Designing a modern mixing console - Part 1 (and introduction): Channel Input

[abbey road d enfer]

High cut filter is definitely an optional item IMHO.

I find the high cut to be quite useful on electric guitars, because I can reduce "fizz" while still having a band of eq to suck out some of the 4-8k resonance that always creeps up. With modern condensers and digital recording, I find vocals can often use some high cut as well. Probably only 6dB/oct slope here.

I agree with you; I must say sometimes I wished I had a variable slope on the LPF, like 2-3 dB/octave starting at 500 Hz or 1k.

The solo system is a bit mystifying to me, since it looks like some sort of combination of "solo in place" which works by muting channels, but at the same time has a stereo solo summing bus pair.

Yes, that is mystifying indeed! :-[ Clearly the solo bus is totally unnecessary. I started by drawing a PFL style setup, but then decided that I really actually hate PFL and much prefer solo-in-place. I then didn't get rid of the solo bus.

You may hate PFL, but I think it's an irreplaceable tool in many circumstances of tracking. I would have both SIP and AFL with two separate switches.

I vote for more than six auxes, but that is just me.  The Ameks I keep mentioning have 16 (!!) which is maybe too far the other direction.  Regardless, each aux send needs an individual pot, and then you get into how to deal with pre/post for each aux.  Pre/post for each aux?  "Group" the auxes so that, say, every two auxes has its own pre/post?  Also, many desks group the auxes into pairs with a pan pot for that aux pair.  One cool thing along those lines the 9098 desks offered on several aux pairs was the ability for the aux pair to be post-main panpot.

The aux sends are going to be a pain. I know it's been mentioned that they usually run unbalanced in the desk, and I know they're usually just for effects which don't need as pristine a signal path,

I don't agree with this statement. FX, and particularly reverb and delay have a tendancy to make minor noises more intrusive.

but I'd hate for noise/buzz to creep in to the system from that path. My initial plan was to run the auxes balanced. It would be easy enough if I had a balancing amp feed all the aux sends, but then I need balanced fader and pan pots for each send. I could avoid the balanced pots, but then I'd need a balancing amp for each send. 

I think you may run impedance-balanced aux bus. Nice compromise.

The desk is a bit of a unique bird; at least, I haven't seen this configuration. It's a bit of a hybrid between a split design and an in-line design. Basically, any input module can be a line input module or a group input module.

In fact, it's not very dissimilar to the Soundcraft TS24, which relies on the same assumption. http://www.soundcraft.com/downloads/user_guides.aspx?pid=95#downloads

 

[Nishmaster]

Revision 2 of the channel block diagram attached.

The insert pre/post switching has been fixed (I think), it now has direct outs, 2 more aux buses, and either solo in place or PFL.

Yikes. At this point, I count 93 bus lines:

L/R Bus x2
Group Outputs 1-16 x2
Group Inputs 1-16 x2
Aux 1-8 x2
SIP Detect
PFL Bus x2
PFL Detect
+18V
-18V
+5V
AGND
SW/LED GND

Plus 4 differential I/O to the outside per card, that makes 101 card connections. That's a hefty card edge or ribbon connection. With my lack of mechanical engineering skill, I may opt for multiple ribbon connections to a main bus board. I'd hate to get significantly far and then find out none of my card edges line up. I'm also thinking that I may go for multiple cards per channel module instead of one large one. That way, I can wholesale bypass/replace parts of the circuit should maintenance be required.

Why not use a small microcontroller to manage all the CMOS switches ? It is not really complex logic system. Maybe require some care to the pcb layout though.

I'd really rather not go down that route. Even a PIC, while cheap, is going to require a CMOS or relay per two inputs. Then it's not so cheap at all. Furthermore, while I am a decent programmer, my assembly is pretty rusty. Noise could also be a problem as well. My solution, while somewhat hokey feeling, does the trick with a minimum of parts.

I agree with you; I must say sometimes I wished I had a variable slope on the LPF, like 2-3 dB/octave starting at 500 Hz or 1k.

