Hi,
I'm currently trying to design a square-wave to sine converter. So far this design works. You can see the vactrols are being controlled by high voltage but this is just for the simulation so I can enter the ohm value I want in the voltage source
What I want to achieve is:
- The input signal is a DC square wave
- I need a DC sine as output
- voltages between 0 and 5V
- needs to work for 0-20hz
You can ignore the voltage divider and 1st opamp at the begining. This is so I can attenuate the input signal with a potentiometer.


When building this circuit, I'll control the resistors with a potentiometer instead. The problem I have with this is that I need 4 knobs (or 1 4-gang knob) and now I'm thinking of adding another opamp so I would need to add 2 more knobs. This doesn't feel like it's the right solution. Vactrols would work but they are expensive. Digital pots would work but that complicates things a lot.
I'm thinking that maybe I could pin down the resistor values and vary the capacitors instead. But I'm not too sure if that's possible.

I've seen people replace a capacitor with two capacitors in series with a voltage source in the middle. Something tells me this could work but I'm not sure why. Can anyone tell me if that's the right direction to go? or is my whole design flawed based on what I need to achieve?

Thank you.

 

Can anyone tell me if that's the right direction to go? or is my whole design flawed based on what I need to achieve?

What level of distortion are you trying to achieve?
What is the intended purpose of the optical devices? If you vary both resistors simultaneously the gain stays the same.

I've seen people replace a capacitor with two capacitors in series with a voltage source in the middle.

Are you thinking of varactors?

 

- needs to work for 0-20hz

"0" makes this very hard. Even 20hz makes this hard.
Numerically Controlled Oscillator will work to near "0".

 

"0" makes this very hard. Even 20hz makes this hard.
Numerically Controlled Oscillator will work to near "0".

I'm sure I could generate a 0Hz sinewave from a 0Hz squarewave very easily!
0.0001Hz would be much trickier!

 

0.0001Hz would be much trickier!

Joke...

As you approach 0 the capacitors need to approach infinity. That is why I think NCO is a better approach.
micro chip NCO

 

This is for an audio application. I don't need 0hz... let's say 1 to 20 then.
As for the distortion, I don't know what a typical tolerance should be. I'd like to be as close as possible but given my very limited experience with this kind of things, I can sacrifice quality for simplicity for now.

I'll take a look at varactors. But the most common thing I see in this type of application, is people doing something like this:


I'm not sure that achieves what I need though.

Ideally, I wouldn't have to tune the RC when the input frequency changes, but I guess the only way to do this would still be through controlling the capacitance or resistance through some other mechanism such as a microcontroller.

 

That would be the circuit one would use if C1 and C2 were varactors.
Varactors have capacitance ranges in the pF region, so won't be much use at <20Hz!