Author |
Message |
e1999
Joined: Apr 25, 2019 Posts: 29 Location: naples
|
Posted: Sun Jun 30, 2019 2:41 am Post subject:
? about building stereo active mixer Subject description: edit - i think the cause of the trouble is power related |
 |
|
Hi there,
Hoping that someone here might be able to help me out. I am attempting to build a stereo mixer, ie take a mono input(s) and pan between left and right outputs.
I've used syntherjack's 4sum mixer
https://syntherjack.net/wp-content/uploads/2018/09/4SUM-portable-mixer-schematic-syntherjack.png
and
the info/schematics from allelectrickitchen
http://www.all-electric.com/schematic/simp_mix.htm
to base the design off of.
When building a mono channel everything functions well. But when I duplicate the channel to create stereo, using TL072 op amps it seems that there is a phasing issue and the two channels when played together sound distorted and cancel parts of each other's signal out.
*edit - i think this is actually a power issue as the the distortion and cancellation happens even when the two signals are not both outputting sound - that is it happens as long as both op amps are receiving power. As soon as I cut power to one amp, the other performs as it should. also, trying out different decoupling capacitors altered the interference between the two op amps.
Any suggestions / ideas / help would be very much appreciated! |
|
Back to top
|
|
 |
PHOBoS

Joined: Jan 14, 2010 Posts: 5795 Location: Moon Base
Audio files: 709
|
|
Back to top
|
|
 |
e1999
Joined: Apr 25, 2019 Posts: 29 Location: naples
|
|
Back to top
|
|
 |
PHOBoS

Joined: Jan 14, 2010 Posts: 5795 Location: Moon Base
Audio files: 709
|
Posted: Mon Jul 01, 2019 6:56 am Post subject:
|
 |
|
good to hear it works now
Quote: | I've read that adding higher value resistors introduces noise into a mixer design. right now I'm using 22k with a 220k on the 1st inverter of the op amp to add quite a lot of gain if needed. Would there be any advantage or disadvantage to using for example 10k and 100k? |
I am not that knowledgeable about opamps so someone else might be able to answer this better. You are correct that using higher values makes it more
susceptible to noise but it also depends on the type of opamp you use. It might affect the slew rate (?) so could depend on the frequencies you use it for.
I don't think you'd notice any difference using 10K/100K instead of 22K/220K though besides a higher current draw (which will still be very low).
Quote: | I wanted originally to use a 12v dc supply, but using the supply without voltage regulator added too much electrical noise into the circuit. So I used 78l06 to produce 6v. Do you think it's worth it for the extra headroom to use instead a voltage regulator rated for 12v? and if so could I still use 12v power supply or would I need to use something larger like 15 or 18v? |
Depends on the level of the signals (and the opamps) but 6V doesn't leave a lot of headroom. With an ideal opamp it could handle -/+3V signals max. but
unless you use a rail to rail opamp (and even then you should not push it to the max) it might only be half of that. There are also some issues with non-linearity.
On 12V you could use a 9V or even 10V regulator instead which will give you a bit more headroom. But unless you use a low-drop regulator a 12V one won't
do much. Better would be to use an AC supply that you can turn into a bipolar DC supply which will give you twice the voltage and an actual GND.
Quote: | finally I've been reading about virtual earth. but still feel like my understanding about the advantages and disadvantages of various methods is lacking. here I used voltage divider with (2) 22k resistors and a 1uf capacitor. can you say anything about possible issues with this solution?
or how it compares to for example the solution I saw in mutable instruments filter designs where they use an LT1054 into 7805 and 7905 voltage regulators? |
It depends on what you use it for although a real GND is preferable. The problem with a virtual/fake one is that it limits the current and as a result anything
you connect to it will more or less affect its voltage level. Opamps generally have a fairly high input impedance so won't be much of a problem. The pan pots
could cause little fluctuations which you might notice as signal bleed through but it's probaly neglectable. It can especially become problematic when you start
with a simple design and keep expanding on it. Connecting several circuits together all with their own fake GND can work if you add capacitors between them
to block any DC voltage. That's also done in this mixer and you will find it on in/outputs of a lot of audio circuits, especially when they use a single supply.
The problem with this is that capacitors aren't a magic solution and only work for a limitted frequency range so will always affect the signal level to some
degree. Also fluctuations in the supply voltage will affect a fake GND as it is derived fromt that. Using an opamp configured as a voltage follower to buffer the
fake ground will already improve it a bit. I can't really say anything about the LT1054 as I don't have any experience with it (besides soldering some MI
modules).
Quote: | also in that design they use something like 16 bypass capacitors, when typically I see people using just 2 or so, do you know what might be the reason for that? is this because it is receiving signals from a cpu? |
It's generally a good practise to provide each chip which it's own bypass capacitors to filter out any noise from and to the supply. Often circuits can work fine
without them but unless you really have to cut costs it's better and easier to just put them in. Also sometimes capacitors with different values are used in
parallel to cover a wider range of frequencies. _________________ "My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube |
|
Back to top
|
|
 |
e1999
Joined: Apr 25, 2019 Posts: 29 Location: naples
|
Posted: Mon Jul 01, 2019 4:48 pm Post subject:
|
 |
|
ok thank you
lots of v useful info here
i will try the op amp configured as a voltage follower
would you use more than one of these op amp voltage followers if you have a circuit that is multiplying in complexity with many ic's / parts in the design?
also in such a design with many parts (on a single supply), would it be in any way helpful to make an op amp voltage follower for the +V power as well? |
|
Back to top
|
|
 |
Grumble