That would be nice. Half pole filter, anyone?

You may hate PFL, but I think it's an irreplaceable tool in many circumstances of tracking. I would have both SIP and AFL with two separate switches.

Thinking about it more, while I've seen the desk as more of a dedicated mix desk than a tracking desk, I'm inclined to think that myself and my fellow engineers at the studio will probably want to track through it, if only for the free eq. So, I've added the PFL and direct outs.

I think you may run impedance-balanced aux bus. Nice compromise.

That is a brilliant idea. I completely forgot about impedance balancing. This is the perfect opportunity for that.

 

[abbey road d enfer]

At this point, I count 93 bus lines

What"s more, you'll need duplicated or triplicated power rails and A LOT of AGND. Ground stiffness is probably the most important single element of performance in a mixer.

 

[Henke]

Getting there! Although, I think you would want to break the signal entering the PFL-bus when not activated...

 

[abbey road d enfer]

Yes, the solo does not need audio bus, only a sense bus. The solo switch should put a logic level on the sense bus AND simultaneously disconnect the Mute circuit from the sense bus.
And there should be a PFL audio bus and a PFL sense bus that switches the monitor source.

 

[Nishmaster]

Yes, the solo does not need audio bus, only a sense bus. The solo switch should put a logic level on the sense bus AND simultaneously disconnect the Mute circuit from the sense bus.
And there should be a PFL audio bus and a PFL sense bus that switches the monitor source.

Perhaps you are looking at the wrong revision Abbey. In the latest posted revision, the arrangement is exactly as you describe, although as Henke mentioned, I totally forgot to disconnect the PFL audio bus.

-Matt

 

[abbey road d enfer]

You're right, I had an older version.
Still, you need a switch between the PF signal and the audio bus.

If anyone has some idea on how to switch 16 balanced inputs while allowing only one input to be active at any given time without digital logic, I'm certainly open to ideas. Dual concentric 16 position rotary switch is about all I can think of.

Maybe use a CMOS MUX?
http://datasheets.maxim-ic.com/en/ds/DG406-DG407.pdf

 

[Brian Roth]

Wow.....a new domain name, and it looks like everything survived the "move"!  Earlier tonight, Matt's Rev 2 diagram was MIA, which made me glad I had downloaded it to my drive so I could look at it off-line.  But, looks like the earlier link works now.

One clarification on one of my earlier remarks, where I referred to "FET switches".  I was basically referring to the various CMOS chips (which are built around FETs internally), but discrete FETs could be also used.

Comments on Matt's Rev 2 signal flow.

1.  The insert in/out switch is not drawn correctly, and won't work as shown.  No biggie, basically just move the SPDT into the "receive" path vs. the "send".  I think my "RothCAD sketch" posted earlier in this thread shows the correct signal flow.

2.  I still don't see a polarity/"phase" switch.  If low and high cut filters are assumed to be useful in this app (vs done "in the box"), I dunno why something as practical as a polarity switch has been left out.

3.  Direct Out....a "good call", although one can argue "pre and post, this and that" for the function.  Yes, let's add MORE switching...LOL!

4.  Earlier comments about having a 16 position rotary switch, or the Maxim 16 "pole" MUX chip precludes any reason for the summing amp shown in the Rev 2 diagram.  Those two options are a "one of 16 choice" and allow no mixing of the various sub groups.  ***IF*** you require two or more of the subgroups to be mixed together within one module, then you need 16 individual mechanical switches, relays, or FET equivalents.  With the latter two options, you also need some sort of "control logic" to toggle the 16 FETs or relays within the possible 65, 536 combinations.

5.  Based on "# 4" above, why not just use the patchbay to poke the desired subgroup back into the module?  That's how I would do it with commonly available desks.

6.  Rev 2 also shows that routing to L/R is "either/or" with the 16 subgroups.  I can think of possibilities where routing to BOTH can be useful, like for parallel processing.  Indeed, trying to avoid internal feedback loops is A Noble Goal, but I can drive ANY desk into internal feedback loops with a few patch cables and a few button presses.  <g>

Best,

Bri

 

[Nishmaster]

Maybe use a CMOS MUX?

