Analogue Summer idea

[tomfairbairn]

Hi all,

This is my first post on GroupDIY, so please be gentle!

By way of introduction - I am a mix engineer (lots of live/FOH work, also recording/production and quite a bit of remote mix work from my modest home setup which is Logic Pro X on a macbook, decent monitors etc).

Must confess to having very limited electronics knowledge, I'm very much an 'end user', although interested to learn!

I'm keen to start building a setup with external preamps and processors, although budget is pretty tight and I'll need to do this over time.

I've been thinking about some variations on an analogue summing unit, which would facilitate use during tracking as well as mix. Here's a brief summary of what I have in mind -

- summing of 16 channels to a master bus with transformers on LR bus
- switchable inserts to incorporate outboard gear (comps, EQs etc)
- broad stroke tonal shaping per input, on switches rather than pots for fast workflow and easy recall
- channel outs to allow use during tracking and mixdown, as well as analogue summed LR
- monitor outputs (ideally including a solo bus)

Having researched online, I reached out to Ian Thompson-Bell who has been very generous with his advice and who kindly put me in touch with Cemal Ozturk of totalaudiocontrol.com who has also been incredibly gracious sharing ideas and options!

Ian suggested I mention the project here to see if anyone has thoughts, questions or can point me towards relevant resources I'll be able to learn from? I know I'll need to first work out a block diagram of the unit, then work out the circuits I needed and possibly see if anyone can suggest available schematics, PCBs or suppliers who have any suitable elements available.

I should mention I have 1 or 2 friends with electronics experience who may be able to help me with the physical construction/soldering etc, but they're not confident on the design side in terms of audio!

All ideas would be welcomed and gratefully received

Thanks all for the huge wealth of knowledge shared here & again thanks to Ian and Cemal for being utter legends!!

Cheers,

Tom
(in Edinburgh, Scotland)

 

[Bo Deadly]

If you have limited electronics knowledge, I don't recommend that you start out creating a 16 channel mixer.

Unfortunately there really are no shortcuts. If I could go back in time and tell myself how to get started, I would say get a USB audio interface and some software like Room Eq Wizard and figure out the details of how to take audio measurements. Specifically, you need to be able to get an FFT spectrum so that you can see noise and distortion of a device. Try measuring some devices. Then build something simple and figure out what it takes to make it's noise and distortion low.

The point of this exercise is that you should realize that there is much more to a successful build than just what you can draw on a piece of paper. You have to be cognizant of impedance at each net and the effects it can have on noise. Specifically, a high impedance net can easily pickup electromagnetic noise from other places like mains cables in the walls. So the layout of parts, shielding, how the wires are dressed and where and how the grounds converge all have to be exactly so or you will get hum.

Start with something small. Make that work perfectly and know it is because you can measure it and know why because you spent the time to understand it. That's where the satisfaction of doing this comes from. Otherwise you could literally buy a 16 channel mixer on Ebay for $300 that will out perform what you made.

 

[Lerok]

As far as places to start with working on a tight budget, I’d recommend a preamp based on something like the THAT 1502 IC - it’s fairly affordable, and doesn’t require an input transformer (unlike most vintage pre designs)

Similar to what square wave suggested, familiarizing yourself with measurement tools will go a long way - there are ways to do ffts with various pieces of software (python, matlab, etc) for maximum control, but with some Google fu it can be done in even excel

And I agree with squarewave, start small - a summing mixer isn’t the hardest project, but can get very confusing very fast

 

[Squeaky]

I started a 16 channel mixer about 6 years ago, nothing too fancy. It is modular and all discrete though. I might finish it next year if I am lucky. When it is finished it will be nothing but an ugly prototype (hopefully it will work better than it looks).

Good advice from Squarewave about audio measurements. One of the problems I have found is that it is really hard to test a mixer, if you only have bits of a mixer. I would do it differently over again.

If you are determined, you could start by trying building the input to a prototype channel strip (the simple line level eq with switching?). Then add a router? How many buses? Pans? sends? What summing approach? Insert points switchable or just switchable in and out or both? Do I need buffering? Where do I need buffering? Impedance, impedance, impedance...

As an exercise, try and conceptualise the grounding scheme.

 

[ruffrecords]

I have had a fairly detailed chat with Tom. He has friends with the necessary expertise to build and test the electronics and his basic block diagram for the mixer is sound. What he needs help with is in identifying suitable off the shelf circuits (preferably PCB based) that he can obtain and have built. It is all line level with no EQ. Ins and outs are on DB25 and there are sensible insert points. I think he probably needs three building blocks:

1. Line input unity gain amp capable of driving a fader and pan pot.
2. Bus amplifiers
3. Output amplifiers

I suggested he ask here simply because I thought the folk here would know better than anyone what (semiconductor) building blocks are available. How about JLM audio Palpigrade for the line in buffers?

