Prototyping Nearness, a minimal panning mixer module


#41

coooooool
panning is the more powerful feature in signal routing :star_struck:


#42

The inputs are AC-coupled in the current design, but that’s something I’ve gone back and forth on. Making them DC-coupled wouldn’t be the only change necessary to turn the module into a good CV mixer, though.

For CV, I think it would be best to use constant-voltage panning, with the inputs spaced evenly between the two outputs:

For audio, constant-power panning is arguably better, and is what I’m using:

I’m not yet sure about the best placement of the inputs, but I know that spacing them evenly doesn’t work well.

So, even with DC-coupled inputs, it would be a pretty weird CV mixer. Plus, for audio, the mix would be more likely to clip and distort when incoming signals have DC bias. (I know of at least one popular Eurorack sound source module that has an inherent DC bias: the Bastl Grandpa.)

Of course, all of this applies only to the “standard” build of the module. If you DIY one, you can change the panning to whatever you like, and also make the inputs DC-coupled (by replacing the capacitor at each input with wire).

Thinking about this made me realize that even “normal” Eurorack audio signals (10Vpp without DC bias) might cause the mix to clip, if they are strongly correlated, e.g. from oscillators tracking the same voltage source. To fix that, I’ve added jumpers to the second revision of the board, that let you attenuate all signals by 6dB, much like on the Intellijel Unity Mixer.


Expander pins, then. At first glance, they seem like a great idea, but for a module like this one, they turn out to be a lot of trouble. (This is the kind of design issue I enjoy reading about, so I’m gonna brain dump a bit here.)

First off, where in the circuit would the pins go? There are two logical places:

  1. In series with each input. This would make possible expanders that do per-input attenuatuation or muting. That means 7 “send” pins, plus 7 “return” pins, plus ground, which adds up to 15 pins, which you’d probably do as 2x8 pins, the same size as the larger variant of the Eurorack power connector.
  2. Added to each output. This would enable chaining multiple modules together without using patch cables. For that, you’d need 2 input pins, plus 2 output pins, plus ground, meaning either 5 or 2x3 pins.

For 7:2, I had to split the circuit across two boards: a front board with all the jacks and a bit of passive per-input circuitry, and a back board with the active circuitry and the power connector (you can sort of see what this looks like in the photos in the first post).

Now, this is hard to know without having seen the schematic, but if you wanted to add expander pins of the first variety, they would need to go on the front board (because on the back board, the signals are already mixed to two buses). But the front board is completely packed - there is simply no space for anything that large. Of course, that could be solved by making the board larger, but then it would no longer fit behind a 3hp panel. Another problem, as identified by @Leverkusen, is that the circuit would still need to work with no expander connected, which means you’d need another 7 jumpers just to make the module work, and if you lost any of them, you would also lose an input.

Expander pins of the second variety don’t present a problem in terms of size: there’s fewer of them, and they would go on the back board, which can be made a fair bit larger without affecting the width of the module. However, making the circuit work equally well with and without an expander connected is trickier here - the expander pins would either need to be buffered multiples of the outputs (which would require more circuitry), or take advantage of the switching on the output jacks, so that inserting a cable in the output jacks would prevent the signal from going to the expander pins. That in turn would require a few more pins connecting the front and back boards. Doable, but not without drawbacks. (And, you could get the same functionality using e.g. a Unity Mixer, at a “cost” of two-thirds of an hp per module.)


#43

One cup of tea and a little bit of math later, it’s DC-coupled again.

It’s nice that the module works both with audio and CV (if a bit curiously), and that it’s simpler to build and modify; and I don’t think that you would generally have a ton of signals with significant DC bias in them, and if you do, 6dB of attenuation may well keep the mix from clipping anyway.


#44

Well, if it clips, as long as it clips interestingly I’d consider that a feature :).


#45

…but if you do, and I experience this sometimes in a modular environment, is it a good idea to have different DC biases on individual stereo channels then? I would be afraid that it alters the stereo image and might let you experience unwanted phase effects. If you want clipping you can always do it later and apart from being easier to built I don’t see much use in a CV mixer with fixed stereo outputs - apart from happy accidents.