Joined: Nov 23, 2015 Posts: 1310 Location: Netherlands
Audio files: 30
|
Posted: Tue Jul 02, 2019 2:43 am Post subject:
|
 |
|
You could also try this: http://www.falstad.com/circuit/circuitjs.html
Here you can enter components and see immediately what happens when you change a resistor value etc.
And as a side-note: You have a 6volt regulator, why don't you use that as the virtual ground and power the opamps with +12 volt and GND? If you are worried about the noise, you should add some capacitors close to the power connections of the opamps (10uF parallel with 100NF), and maybe a 10uH coil in series in the +12 volt power line. _________________ my synth |
|
Back to top
|
|
 |
PHOBoS

Joined: Jan 14, 2010 Posts: 5795 Location: Moon Base
Audio files: 709
|
Posted: Tue Jul 02, 2019 12:56 pm Post subject:
|
 |
|
Quote: | would you use more than one of these op amp voltage followers if you have a circuit that is multiplying in complexity with many ic's / parts in the design? |
hmm, I never really thought about doing that but I guess it would be an improvement over using a seperate 'GND' for every circuit.
There'd still be small differences between them as opamps aren't perfect but for certain applications it could be an adequate solution.
Personally I'd go for a bipolar supply if possible when a circuit reaches a certain complexity. But for a simple circuit or one that runs
mostly on a single supply a fake GND can be very useful. Take your mixer for example, if you keep it small a simple it will work fine
but if you'd want to add some tone control, more channels, maybe some other options you will sooner or later run into problems.
Quote: | also in such a design with many parts (on a single supply), would it be in any way helpful to make an op amp voltage follower for the +V power as well? |
That wouldn't have any use. For a fake GND created with a voltage divider it is an improvement because the output will not fluctuate
much with varying currents. If you just use the resistors then any variation in current you draw from it will cause its level to change
(it's just ohm's law). The V+ line is already more stable and capable of delivering a lot more current without fluctuations than an opamp
would. There's also the problem that you'd need an even higher supply voltage to power the opamp
Using a voltage regulator as grumble suggested is also a good alternative. It has its own internal voltage reference so doesn't depend on
the supply voltage, and can handle more current than a standard opamp. Instead of a coil you can also add a small value resistor (10 ohms)
in series, or use both. I too have had problems with unregulated supplies in the past and adding capacitors doesn't always solve it. _________________ "My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube |
|
Back to top
|
|
 |
JovianPyx