That solves half of the problem only. Since the input on those is a 4 bit logic input, if two switches are engaged, the cmos will be selecting the binary addition of both of the switches. The problem is not so much only activating one signal path as it is disengaging any other switches that may still be active.

The only other solution I see other than my cheesy sum amp is to use some sort of arrangement of momentary switches into a PIC, and then the PIC code outputs the correct signal for the mux to switch. Conceptually quite simple, but in practice quite above my head. I'll have to get into switch debouncing, assembly code for the PIC, how to keep the PIC noise out of the equation, and I'd also need to source one with enough analog I/O pins, 16 for the switch inputs and 8 (mux x2) for the ouputs. Then there's the issue of indication, which means I'd probably need another mux to switch a led driver to the correct led.

Wow.....a new domain name, and it looks like everything survived the "move"!  Earlier tonight, Matt's Rev 2 diagram was MIA, which made me glad I had downloaded it to my drive so I could look at it off-line.  But, looks like the earlier link works now.

When this is all said and done, I'll post this stuff up to my permanent website as well, so it doesn't get lost to history.

1.  The insert in/out switch is not drawn correctly, and won't work as shown.  No biggie, basically just move the SPDT into the "receive" path vs. the "send".  I think my "RothCAD sketch" posted earlier in this thread shows the correct signal flow.

I'm probably being dense, but I'm not seeing a signal flow difference between these two. In mine, I break the send amp out of the circuit, in yours, you break the receive amp, but that's the only difference I can detect.

2.  I still don't see a polarity/"phase" switch.  If low and high cut filters are assumed to be useful in this app (vs done "in the box"), I dunno why something as practical as a polarity switch has been left out.

I meant to include that, yes. No reason to leave such a simple item out.

3.  Direct Out....a "good call", although one can argue "pre and post, this and that" for the function.  Yes, let's add MORE switching...LOL!

Yeah, this path has just about enough switches for my taste. While I can think of reasons to have direct out pre-eq, they are vanishingly small. Since there's no mic amp, what would be the point of sending anything pre-eq out of the desk? Just send your rack pre to tape and leave the desk out of it.

4.  Earlier comments about having a 16 position rotary switch, or the Maxim 16 "pole" MUX chip precludes any reason for the summing amp shown in the Rev 2 diagram.  Those two options are a "one of 16 choice" and allow no mixing of the various sub groups.  ***IF*** you require two or more of the subgroups to be mixed together within one module, then you need 16 individual mechanical switches, relays, or FET equivalents.  With the latter two options, you also need some sort of "control logic" to toggle the 16 FETs or relays within the possible 65, 536 combinations.

If this was some sort of switch matrix configuration, I would agree. Even still, the maximum number of total configurations are 16x48, or 768 combinations. However, I don't see the necessity for control logic here for mixing within one module. Simply select any number of the groups and they all flow into the summing amp. Perhaps you were meaning "former" instead of "latter"?

5.  Based on "# 4" above, why not just use the patchbay to poke the desired subgroup back into the module?  That's how I would do it with commonly available desks.

This is certainly a valid thought, and I'm glad somebody brought it up as I was considering it also. Most current big name desks deal with groups in the way you describe, so perhaps I'm over-complicating things. The flexibility would be nice, but the cost-benefit may not make it worthwhile. It certainly is going to be a metric arseload of switches, to be sure.

6.  Rev 2 also shows that routing to L/R is "either/or" with the 16 subgroups.  I can think of possibilities where routing to BOTH can be useful, like for parallel processing.  Indeed, trying to avoid internal feedback loops is A Noble Goal, but I can drive ANY desk into internal feedback loops with a few patch cables and a few button presses.  <g>

Sorta. The channel outputs in to the groups can be multiply selected, so any one channel can go to multiple groups, including L/R main bus. The only time where this doesn't apply is when group mode is selected to allow a group input to the channel, then the channel automatically goes to L/R only. Are you saying it might be nice to have the output of a group go to another group instead of L/R? I have done that sometimes, but even with parallel processing I usually have to buses side by side into L/R as opposed to going into another bus then into the L/R bus. It's also possible I'm misunderstanding you entirely.

