Best Place for Volume Pot?

Thread Starter

AndrewCE

Joined Jun 27, 2009
14


I want each channel of this mixer to have its own volume control, and i'm wondering what is the best place for it. Putting it before the input buffers would alter the loading characteristics, and I don't want that. Putting them after the buffers (before the 10K resistors), changing the volume would alter the mix of the channels, which arguably is what I'm trying to do, but I can see that getting complicated. Opamp-gain-altering pots might work, but won't they get noisy? any opinions? What is the best way to have a volume control on these channels?
 

KL7AJ

Joined Nov 4, 2008
2,229


I want each channel of this mixer to have its own volume control, and i'm wondering what is the best place for it. Putting it before the input buffers would alter the loading characteristics, and I don't want that. Putting them after the buffers (before the 10K resistors), changing the volume would alter the mix of the channels, which arguably is what I'm trying to do, but I can see that getting complicated. Opamp-gain-altering pots might work, but won't they get noisy? any opinions? What is the best way to have a volume control on these channels?

Pots always generate some noise, so it's always best to have some amplification before any pots....so the SNR is already established beforehand. Right after the input buffers is generally a good choice.

eric
 

Thread Starter

AndrewCE

Joined Jun 27, 2009
14
Pots always generate some noise, so it's always best to have some amplification before any pots....so the SNR is already established beforehand. Right after the input buffers is generally a good choice.

eric
But then, as I turn the volume pot down for a channel, that 10K resistor becomes effectively larger, changing the mix of the signals (I think the channel whose volume I reduce will actually get MORE space in the mix, so the volume pot would be working against itself).
 

daviddeakin

Joined Aug 6, 2009
207
Firstly you already have 1M resistors in a feedback loop, which is very noisy indeed! Can you adjust the deign to use lower values, like 10k?

Also, adding the pots at the input, before the opamps, will not alter the input impedance as you adjust them (not very much anyway).

Using a pot to vary the gain of an opamp is a very suitable method, provided it's not a high value like 1M! It has the advantage that as you reduce the gain you improve the linearity of the stage.

Alternatively, if you rearrange that mixer opamp as a virtual earth (inverting) mixer then you can put the pots after after each of the first opamps and it won't affect the ratio of mixing when you adjust the pots.
 

Thread Starter

AndrewCE

Joined Jun 27, 2009
14
Firstly you already have 1M resistors in a feedback loop, which is very noisy indeed! Can you adjust the deign to use lower values, like 10k?

Also, adding the pots at the input, before the opamps, will not alter the input impedance as you adjust them (not very much anyway).

Using a pot to vary the gain of an opamp is a very suitable method, provided it's not a high value like 1M! It has the advantage that as you reduce the gain you improve the linearity of the stage.

Alternatively, if you rearrange that mixer opamp as a virtual earth (inverting) mixer then you can put the pots after after each of the first opamps and it won't affect the ratio of mixing when you adjust the pots.
Thanks,

First, if I use a 10K resistor in my feedback loop, I will need to use a 10K for the other resistor to keep the gain at 1. At that point, the input impedance of that buffer is only 10K! Did you overlook this or is there a way around it (is 10K even that bad?)?

If I use variable gain stages, those noninverting buffers cannot get down to a gain of 0, which is a feature that I want.

Are you sure the pots wouldnt affect the ratio of signal mixing? Because I'm pretty sure the gain of that last opamp with respect to any one of the three given inputs is dependant on that 10K value from one opamp to the next. If there is, say, a 50K volume pot, when it is turned down halfway (lets assume linear for the sake of easy math), the series resistance on the line is 25K+10K. So if the feedback resistor is 10K, the gain is 10K/35K= 2/7. See, with the inverting opamp at the end, changing the volume alters the gain (arguably now the change is in the right direction, but still I'd like to not have it for the sake of simplicity and pot linearity). Am I asking too much? Is this the way commercial mixers are actually wired? I havent found any convincing schems.
 

daviddeakin

Joined Aug 6, 2009
207
Thanks,

First, if I use a 10K resistor in my feedback loop, I will need to use a 10K for the other resistor to keep the gain at 1. At that point, the input impedance of that buffer is only 10K! Did you overlook this or is there a way around it (is 10K even that bad?)?
Well you didn't say what you want to plug into this circuit! Do you need a 1M input impedance? Could you not rearrange that third input as a non-inverting amp, so that you can put low values in the feedback loop and still use a 1M input impedance? Do you only have 4 opamps, or could you use more?


If I use variable gain stages, those noninverting buffers cannot get down to a gain of 0, which is a feature that I want.
True. Perhaps a master volume at the output...?

Are you sure the pots wouldnt affect the ratio of signal mixing?
As you have it drawn now they would affect mixing, yes, but if you used a virtual earth mixer instead then the channels won't interact anymore, the mixing would stay the same regardless of vol setting. Each channel would "think" that it is driving a resistance to earth, it wouldn't "see" the other pots/resistors.


EDIT: I just spotted that you are running this from 9V. Is this for electric guitar? I'm sure we can come up with something. You don't even need to worry at all about relative phasing with guitar.
 
Last edited:

Thread Starter

AndrewCE

Joined Jun 27, 2009
14
Well you didn't say what you want to plug into this circuit! Do you need a 1M input impedance? Could you not rearrange that third input as a non-inverting amp, so that you can put low values in the feedback loop and still use a 1M input impedance? Do you only have 4 opamps, or could you use more?


True. Perhaps a master volume at the output...?


As you have it drawn now they would affect mixing, yes, but if you used a virtual earth mixer instead then the channels won't interact anymore, the mixing would stay the same regardless of vol setting. Each channel would "think" that it is driving a resistance to earth, it wouldn't "see" the other pots/resistors.


EDIT: I just spotted that you are running this from 9V. Is this for electric guitar? I'm sure we can come up with something. You don't even need to worry at all about relative phasing with guitar.
That new design you showed is what would seem natural for a mixer, but the one "weird" feature I want in this one is that one of the inputs is inverting. That way I can mix pedals (you guessed right, i am a guitarist) in an out-of-phase way to experiment with sounds.

Also, with your design, it seems that even with the volume control full up, in tandem with the 10K resistor at least half of the volume is thrown away to ground.

See?
 
Last edited:

daviddeakin

Joined Aug 6, 2009
207
That new design you showed is what would seem natural for a mixer, but the one "weird" feature I want in this one is that one of the inputs is inverting. That way I can mix pedals (you guessed right, i am a guitarist) in an out-of-phase way to experiment with sounds.
In that case just feed one of the vol pots into the non-inverting input of the mixer; that will achieve the same thing.


Also, with your design, it seems that even with the volume control full up, in tandem with the 10K resistor at least half of the volume is thrown away to ground.
See?
In order to have attenuation the signal has to "pass through" a series resistance, followed by a resistance to ground (your top diagram). In your second diagram there is a resistance to ground (which does nothing), followed by a series resistance. If the input impedance of the next stage is high, that series resistance does nothing- no attenuation. In the case of the mixer it feeds a virtual earth; current just flows into the virtual earth through a (10k) series resistance, it goes straight to the input of the mixer opamp.
I hope I'm explaining it well enough, but I assure you, the vol pots work perfectly normally with no loss in the max position. Loads of mixers circuits are built that way.
 
Last edited:
Top