbon-cable) plug, which would serve to carry the power and summing busses between modules.

Lets see how many busses we'd need:
1 +v
2 ground
3 -v
4 aux1
5 aux2
6 left
7 right
8 solo summing
9 solo enable

And then at least double that: the power rails could get 2 or 3 lines each, and using the Forssell-style summing doubles the number of lines for the summing busses. Interleaving the real and "sense" busses might also help hold crosstalk down. Ribbon cable is cheap, let's use that to our advantage.

The channel card itself would be maybe 4.5", so it would fit behind a 3U rack panel...or something like the Paia/Blacet FrackRacks. So a 3U unit with the cards behind, with a second panel for the faders (sized to match the faders, naturally).

The summing cards would be if a similar design, and the usual design would get 4 of them: left, right, aux1, aux2. The cards would be the same, selecting which buss off the ribbon they sum via a jumper. There might be some slight tweaks to the cards...the auxes might get solo switches that the stereo busses wouldn't...just a pushbutton that doesn't get populated.

There might be one more card: a solo/control room card, maybe with a mono switch, input select, the master solo logic, and output speaker switching? What other facilities do people like?

What do people think of the following components:
-pots: the 1" garden variety Alphas from Mouser...about a buck per stage in quantity.
-switches: the ITT MTH series of DPDT pushbuttons...again about a buck each in quantity.


There are a couple of tweaks to NYD's design that my tastes lean towards:
-Rearrange the front end. Add some 22K resistors to turn the voltage follower into a unity-gain balanced input. The fader and auxes would hang off it's output. The mute would then only need a single pole of the switch. The other pole could switch an LED.
-The ground summing mix buss implementation (called impedance balanced, above) is attractive, for the cost of another line on the ribbon, and a handful of resistors.
-Move the 47R on the output inside the opamp feedback loop.
-Solo switches on the aux outputs.

And to answer a couple of Enthalpy's questions:

--Go read this whitepaper on summing. It'll answer some of your questions:
http://www.forsselltech.com/summing buss.htm

--The original schem is probably omitting some pieces of the solo circuits. The "PFL" output from the channel leads to another summing amp. The switch itself should be a double-pole, with the second pole used to drive a DC level on another buss, with another (DC coupled) summer, and maybe a comparator. That signal drives the select of a relay, which would switch from the mix to the solo signal. Clever wiring would put an LED next to the solo switch.

If we moved the solo feed to be after the panpot, and used a 4-pole switch onto a pair of busses, we could have in-place solo.

And there are a few other things that need to be fleshed out:
-A coherent grounding scheme...I keep catching myself designing in ground loops...
-Which slots would need a DOA/monolithic dual footprint?
-How many caps, of which sorts, where? How overboard should we go? Lytics with film bypass on each opamp output? Phase compensation on each feedback loop? Big lytics for power supply reservoir on each module of course, but also bypassed on each chip?

 

[lofi]

i got a feeling you guys are going to make this 'big'! can i play too, when its done, that is

 

[Mlewis]

Which slots would need a DOA/monolithic dual footprint?

i think it only really makes sense for amps with significant gain (such as the Mix Amps) to be DOAs.

i'd definitely like to see the Pseudo ballanced/impedance ballanced/whatever-its-called ballanced mix buss.

 

[Joechris]

Good to keep this simple, but a extra group or two is a nice mix tool.
Solo in place is good. PFL must be a leftover from the radio days.
JLM relay system for the PM 2000 seems like a nice solution.
One relay for both mute and SIP.
My two cents
Jo

 

[NewYorkDave]

I'll see about adding a subgroup and quasi-balanced (ground-sensing) mix amps, as well as addressing the issues Boswell raised. But I have many irons in the fire these days and little time to devote to some of these projects.

I won't get involved in "mission creep" beyond the items just mentioned, so it will probably be up to someone else to truly flesh it out into the more complex mixer some of you seem to desire. That does beg this question, though: why not just buy a mixer and bypass the preamps, if you need all that routing and monitoring capability?