-Matt

 

[abbey road d enfer]

Maybe use a CMOS MUX?

That solves half of the problem only. Since the input on those is a 4 bit logic input, if two switches are engaged, the cmos will be selecting the binary addition of both of the switches. The problem is not so much only activating one signal path as it is disengaging any other switches that may still be active.

I was thinking using a grey-code switch. But I think I don't really get what you're trying to do. I read that you want mutually-exclusive but you want to have the possibility to add several...?

1.  The insert in/out switch is not drawn correctly, and won't work as shown.  No biggie, basically just move the SPDT into the "receive" path vs. the "send".  I think my "RothCAD sketch" posted earlier in this thread shows the correct signal flow.

I'm probably being dense, but I'm not seeing a signal flow difference between these two. In mine, I break the send amp out of the circuit, in yours, you break the receive amp, but that's the only difference I can detect.

I confirm it won't work, unless you consider joining the output of the insert return amp and the mic pre's output to be normal operation.

 

[Nishmaster]

I was thinking using a grey-code switch. But I think I don't really get what you're trying to do. I read that you want mutually-exclusive but you want to have the possibility to add several...?

No no, I'm not really explaining myself terribly well, my apologies. I was just debating the different approaches. There is no design necessity to have several groups summed within a module, that is simply a side effect of using simple pushbutton switches to select the group input via the design drawn. Currently, you press button 1, and group one gets sent to the summer at the front of the module. You press group 2 and now groups 1 + 2 are sent to the summer at the front of the module. Pressing group 2 does not cancel group 1. This is probably the second least expensive and complex option, with having all this be done on the patchbay as probably the simplest overall.

My ideal design goal would be to have the group inputs for the module be selectable by individual numbered pushbuttons which would somehow allow only one input to be selected at a time, ie momentary pushbuttons, some kind of microcontroller, and one or several muxes. You would press button 1, it would light up and engage the switch to select group 1 as the input of the module. You could then press, say, button 2, and button 1 would cease lighting up, group 1 would disengage from the channel, and group 2 would be selected. That is my fantasy interface for that part of the desk.

A grey-code 16 position rotary encoder to mux would work also (that's a great idea Abbey), with the caveat of then needing some kind of indication as to which group was selected. Dual numeric segmented leds, maybe, but that can get pricey, and we're back to logic of some kind.

I confirm it won't work, unless you consider joining the output of the insert return amp and the mic pre's output to be normal operation.

Er, well, no. Quite true. Getting ahead of myself as usual.

 

[abbey road d enfer]

My ideal design goal would be to have the group inputs for the module be selectable by individual numbered pushbuttons which would somehow allow only one input to be selected at a time, ie momentary pushbuttons, some kind of microcontroller, and one or several muxes. You would press button 1, it would light up and engage the switch to select group 1 as the input of the module. You could then press, say, button 2, and button 1 would cease lighting up, group 1 would disengage from the channel, and group 2 would be selected. That is my fantasy interface for that part of the desk.

I've done that eons ago for monitoring controllers. Originally used a chip made for TV channel selection, but when it became obsolete because all TV's had a built-in microcontroller, I did that with 4011's connected as RS flip-flops. A single 4011 would cover 2 push-buttons. I would need to think a bit or excavate my archives, but it's pretty much textbook logic. And yes, the way you've expressed it, it's logic; no wonder you have to have some logic in there...

 

[Henke]

Just an idea if you decide to go with the group/bus-strategy shown in v.0R2. If you add one more pole to the bus assignment switches and pass the corresponding group input through this pole (breaking the signal when engaged) before it reaches the group input selector, then you're free to route the group to all the other buses.

 

[Nishmaster]

Revision 3 of the channel block diagram attached.

