Makeup Gain Stage for Passive Summing Mixer Using ETI 430 Line Amplifier

[JohnRoberts]

The floating un-used inputs was meant for active summing,

a workable technique for active combining to preserve differential ground referencing, is to back ground de-selected feeds. Of course the noise gain stays the same, but the differential math is still OK. .

you can do the same thing with passive summing but the level will change, I still think that after all that Ian has commented, its more about "patching up " the circuit so it works as good as possible  and the level differences are less noticeable, whilst doing active summing is pretty straight forward. All in all I see more disadvantages with passive summing than active summing, but to each his own.

I have followed the fad (sorry) of passive summing...

Ian nailed it... Its about using non-linear preamps with euphonious colorations.

JR

 

[cpsmusic]

Hi Everyone,

Didn't realise that my question would start such a detailed discussion! And no, I'm not scared (well, maybe a little bit  ).

So to answer a couple of questions:

Lots of people are using passive summing mixers, from DIY through to ridiculously expensive, and all claim that they make their mixes sound "better". I'm skeptical but as they're relatively simple to make I thought I'd try one myself.

The reason I'm interested in the "less-than-perfect" make-up stage is because I'd like to be able to drive the mixer into gentle distortion. I'm still not exactly sure about the best way to do this. Maybe an amplifier that distorts at the end of its range. Or possibly a guitar pedal like a Hudson Broadcast or a Mona Lisa. Both of these supposedly produce light distortion reminiscent of vintage mic preamps.

Cheers,

Chris

 

[cpsmusic]

BTW, has the forum been down lately? I haven't been able to connect for a few days. Seems ok now.

 

[cpsmusic]

Forgot to mention that another thing that appealed to me about the ETI430 was that it runs off a 9V battery.

 

[ruffrecords]

Forgot to mention that another thing that appealed to me about the ETI430 was that it runs off a 9V battery.

Glad you are backl.

What are going to feed the output of the gain make up amp into.

Cheers

Ian

 

[efinque]

I built the summing bus like this.. it's stereo with 5 channels plus separate L/R grounds.

It's made of tag board with plastic PCB "turrets" and hookup wire. Those are very closely matched 10kOhm resistors, but I changed them to 220ohm afterwards (it's easy to play around with different values but matching them is another thing)

 

[cpsmusic]

Glad you are backl.

What are going to feed the output of the gain make up amp into.

Cheers

Ian

The plan is to send it back into my computer via a Scarlett 18i20 audio interface. It will probably go back in via some line level gear (Warm Audio WA273-EQ, DBX-162SL).

 

[abbey road d enfer]

I think summing resistors are mainly used for reducing crosstalk between channels, and to some extent matching the mix bus level to the output

You are definitely wrong about this. resistors are needed because you can't connect signal sources in parallels -unless their own impedance is large, whivj is not the usual case.

You don't necessarily *need* them

yes you do.

 

[user 37518]

I have followed the fad (sorry) of passive summing...

Ian nailed it... Its about using non-linear preamps with euphonious colorations.

JR

I thought you were more of a "transparent" kind of guy. But thats ok.

 

[JohnRoberts]

I thought you were more of a "transparent" kind of guy. But thats ok.

I am.... Not only have I never seriously considered making a passive mixer, I am not a fan of audio paths with intentional distortion, euphonious or otherwise. I always cook with a clean pot, and then add spices.

When the popularity of passive mixing arose (years ago) I paid attention to see if there was any there there to support the claims. In my judgement there isn't.

I wrote an article in 1980 describing a better way to combine channels, but even that is made moot by modern digital technology. Now decades later I am pretty used to people discounting my advice.

If the OP wants to roll his own using a low-fi discrete transistor gain stage, enjoy. Don't forget to mention that the two transistor gain stage is class A. That should impress somebody.  8)

JR

 

[cpsmusic]

I have another question - if I wanted to add a Gain control (rather than the 20/40dB switch) how would I do that?

 

[JohnRoberts]

I have another question - if I wanted to add a Gain control (rather than the 20/40dB switch) how would I do that?

Putting a 10k pot in place of R8 (rheostat wiring), will give you roughly 20dB to 40dB range.

JR 

 

[abbey road d enfer]

Those are very closely matched 10kOhm resistors, but I changed them to 220ohm afterwards (it's easy to play around with different values but matching them is another thing)

220r seems a bit extreme, since the sources will see a load of about 500 ohms, which they may not feel good about.

 

[Newmarket]

BTW, has the forum been down lately? I haven't been able to connect for a few days. Seems ok now.

I think so. I also couldn't connect for a while.

 

[efinque]

220r seems a bit extreme, since the sources will see a load of about 500 ohms, which they may not feel good about.

The inputs are fed with mic/phono/line preamplifiers via 100kOhm pots.. the summing bus then feeds a bypassable master EQ (which I believe is +/-0dB) and an additional line preamplifier (that has trimmers) which goes into the mixer outputs.

