That would work, but i for some reason thought the 3x/1x stage would be pre-mix.
If this is to still make sense, the 3x/1x switch would have to manipulate R2 (R39 in your schematic) otherwise the clipping will provide no help in the 3x case.
I have not tested or simulated this specific circuit, but apart from that it looks good
The 3x soft key function gives the extra effect. Lars said they use this in their modules too.
Maybe I did not understand this correctly. There are some clipping blocks in those modules.
Color Chords example:
I just want to make sure there is no 9v situation possible, so that is why I placed the soft clip portion there (where the max voltage would be +3v)
But I’m really in new waters here, so correct me if I’m wrong!
For the 3x / 1x switch I have to choose a location where I don’t have to switch between too many resistors (so after the mixing stage) so I can use a simple switch (spdt or dpdt) to change the whole ‘column’ at once.
This could in theory also be at the first mixing stage.
Someone else might be able to chime in on exactly what the LZX designs do, but i think what we want is for the gain to be pre clip.
This should be possible by having the switch switch between R2 values which is possible with a SPDT or even SPST
ok. I’ll try some setups in the simulator.
The Cadet II encoder has a brilliant clipper circuit that I have re-used in a few of my designs. It’s definitely worth spending the time to wrap your head around its subtleties and cleverness.
here is a simulation. it clips the top , but not the bottom of the waveform.
I still think the cadet precision clippers are a bit overkill for this application, but the problem you’re seeing is because you are measuring the output at the wrong node
ha, you are correct.
this is better, but I can’t seen to get the right clipping settings for a ‘max state’ of the matrix mixer, as in 9 volt. even 3 volt is turning the waveform in a squarewave
I’m going to try your suggestion, and think about it some more.
It’s going to need some tweaking, but i really like the idea of being able to overdrive the mixer and then attenuating it just a bit back down to get some nice soft nonlinearity.
As long as the parameters are right it would still be (mostly) linear from 0-1V for a more clinical standard case. (while still keeping the output from going too far outside the limit)
This seems to work (with the gain switch), I not math whizz: I do not seem to be able to get into the 1v range.
note: Not sure if the switch config is correct yet. was this what you meant?
edit: this one is a bit better.
with an extra gainstage at the end. this works pretty good in most scenarios. but I have the feeling this method is not optimal, I mean, it might be possible to use less opamps.
but it has interesting waveshaping options when using the attenuverters.
I was thinking pretty much this
ah, that is clever! I will simulate it, to see what this does to the waveshapes and levels
this works. the ranges are a bit fiddly. but workable I think. playing with the values a bit more…
now I have to see where I put the gain pot. it was supposed to go where the 1.4k resistor is now. but that won’t work I guess. mmmmh
I’m not sure if the master gain being post-limit is such a bad thing, as that would allow you to attenuate after the soft clip?
Jack connection ideas:
A: there are 3 inputs, I can give them ‘direct multiple outputs’ , which save you patching time.
The ‘direct multiples’ would just be connected to the input jacks. like regular multiples
I have to see if I can work something out with the normalization, as I wanted to connect +1v to the switch of the 3 input jacks, for matte color outs
B: the output jacks. I can make multiples here too. for easy patching
like this:
this would mean that there would be 12 jacks.
a) I don’t think the final set of amps are needed, the output of the last variable gain stage is already a strong driver.
b) Any specific reason why the pot in the final gain stage is in the feedback loop rather than just scaling the input to ground before the amp?
c) Rather than having multiple outputs, could we have unity gain inputs for each of the colors?
Otherwise, this is starting to look good 