Cheers

Ian

 

[sahib]

I have told Tom that I already have small, stand-alone dual opamp boards with which he can do all of the circuit blocks you mention. I suggested that he drew a block diagram of his signal chain and then identify the circuit blocks he needed.  The picture is attached. It sits vertically, or horizontally depending on the orientation of the header. Many stages can be built literally on a perfo board and hard wired in between.

 

[jacomart]

I recommend you to read this series of articles about mixing consoles design, you can get a lot of interesting ideas there! Oldies (But Goldies!)

https://groupdiy.com/index.php?topic=77353.0


Cheers
JM

 

[Bo Deadly]

Again, I don't think block diagrams are a problem. The problem is going to be laying everything out so that it's noise free. If it's not, what is the point? Just getting power to everything with return follows supply rules is not trivial. If its a conventional virtual earth summer, then the summing bus connections and ground bus need to be super solid and clean. Mixers are not a good first project. Make a DI. Make a THAT 1512 mic pre. Do something quick and easy first and get the satisfaction of actually using it this year.

 

[wlinart]

I have had a fairly detailed chat with Tom. He has friends with the necessary expertise to build and test the electronics and his basic block diagram for the mixer is sound. What he needs help with is in identifying suitable off the shelf circuits (preferably PCB based) that he can obtain and have built. It is all line level with no EQ. Ins and outs are on DB25 and there are sensible insert points. I think he probably needs three building blocks:

1. Line input unity gain amp capable of driving a fader and pan pot.
2. Bus amplifiers
3. Output amplifiers

I suggested he ask here simply because I thought the folk here would know better than anyone what (semiconductor) building blocks are available. How about JLM audio Palpigrade for the line in buffers?

Cheers

Ian

JLM also has the aux8 and sum8 building blocks for a mixer.

 

[sahib]

Again, I don't think block diagrams are a problem. The problem is going to be laying everything out so that it's noise free. If it's not, what is the point? Just getting power to everything with return follows supply rules is not trivial. If its a conventional virtual earth summer, then the summing bus connections and ground bus need to be super solid and clean. Mixers are not a good first project. Make a DI. Make a THAT 1512 mic pre. Do something quick and easy first and get the satisfaction of actually using it this year.

I hear what you are saying.

But Tom's problem is that he does not know what he exactly needs. So, get something on paper so that we could all comment and develop it into something. I have already given him the buss structure of a rack I designed a couple of years ago for a rack mount mixer as a starting point. It is in one of the topics here. So, he does not need to do anything about that.

 

[analag]

Many ways to do the same thing. Not saying how you should do it, but I would do it like this. Quantum 4848, into passive summer, amplified by something like a John Hardy M1. Panning done in software, Quantum allows you to bypass pre/ line amp and go straight to converter. Sounds awesome. I did it like this. Bunch of hardware pres and compressors (many diy because I much prefer how my stuff sounds, lol) into Quantum TB into passive box into custom solid state M1 style pre. From there I can patch into a couple tube and a VCA compressor which included the 30 tube monstrosity, from there I return to 2 in 6 out distribution mixer so I can drive headphone amp, monitors etc. Everything is 100% balanced. I chose the M1 style pre because it presents transformer action at it's finest level. 1:2 in and 1:1 out. The ugly sound of the high ratio winding is circumvented leaving you with size, depth and all the other good stuff we like.

 

[Newmarket]

If you have limited electronics knowledge, I don't recommend that you start out creating a 16 channel mixer.

I have to second that. I see that the OP has a good idea and resources/people to help but still...16 channels...

I'd suggest to go for eight channels at most first. It's reasonably likely that you'll find something that you want to change - physical channel width / ergonomics etc apart from the electronic details. Having got an eight channel block to your satisfaction a second one could be built to give sixteen channels in all and they could be 'bucket summed' . Then you can call it your 'Mini-Focusrite' 

On the summing detail. Assuming it's not using 'passive' summing then I say look at some form of balanced summing bus/amp. Take a look at the DDA stuff.

I'd also like to congratulate the OP for the correct spelling of "Analogue"  ;D

 

[Bo Deadly]

I have to second that. I see that the OP has a good idea and resources/people to help but still...16 channels...

I'd suggest to go for eight channels

Clearly I'm not breaking through on this one.

