Active matrix mixer... Again!?

[Nicotep]

Trying to build me an active matrix mixer. Read all I could about it, including in this forum. Learned much about impedance but it looks like not enough.
Pladask makes a great 4x4 one but I think eh, with all these opamps sleeping here, I'd better make it myself.
Then come the troubles : how come I don't even succeed to make a basic active mixer? Tried different schematics and none of them worked on my breadboard...
Tried the ggg mini-mixer. Was noisy, and much worse if I put buffers before the inputs.
Tried Rod Elliott's, tried Ken Stone's. Succeeded to produce either silence or distorted sound.
I can't understand why it's so difficult, with so few components.
Would someone point me at a nice, simple method ?
Sure, I'd like to have it able to handle hi-z as well as line levels, but if I could only have a starting point which sounds good with low impedance, then complexify later...
Grrrrrrr...

[Sooner Boomer]

What are you using to do the actual switching? How is that controlled?

For impedance control, you're pretty much going to either need an op amp, or some sort of FET buffer.

Have you tried tracing the signal through a circuit to see exactly where the distortion occurs? And then break the connection to the following curcuit to see which side the problem lies?

[Nicotep]

What are you using to do the actual switching? How is that controlled?

I plan to use only pots for sending to the outputs (well, later i'd love to modulate them with LFOs...but that's for MUCH later).
Rod Elliott's design https://sound-au.com/project129.htm or Ken Stones http://synthpanel.com/modules/cgs33_matrix_mixer.html are the type I want to achieve.

For impedance control, you're pretty much going to either need an op amp, or some sort of FET buffer.

Sure. Though I'm not 100% confident with the sole concept of "buffering"...

Have you tried tracing the signal through a circuit to see exactly where the distortion occurs? And then break the connection to the following curcuit to see which side the problem lies?

Nope. And I don't have an oscilloscope...

[ElectricDruid]

I plan to use only pots for sending to the outputs (well, later i'd love to modulate them with LFOs...but that's for MUCH later).
Rod Elliott's design https://sound-au.com/project129.htm or Ken Stones http://synthpanel.com/modules/cgs33_matrix_mixer.html are the type I want to achieve.

I think the problems you're having might be down to the power supply. You don't mention what kind of a supply you're using, but both of those circuits use a nice bipolar +/-15V supply. The classic audio circuit supply, in fact. We mostly don't have such luxuries for stomp boxes, but such a supply makes working with op-amps a lot easier. None of that awkward messing around with virtual grounds and mid-point biasing and so on.

So my advice is to start off with the power supply. If you get that sorted, I think you'll find the other stuff just magically starts to work for you a lot more easily.

[Nicotep]

I think the problems you're having might be down to the power supply. You don't mention what kind of a supply you're using

That's right, and that was exactly the source of my problem. Each time I was testing it with 9V-usb-rechargeable batteries (I know...) and they were producing a strong whine. So I switched to a tired 9V alkaline hanging out and... no noise !
I first tried a basic ggg mini-mixer with virtual ground, then used a charge pump with a 7660s and it worked too though I didn't notice much difference.

both of those circuits use a nice bipolar +/-15V supply. The classic audio circuit supply, in fact.

The reason I didn't try it yet is I'm living in an 'isolated' place (it's good to be a pirate on such strange societal-sickness moment, btw) with mere solar power, which I avoid to convert to 220V. So I don't have many power supplies here.
It's not always convenient, but naively makes me proud of my relative sunpowered successes when they happen.

We mostly don't have such luxuries for stomp boxes, but such a supply makes working with op-amps a lot easier. None of that awkward messing around with virtual grounds and mid-point biasing and so on.

I didn't know they were evil, much too noob for that. Until now I thought that was a simple workaround that does the trick.

So my advice is to start off with the power supply. If you get that sorted, I think you'll find the other stuff just magically starts to work for you a lot more easily.

Thanks for your advice. Now after what you wrote, I'm wondering if a +17/-17 made out of a couple of 7660s/MAX1044 is a good idea or will bring back some noise...?

[ElectricDruid]

both of those circuits use a nice bipolar +/-15V supply. The classic audio circuit supply, in fact.

The reason I didn't try it yet is I'm living in an 'isolated' place (it's good to be a pirate on such strange societal-sickness moment, btw) with mere solar power, which I avoid to convert to 220V. So I don't have many power supplies here.
It's not always convenient, but naively makes me proud of my relative sunpowered successes when they happen.

Well, how about a +/12V supply with a couple of 12V batteries then? That'd make any op-amp happy.

We mostly don't have such luxuries for stomp boxes, but such a supply makes working with op-amps a lot easier. None of that awkward messing around with virtual grounds and mid-point biasing and so on.

I didn't know they were evil, much too noob for that. Until now I thought that was a simple workaround that does the trick.

Not evil exactly, but "workaround" does sort of suggest that they're a problem, doesn't it? Mostly a virtual ground *does* do the trick, but making a circuit work on a single supply is definitely one more thing to think about. Sometimes it's nice not to have to.

Now after what you wrote, I'm wondering if a +17/-17 made out of a couple of 7660s/MAX1044 is a good idea or will bring back some noise...?

There's a load of stuff on this site about bipolar supplies made with one or another power-convertor chip, so I recommend a search.
I would say that those chips depend on a high frequency oscillator (which might or might not be a problem) and also usually have a limited current supply. Too much current drain on the newly-created negative supply pulls the voltage down and you finish up with an asymmetric +17/-14V supply or similar. Not that op-amps care much about that either, but you need to watch the headroom.

[Nicotep]

Hi Tom,

Well, how about a +/12V supply with a couple of 12V batteries then? That'd make any op-amp happy.

Haha you're right... I never used my huge batteries as a direct power supply for those mini-circuits...

Not evil exactly, but "workaround" does sort of suggest that they're a problem, doesn't it? Mostly a virtual ground *does* do the trick, but making a circuit work on a single supply is definitely one more thing to think about. Sometimes it's nice not to have to.

I finally managed to make a rather clean mixer (based on ggg's) which reacts ok with my different attempts, now that it's stripboarded and not on my messy breadboard anymore : 9V, 2x9V is series, and even the chinese usb-rechargeable-9V works (a bit noisier). The 7660s voltage doubler is ok too, but not with a 1044 which whines like a summer cat.
Slowly learning...