Joined: Nov 20, 2007 Posts: 1988 Location: West Red Spot, Jupiter
Audio files: 224
|
Posted: Tue Jul 02, 2019 3:34 pm Post subject:
|
 |
|
The use of virtual grounds is a design choice that should be made with care and complete information. There are a few scenarios where they work well, but many more where they don't. Small designs that use low currents are most likely to tolerate virtual grounds. When you are designing something, it is always a good idea to start by using a good quality bench power supply that is already dual voltage and more than beefy enough for the current needs.
Virtual grounds are likely to cause problems in circuits where the design is not solidified and is growing. For these, the bench supply will not introduce new problems related to virtual grounds so that the development works properly (functionally as well as troubleshooting) until the idea has been completed. Once that is finished, if there is truly a need for virtual ground, then the bench supply can be replaced with a single supply and virtual ground and then regression tested. However, as the design grows it becomes less and less friendly to the virtual ground. For battery operated designs, consider making a dual supply with two 9 volt batteries as opposed to one 9v battery and a virtual ground. The +9/-9 supply will give more headroom comfort and will have a much more stable ground than the virtual ground method. If you begin a design with a virtual ground, once the current needs grow beyond the capacity of a virtual ground you will experience problems, but they are weird problems that will make you chase your tail mucking with things that were actually quite well designed. Failure of a virtual ground can show up in the oddest places and will flumox the designer. This is what happened to Cynosure with the CMOnSter project (somewhere here on EM). It was fine until a certain point where adding things just wouldn't work. Once he replaced the virtual ground supply with a real dual supply, it all worked as designed.
As PHOBoS pointed out, the problem with virtual ground is current. While it may be used as a voltage reference, there is still current draw. Virtual grounds can be made more stable (as PHOBoS points out) with an opamp voltage following buffer, but only to a limited degree. Most opamps are comfortable sourcing or sinking around 10 mA or so. If more than that is needed, the voltage of the virtual ground can shift. That injects noise and that noise will go everywhere the virtual ground is used.
In my own design world, I would never want to use a virtual ground for any mixer, but that's just me. _________________ FPGA, dsPIC and Fatman Synth Stuff
Time flies like a banana. Fruit flies when you're having fun. BTW, Do these genes make my ass look fat? corruptio optimi pessima
|
|
Back to top
|
|
 |
AlanP
Joined: Mar 11, 2014 Posts: 746 Location: New Zealand
Audio files: 41
|
Posted: Tue Jul 02, 2019 11:41 pm Post subject:
|
 |
|
Not to be a downer, but is there any reason you need to use single rail supply?
If the circuit as a whole is relatively low current, you can use chips like the MAX1044 (or better) to supply low current circuits on the negative rail, respecting the limitations of the chip. |
|
Back to top
|
|
 |
e1999
Joined: Apr 25, 2019 Posts: 29 Location: naples
|
Posted: Wed Jul 03, 2019 5:43 am Post subject:
|
 |
|
Thanks to everyone for all the great info & advise!
I am going to try all of the suggestions:)
Just as an experiment, I tried voltage follower in several parts of the circuit for the + power rail and it got rid of all the noise that the caps to ground hadn't been able to handle although it reduced the + rail to something like 11.77 instead of 12v. Still better than the 6v from the 6v voltage regulator I happened to have. Anyways, I don't think I'll use that as a final solution-at least for the + rail, but good to know about regardless.
Also I get what you all are saying about dual supply vs virtual ground. And it totally makes sense that using dual supply is objectively better, especially for sound applications.
However, I'm new to this and I've read a lot of warnings about jumping straight into building a dual power supply, without the requisite knowledge and understanding. As well there are a few practical/logistical reasons (ie travel a lot and dc supply is smaller, lighter, easier to replace if broken, no step up/down converter etc...) that make a DC supply(if possible to use) in some ways preferable for the current project I'm working on.
As a side note, I've noticed that there are quite a few commercial synths now being made that run off DC supply, such as Moog Mother 32 or Dfam for instance. At the same time I've heard that traditional AC dual supply is still more reliable for modular synths and mixers.
Does anybody have any insight or experience on dual supply created from single supplies for synths?
Is a dual supply that is created from a switching psu such as doepfer psu-3 also using a virtual ground? |
|
Back to top
|
|
 |
PHOBoS

Joined: Jan 14, 2010 Posts: 5795 Location: Moon Base
Audio files: 709
|
Posted: Wed Jul 03, 2019 1:57 pm Post subject:
|
 |
|
Quote: | Just as an experiment, I tried voltage follower in several parts of the circuit for the + power rail and it got rid of all the noise that the caps to ground hadn't been able to handle although it reduced the + rail to something like 11.77 instead of 12v. |
Most interesting, if anything I would have expected it to make it worse. I guess it might be limiting the current and having a
similar effect as placing a resistor in series.
Here's the schematic for a single supply mixer I made couple years ago. Don't pay to much attention to the component values
but as you can see I placed capacitors after the panning pots instead of the level pots. Because of this that section doesn't need
a fake ground and I only used it as a reference voltage for the opamps.
Quote: | Does anybody have any insight or experience on dual supply created from single supplies for synths? |
I don't really have any info on that but they might use something similar to the chip AlanP posted which means it keeps the actual ground
and adds a negative supply. (btw you can also make a dual supply with 2 seperate DC supplies) _________________ "My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube |
|
Back to top
|
|
 |
e1999
Joined: Apr 25, 2019 Posts: 29 Location: naples
|
Posted: Thu Jul 04, 2019 8:37 am Post subject:
|
 |
|
Hey thats smart.
Thanks for sharing your schematic!
Have to put this project on hold for the next week bc other things, but will try this out when I resume! |
|
Back to top
|
|
 |
|