The EQ may take it but I was a bit worried about the master preamp, however it did quite fine (iirc at some point I had resistors in the bus output, I was thinking 1kOhm summing would do but you can always ease a bit on the channel trim/volume, the trim controls affect the PFL volume though)

It may seem a bit odd to amplify a signal and then at the next stage limit it with resistors but afaik it's how analog mixers are designed (the crosstalk, summing to mono prior to the bus, matching impedances etc)

Sourcing a suitable preamp is another thing.. the one I used throughout the build is now discontinued/obsolete. It depends on what your inputs are, eg. mic/instrument (which is when you'd look at low R values if it's a passive bus, unless ran with an active DI) or line level (which is when you'd need larger R values and more gain to compensate)

Summing busses with mixed-value resistors are called scaling (ie. you'd know what's plugged into it and intentionally pick suitable R values, although there's a risk of crosstalk due to some channels bleeding into the others, say in a stereo bus the mic inputs are 1kOhm and the rest of the bus is 10kOhm.. the mic channel goes into both L/R which forms a 2k shunt, one way to work this around is to use a pan pot I think or not to max out your mic vol if it's a dual gang pot)

Or use diodes.. forward voltage becomes your enemy as you need an active stage plus it may gate the signal I think (~0,7V is lowest I've seen, anything below that will be cut off as the diode doesn't conduct)

 

[ruffrecords]

I am becoming confused at to the overall scheme of what you intend. 100K pots seem out of place to me. Can you post a schematic or at leas a block diagram of the while thing?

Cheers

Ian

 

[Winston OBoogie]

Haven't read this whole thread but saw Steve Dove's name crop up.
I believe it was Steve who pointed out some decades ago that, all other things being equal, there is a slight noise advantage for voltage summing when channels summed is less than 10. 

On forcing the bus low to ameliorate the changes in gain when channels are switched in or out, besides old Neve desks, Dick Swettenham (Helios) wrote about this in conjunction with his "free grouping" idea whereby any input channel (with mic amp) could pick up a bus and be used as a group module.
No two Helios desks were alike and most did not use this technique, but I have seen one that did so there may have been a few others.

There are lots of folks that make great records that now mix solely in the box.  There are probably just as many making great records who use some type of external boxe(s), either for specific sub groups, and/or to sum the final stems.

Only about .0001% of the record buying public will care how you got there so, whatever works. 

 

[Winston OBoogie]

P.S.  The earliest form of current summing I've seen was from the 1930's.
In its simplest form, imagine a series of triodes, each having their own bias scheme and input signal.  Tie all the anodes together and use one common resistor (beefy) for the current to voltage.  Parallel Plate mixing is how RCA described it. 
There were more elaborate forms of this,  but they are probably TMI for here.

Another ancient form of mixing used a transformer for combining 2 sources, a hybrid coil.  The REDD.37 and REDD.51 desks used this in conjunction with constant Z mixing.   

More useless, but may be interesting info from the digits of D.J.H.

 

[ruffrecords]

Haven't read this whole thread but saw Steve Dove's name crop up.
I believe it was Steve who pointed out some decades ago that, all other things being equal, there is a slight noise advantage for voltage summing when channels summed is less than 10. 

On forcing the bus low to ameliorate the changes in gain when channels are switched in or out, besides old Neve desks, Dick Swettenham (Helios) wrote about this in conjunction with his "free grouping" idea whereby any input channel (with mic amp) could pick up a bus and be used as a group module.
No two Helios desks were alike and most did not use this technique, but I have seen one that did so there may have been a few others.

Just to pick up on the Neve aspect, Neve used passive summing in all their classic mixers from the early days. I am not quite sure exactly what you mean by force the bus low to ameliorate gain changes but they did not slug the bus to do this. Passive mixing works fine without any bus slugging or gain changes as long as you ensure each bus resistor is fed from a very low impedance. Neve typically used 15K bus feed resistors and their bus driving amps typically had an output impedance of a few ohms. Unselected bus feed resistors were grounded so again they had near zero source impedance. I use a similar technique in the simple versions of my tube mixers where the bus is actually slugged with the 10K bus master fader and nothing else.

The typical 10468 transformer used across the bus as the input to the Neve bus amp worked best from a source impedance of about 300 ohms. The bus was slugged to meet this requirement.

Cheers

Ian

 

[abbey road d enfer]

It depends on what your inputs are, eg. mic/instrument (which is when you'd look at low R values if it's a passive bus, unless ran with an active DI) or line level (which is when you'd need larger R values and more gain to compensate)

The more you write, the less I understand. Is it the mark of genius?

Summing busses with mixed-value resistors are called scaling

In 50+ years of pro audio, it's the first time I meet this concept. I love it, an old dog like me learning new tricks.

Or use diodes.. forward voltage becomes your enemy as you need an active stage plus it may gate the signal I think (~0,7V is lowest I've seen, anything below that will be cut off as the diode doesn't conduct)

Now you are getting absurd...