I'd also like to congratulate the OP for the correct spelling of "Analogue"  ;D

Ehhh, pronouncing it analogooee doesn't roll off the tongue very easily 

 

[JohnRoberts]

Look for an existing design to copy and just do it... you will learn what you didn't know by doing.

JR

 

[ruffrecords]

Clearly I'm not breaking through on this one.
Ehhh, pronouncing it analogooee doesn't roll off the tongue very easily 

Huh?? This is English we are talking about. There is near zero correlation between the spelling and the sound. Only Americans use phonetic spellings  Think about through, rough, thorough, though - they all end in 'ough' but everyone is pronounced differently. What about no and know; completely different spelling but sound identical. I sometimes wonder how English became so widespread.

Cheers

Ian

 

[tomfairbairn]

Hi folks,

Thanks so much for all the responses! Great to see a variety of opinions and approaches

I don't want it to seem like I underestimate the amount of knowledge and skill that goes into a complex build...not at all, I have the utmost respect for this! Therein is the challenge as my interest is primarily as a mix engineer, with a specific workflow in mind.

The aim of this project is very much to figure out if I can create/afford a design that could be built primarily by a friend that will tick the boxes I want as an 'end user'. Along the way I'm keen to learn what I can 'under the hood' but the goal is a unique piece of hardware that encourages a workflow, rather than the goal being for me to get to grips with audio electronics!

I totally appreciate there are no shortcuts and a lot of hard work


To clarify a few small points:

- I do have a firm idea of what I want to achieve in terms of audio signal flow (inputs, insert points, simple internal EQ, buses, outputs etc). What I don't know is how this looks from an electronics POV or at circuit level

- I'm not intending to build any mic preamps, these would all be external and connected at line level (potentially in one of the TAC 500 series racks)

Ian mentioned three potential semiconductor building blocks:

1. Line input unity gain amp
2. Bus amplifiers
3. Output amplifiers

I had already seen the JLM "Palpigrade" product, although the dual opamp boards Cemal mentioned seem like a great option if 1 simple board type can be configured for all of these tasks? I will try to calculate how many of these boards I think I'd need!

I know the JLM product is 8way per card and even with that I would need a lot of these circuits for the spec I have in mind!! (effectively 16:2:2, with direct outs and the inputs derived from 32 channels of D Sub inputs, switchable between A/B sources)

Does anyone have experience of the JLM aux8 and sum8 boards, or an alternative suggestion?


Thanks @jacomart for that suggestion I am reading through the pdf you linked!!


I'll keep researching and hope I may be able to visit Cemal in Glasgow at some point. In the meantime I'm certainly open to any other questions, suggestions etc

Cheers, T

 

[Newmarket]

Clearly I'm not breaking through on this one.

If you mean that's not 'starting small' as you advocate then I'd agree and would also advise that as best course. But OP seems to have greater short term ambition so I was compromising there particularly in view of what Ian said about their discussions.
I assume in reality that development would be incremental in any case eg build a 'channel' and test input to output. Duplicate channel and then sum both channels to establish functionality and performance and so on.

 

[analag]

A Mackie CR1604 is a cheap alternative, but if you're just itching to get a custom rig then money and time is all that is required.

 

[tomfairbairn]

Taking to heart some of the advice I have received, I've had a go at a rudimentary block diagram showing signal flow for a single channel

I would be grateful to hear from anyone who's able to suggest available PCBs that might be helpful

Happy to answer any questions of course

Cheers!

Tom

 

[wlinart]

I'm gonna repeat the suggestion of JLM, because they's easy to make and not too expensive.
For the input switcher, you could use this one: https://www.jlmaudio.com/shop/8-channel-relay-pcb.html with this one: https://www.jlmaudio.com/shop/tascam-/-idc-to-xlr-/-trs-in-out-pcb.html for the xlr connections. With 2 of these cards you could also build the insert relays.
And then for the balanced to unbalanced you could use this: https://www.jlmaudio.com/shop/palpigrade-8-channel-balanced-in-to-unbalance-output-kit.html. For the direct outs, this one: https://www.jlmaudio.com/shop/tardigrade-8-channel-unbal-to-balance-output-pcb.html and for the summing https://www.jlmaudio.com/shop/sum8-pcb.html?display_tax_prices=1, which would give you 8 channels of summing, while you need 5 (2 x stereo mix bus, 1x solo).
The big advantage of using all these is that they interconnect with a single cable. That makes wiring super easy. This is as far as i know the closest you will get in kit form to what you want.