Increase duty cycle on Sallen-Keys filter

[[email protected]]

I am using a LM324 op amp for this filter. R1 and R2 are 40k. R3 and R4 are 3k

**broken link removed**

I am inputing a sine-wave and getting an output riding on the DC level with the negative alternation dipping below the applied voltage. I assume this is how this filter is suppose to work, and have used an inverting op amp to get the alternation to ride on 0V and rise above 0V.

This has been working well, but the duty cycle is quite low on the output of the filter. How can I get a decent duty cycle out of this filter?

 

[crutschow]

You should not get an output riding on a DC level. For a sine-wave the output should just be an undistorted version of the input waveform, either larger or smaller depending upon the frequency, the filter roll-off point, and the filter gain.

What gain are your using (R3 and R4)? What are your supply voltages to the 741?

I don't understand your question about duty-cycle. What is duty-cycle as applied to a filter?

 

[audioguru]

Since the opamp has a dual-polarity supply then the input DC reference voltage from the source to R1 should be 0V. Then the output will also have a DC voltage of 0V with the signal swinging positive and negative.

A 741 opamp works poorly above 9kHz and an LM324 operates poorly above 2kHz and has crossover distortion.

 

[[email protected]]

Hmmm, I turned on the circuit again today and the filter was outputing a near perfect sine-wave. Strange

 

[MrAl]

Hi,

Did you switch to a different input frequency? What are the values of your caps? What frequency are you driving it with?

With R1=R2, R3=R4, and C1=C2 the following holds:

the 3db down point is at:
f=sqrt(sqrt(5)+1)/(2*sqrt(2)*pi*R1*C1)

and the 0db point is at:
f=sqrt(sqrt(16*Vin^2-3)+1)/(2*sqrt(2)*pi*R1*C1)

and it looks like the amplitude reaches a max at the point:
f=1/(2*sqrt(2)*pi*R1*C1)

where the response is:
VoPeak=4*Vin/sqrt(3)

or that in db:
VoPeakdb=20*log(VoPeak)

and the response at DC is:
2*Vin

therefore the peaking factor is:
2/sqrt(3)

which means the output peaks at about 15 percent higher than nominal before it dips down toward the -3db point.
The -3db point is the point where the response is 3db down from nominal.


But this assumes that you are not exceeding the limitations of the 741 in any way.
You'll have to tell us a little more about your circuit if you want more information about it.