I would need to think a bit or excavate my archives, but it's pretty much textbook logic. And yes, the way you've expressed it, it's logic; no wonder you have to have some logic in there...

I know, it's silly. I have a good programming background with a solid understanding of software logic, just no exposure to hardware logic. But, more so than logic, which isn't scary by itself, is the threat of noise it brings with it, as well as the necessity of yet another separate ground and dealing with clocking, which I have never done. As this is already a huge undertaking, adding yet another entire field of understanding to the project only serves to increase the probability that I will end up designing a piece of junk.

Just an idea if you decide to go with the group/bus-strategy shown in v.0R2. If you add one more pole to the bus assignment switches and pass the corresponding group input through this pole (breaking the signal when engaged) before it reaches the group input selector, then you're free to route the group to all the other buses.

That is indeed quite clever. I have to wonder if capacitive crosstalk within the switch would be a problem with that configuration, though. Even in the best circumstances some signal would get back to the bus, although at a much reduced level.

 

[Brian Roth]

A thumbwheel switch (hexadecimal coded) should work for the group input selector in conjunction with the Maxim 16 input mux chip discussed earlier.  But, this is a "one and only one" type of selection, so you can't sum two or more groups inside of an input module.  To do that summing feature, the Maxim mux won't work anyway; you would require 16 individual switches of some sort to select among the 65K+ (16 squared) possible combinations.

Also, a quick check of thumbwheel switch catalogs shows the numbering is 0-15 or  0-9, A-F.  Maybe someone sells a thumbwheel numbered 1-16.....

I still "vote" to make the group return a patch into the module, which eliminates the need for the switching AND the need for the "line/group" selector switch.

Best,

Bri

 

[abbey road d enfer]

Revision 3 of the channel block diagram attached.

I would need to think a bit or excavate my archives, but it's pretty much textbook logic. And yes, the way you've expressed it, it's logic; no wonder you have to have some logic in there...

I know, it's silly. I have a good programming background with a solid understanding of software logic, just no exposure to hardware logic. But, more so than logic, which isn't scary by itself, is the threat of noise it brings with it, as well as the necessity of yet another separate ground and dealing with clocking,

Do you know there's static logic (gates, diodes or relays sitting there doing nothing but wait for the push of a button or an electrical stimulus), as opposed to sequential logic (micro controllers with their noisy clocks). 

 

[JohnRoberts]

CMOS logic has coexisted with analog audio forever, and yes it mostly sits there doing nothing. The attraction to CMOS logic is the higher rail voltages and low current draw when not switching.

Micro clocks up at several MHz should not be a huge issue with careful handling.

The design heavy lifting will always be the purity of the analog path, when switching audio signals and the like.

JR

 

[boji]

The design heavy lifting will always be the purity of the analog path, when switching audio signals and the like.

Simplicity would win in the respect, right? Due to my inexperience with audio signals, I have opted to use relays over IC's to simplify what is around my audio path. Or when something cool can be done with logic, have that logic be off board, running to relays on board. This hurts pcb real estate and the pocketbook, but I think I'll run into to fewer problems with only two types of ground and power schemes to muck with. 

Have you considered going in this direction Nish? You could satisfy your programming itch with micro controllers and branch that out to a simplified, analog input card.

-Don

 

[Brian Roth]

Have you considered going in this direction Nish? You could satisfy your programming itch with micro controllers and branch that out to a simplified, analog input card.

That's how Otari (Sound Workshop) handled things on their Concept (and other) desks.  The "digi racket" cards were enclosed in a metal shell beneath the VU meters, and "static" logic control lines were fed through into the analog audio modules via an "Euro" multipin connector set.  Those control lines then switched on/off the multiplicity of 4053 CMOS switches, as well as receiving contact closures from various front panel switches.  Hence, no clocking signals on the analog circuit card.

Best,

Bri

 

[Brian Roth]

Well. this thread has died.....

question back to Nish/Matt....

I am thinking "further ahead"....

For the EQ section...why not use a 500 format module?

In your I/O design, it seems that a 500 could fit above/below the rest of the module?

Best,

Bri