Author Topic: Mixing Synth Audio  (Read 2326 times)

0 Members and 1 Guest are viewing this topic.

Offline FlyingHackerTopic starter

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: us
  • You're Doing it Wrong
Mixing Synth Audio
« on: August 18, 2017, 09:11:47 pm »
I am working in a Synth project with three CD4046 VCOs used for harmony. To mix the output of those three VCOs should I use coupling capacitors and then 47K resistors as a passive mixer? Or resistors first and then a single series cap to remove the D.C. bias? I am guessing the output from the 4046 should be low enough impedance that a passive mixer is fine, rather than a summing OpAmp.

Open to any thoughts on this. Thanks.
--73
 

Online floobydust

  • Super Contributor
  • ***
  • Posts: 6928
  • Country: ca
Re: Mixing Synth Audio
« Reply #1 on: August 19, 2017, 01:59:40 am »
You could use a simple resistor summation, the disadvantage is output impedance of the sum (all three) is higher.
i.e. you use three of 10k resistor, so your output drive is basically 10k/3 or 3.33k which may not be mediocre for the audio gear downstream.

There is no DC component or bias on the VCO output - it's just a square wave? No need for a capacitor then.

The CD4046 VCO output is a bit weak, is does not source much current. Running off 5V with 1mA load it drops to 4.6V, 2.5V at 3.2mA.



 

Online DaJMasta

  • Super Contributor
  • ***
  • Posts: 2289
  • Country: us
    • medpants.com
Re: Mixing Synth Audio
« Reply #2 on: August 19, 2017, 03:13:01 am »
If you wanted the square wave aligned for a split supply amp, then you could use a cap, though you can just sum normally then put your capacitor.

For the aforementioned reason, I'd just use a summing amp instead of a passive sum just because you maintain your capacity to drive some load and assure that the impedance the CD4046s are driving.  Simplifies plugging whatever is needed down the chain, and in the event you need some gain, you can do that easily.
 

Offline FlyingHackerTopic starter

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: us
  • You're Doing it Wrong
Re: Mixing Synth Audio
« Reply #3 on: August 19, 2017, 03:20:49 am »
Thanks for the replies.

My VCO outputs a positive going square wave. So it does have an average positive D.C. Voltage.

Ok, a summing amp sounds like a better solution then. I tried it with 47K resistors in a passive mix, with a single cap afterwards before feeding into the audio amplifier, and it did work out fairly well. I think I want to buffer the output anyway. So summing would do that at the same time.
--73
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 19347
  • Country: gb
  • 0999
Re: Mixing Synth Audio
« Reply #4 on: August 19, 2017, 01:09:48 pm »
You could use a simple resistor summation, the disadvantage is output impedance of the sum (all three) is higher.
i.e. you use three of 10k resistor, so your output drive is basically 10k/3 or 3.33k which may not be mediocre for the audio gear downstream.
Actually the impedance is not the sum of all outputs, but the reciprocal. The admittance is equal to the sum of all output admittances.

Quote
There is no DC component or bias on the VCO output - it's just a square wave? No need for a capacitor then.
There is a DC component, unless the VSS pin on the IC is equal to the negative of VCC pin. Assuming 0V on VSS, the DC component on the output is equal to half of VDD.

Quote
The CD4046 VCO output is a bit weak, is does not source much current. Running off 5V with 1mA load it drops to 4.6V, 2.5V at 3.2mA.
Yes, use high value resistors or a buffer.
 

Offline FlyingHackerTopic starter

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: us
  • You're Doing it Wrong
Re: Mixing Synth Audio
« Reply #5 on: August 19, 2017, 01:40:20 pm »
Thanks for the reply, Hero999.

Yes, there is a D.C. Component because my Vss is indeed 0V.

Interestingly, the 4046 seems to put out plenty of level for my amp (an old Peavy guitar amp). I had to put a pad in there to get the level reasonable for the amp. I know this is the voltage not the current, though, and the pad is just sucking away for current too.

I will do a summation amp with something like an NE5532. I will keep the signals positive until the final output where I will put a blocking cap.
--73
 

Online DaJMasta

  • Super Contributor
  • ***
  • Posts: 2289
  • Country: us
    • medpants.com
Re: Mixing Synth Audio
« Reply #6 on: August 19, 2017, 02:54:05 pm »
An NE5532 is a good choice for audio stuff, but since it won't work off a single supply, it may simplify the design a bit to go with a single supply capable amp.

There are good, fairly inexpensive audio opamps that do work off a single supply, though.  All you'd need is a capacitor on the output (of enough size that it can drive the load without losing too much low frequency) and you could avoid needing a negative supply rail for the device - though to keep noise out you may consider local regulation for the amp or something like an inductor before some bulk bypass.
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 19347
  • Country: gb
  • 0999
Re: Mixing Synth Audio
« Reply #7 on: August 19, 2017, 04:28:02 pm »
An NE5532 is a good choice for audio stuff, but since it won't work off a single supply, it may simplify the design a bit to go with a single supply capable amp.
Sorry, that's nonsense. The NE5532 is perfectly capable of operating of a single supply. The only issue which may be a show stopper in some single supply applications, is the common mode range doesn't include the negative rail. Keep both inputs biased around half the supply voltage and the NE5532 will work perfectly from a single supply.