What’s wrong with a simple fixed panning mixer? It does not have to do everything but fits everywhere as the Intelljel Unity Mixer does for your CV.


#46

do you think DC offset would be a problem though? i imagine this will likely be used at the end of a chain with a filter in there somewhere that should remove any DC from audio signal.

what i like about it being DC coupled is that it benefits from the beautiful simplicity of the concept as well, having 2 modulation sources that mix from 7 others where you control how much of it goes to each channel simply by repatching into a different input. that and the fact that it would pair so well with just friends in envelope or LFO mode…

i do agree though that if it makes it work less well for audio signals then maybe it’s not worth it, or maybe it could be a switch on the back or 2 different versions, one for AC and one for DC.


#47

I think priority should remain audio mix/panning… that is what made the original module and the design idea so appealing!

Certainly a CV version of the module can be also designed and made, as it’s own thing…

But what I really like about the original idea is it’s minimal elegance.
Dare I say: very much feeling like something that effortlessly fits in the monome universe.


#48

Following this with great interest!
I planned to build an euro Mid/Side encoder this afternoon, (Worng LRMSMSLR), your panning mixer would be a perfect match! :slight_smile:

Personally i’d also prefer it to remain an audio mixer, with constant-power panning.

Then of course the ideal would be to have a PCB that could accommodate both constant-power & constant-voltage panning, and the builder could chose to build it for audio or CVs, but I have no idea if this is possible (?).
Having a jumper on the back that switch to one or another (bypassing AC-coupling caps when using constant-voltage), would be cool too, though I’m not sur it’s worth the trouble (can’t really imagine how i’d use it with DC signals…).

Edit: Sorry, i just saw that scanner_darkly suggested it already.


#49

How are things going with this guy? Planning on releasing and/or open-sourcing any time soon? I’m speaking primarily based on my own greed of wanting a few in my rack - but I’m sure others have been wondering as well. :wink:

Thanks again for the cool idea and sharing your thoughts and progress! :slight_smile:


#50

​​​​​​​​​​​​​​​​​​​​Yes. :slight_smile:


#51

Now.

The design can be tweaked indefinitely, but it’s now at a point where I’m happy with it, and am happy to share it.

Most importantly, I’ve moved the output jacks to the opposite ends of the panel, as originally suggested by @laborcamp. I am really happy with this change, and how it strengthens the underlying metaphor: the closer an input is to an output, the greater its influence on it.

Many small tweaks. There are jumpers on the back board to optionally attenuate the mix by 6 dB. The back board is bigger, and some of the components larger, for easier soldering.

The whole thing is of course the pan positions of the inputs. I’ll write more on that later.

If you’re feeling adventurous, you can find the project files on GitHub. You can get boards made wherever you like, but for convenience, here are OSH Park links for the front and back boards. You need one of each.


How do I achieve that lovely matte finish on eurorack panels?
#52

Cool! It’s gorgeous! Ordered 3 boards to Berlin! (Anyone in Berlin who wants to build this as well?)


#53

If anyone in the UK is building one of these and wants to build another one, drop me a PM.


#54

I might actually be interested to build a few of these, if anyone else is up to get some! Might be worth doing some group order for the board/components as well if people want to do the DIY route.


#55

So awesome! Thanks for all of the hard work and sharing the project so openly! This is a wonderful
addition to the ecosystem. I’m super excited to add a number of them to my cases!!

I’m also planning on building some units and could be convinced to share a few (or more than a few). :wink: :wink: I’ll post an interest link after I’ve built the prototype and have formed up my plans.

Thanks again!!!


#56

It’s so … elegant. Well done!

PS: If anyone is building them, drop me a note as I have no talent in that area.


#57

anyone in the USA planning to order boards and will have a spare set?


#58

Nice work. Any hints on where to order front panels from? Using the .svg on Github.


#59

Wonderful work, thanks much for sharing!

If anybody here in Germany would be so kind to build one for me, please drop me a line!


#60

Hmm having trouble sourcing the 81-GRM21BR61E226ME4L caps. Anyone got any ideas where I can get some from, or suggest alternatives?