Thanks for the reply, Hero999.

Yes, there is a D.C. Component because my Vss is indeed 0V.

Interestingly, the 4046 seems to put out plenty of level for my amp (an old Peavy guitar amp). I had to put a pad in there to get the level reasonable for the amp. I know this is the voltage not the current, though, and the pad is just sucking away for current too.

I will do a summation amp with something like an NE5532. I will keep the signals positive until the final output where I will put a blocking cap.
What's the supply voltage? The only issue you may have is the NE5532 is not specified to work below a total supply voltage of 10V. You may find it works at lower voltages in practise but it could be unreliable.
http://www.ti.com/lit/ds/symlink/sa5532.pdf

EDIT:
I've just found another datasheet from ON Semi which specifies the NE5532 down to 6V.
https://www.onsemi.com/pub/Collateral/NE5532-D.PDF

The Philips and the original Signetics data sheets also say 6V, so you can be pretty confident down to that voltage. Lower voltages will be marginal and would need testing.
http://www.ben.cz/_d/datasheet/ne5532.pdf
https://4donline.ihs.com/images/VipMasterIC/IC/PHGL/PHGLD076/PHGLD076-4-100.pdf?hkey=EF798316E3902B6ED9A73243A3159BB0
« Last Edit: August 19, 2017, 04:37:44 pm by Hero999 »
 

Online floobydust

  • Super Contributor
  • ***
  • Posts: 6928
  • Country: ca
Re: Mixing Synth Audio
« Reply #8 on: August 19, 2017, 05:31:26 pm »
You could use a simple resistor summation, the disadvantage is output impedance of the sum (all three) is higher.
i.e. you use three of 10k resistor, so your output drive is basically 10k/3 or 3.33k which may not be mediocre for the audio gear downstream.
Actually the impedance is not the sum of all outputs, but the reciprocal. The admittance is equal to the sum of all output admittances.

I'm not sure what you are saying is the output impedance.
Three logic gate (push-pull) outputs tied together through 10k resistors, I apply Thévenin's theorem shorting voltage sources, the impedance seen looking back is 10k/3=3.33k  (ignoring the gates' ~780R non-symmetric output impedance for simplicity's sake).

Just trying to make sure the technique is reasonable, as that is what I use.
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 19347
  • Country: gb
  • 0999
Re: Mixing Synth Audio
« Reply #9 on: August 19, 2017, 05:36:48 pm »
You could use a simple resistor summation, the disadvantage is output impedance of the sum (all three) is higher.
i.e. you use three of 10k resistor, so your output drive is basically 10k/3 or 3.33k which may not be mediocre for the audio gear downstream.
Actually the impedance is not the sum of all outputs, but the reciprocal. The admittance is equal to the sum of all output admittances.

I'm not sure what you are saying is the output impedance.
Three logic gate (push-pull) outputs tied together through 10k resistors, I apply Thévenin's theorem shorting voltage sources, the impedance seen looking back is 10k/3=3.33k  (ignoring the gates' ~780R non-symmetric output impedance for simplicity's sake).

Just trying to make sure the technique is reasonable, as that is what I use.
Yes, that's true but you said:

You could use a simple resistor summation, the disadvantage is output impedance of the sum (all three) is higher.

You meant lower.
 

Online floobydust

  • Super Contributor
  • ***
  • Posts: 6928
  • Country: ca
Re: Mixing Synth Audio
« Reply #10 on: August 19, 2017, 05:50:23 pm »
You could use a simple resistor summation, the disadvantage is output impedance of the sum (all three) is higher.
i.e. you use three of 10k resistor, so your output drive is basically 10k/3 or 3.33k which may not be mediocre for the audio gear downstream.
Actually the impedance is not the sum of all outputs, but the reciprocal. The admittance is equal to the sum of all output admittances.

I'm not sure what you are saying is the output impedance.
Three logic gate (push-pull) outputs tied together through 10k resistors, I apply Thévenin's theorem shorting voltage sources, the impedance seen looking back is 10k/3=3.33k  (ignoring the gates' ~780R non-symmetric output impedance for simplicity's sake).

Just trying to make sure the technique is reasonable, as that is what I use.
Yes, that's true but you said:

You could use a simple resistor summation, the disadvantage is output impedance of the sum (all three) is higher.

You meant lower.

Ahh I see, I did mean higher compared to taking it directly off one VCO's output, one gate is lower Z. But my coffee is not finished.
I imagined OP started with one VCO and added two more. But not knowing what is downstream- just a raw audio output or going to VCA/VCF etc. a resistor summer could work great or upset things further down the chain.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf