square to sine convertor

[rohandayal]

I want to convert a square wave to a sine wave of same amplitude and frequency.Anybody help me out

[audioguru]

Hi Ro,
Welcome to our forum.
To convert a square wave to a sine wave, just filter its harmonics with a low-pass-filter. It must have a steep slope to pass the fundamental but attenuate the 3rd harmonic and above. 50-50 square waves don't have even-numbered harmonics. Use at least a 4-pole Butterworth LPF.

The web is full of and I have made converters using a switched-capacitor LPF IC. My extremely low distortion sine-wave generator uses a square-wave with over-sampling and a switched-cap LPF IC. The amplitude of its sine-wave is rock solid.

[rohandayal]

could any one tell anyone tell how to reduce harmonics and noise signals in a sinusoidal wave using a PLL4046 chip?

[audioguru]

Hi Ro,
A PLL has a square-wave output so you would lose any amplitude changes of a sine-wave it is locked in frequency with. A PLL is useful to recover a sine-wave that is buried in noise or has frequency jitter.

A PLL output might have frequency overshoot and decaying frequency hunting with quick frequency changes of the sine-wave, or follow the frequency changes too slowly.

Then you just filter-out the PLL's square-wave's harmonics with a low-pass-filter like previously to end up with a pure sine-wave. ;D

[Kevin Weddle]

If you could modify the square wave beforehand you would be better off. It requires that you use logic to insert a DC level of zero volts for a time at the zero crossing. There is a reason for converting a square wave to a sine wave and the applications are limited. It would seem very appropriate to construct a sine wave from an oscillator. I would like someone to tell me how it is done. I mean how do they actually get the value of the sine wave right. As I understand it, the output of an oscillator is basically a single rate of change. Sort of like a triangle. How do they accomplish the effect that gives rise to the rounding at the peak.

[gg4rest]

Here is a good article I found on constructing sinewave oscillators.

http://www.web-ee.com/primers/files/AN-263.pdf

Hope it helps.

[audioguru]

Thanks GG, I forgot about that appl note. Here's a modern one:
http://www-s.ti.com/sc/psheets/sloa060/sloa060.pdf

[Guest Alun]

What frequency?

Try this circuit the formula for the frequency is 1.38/(RC) Just make sure the capacitors and resistors drawn in pink and labeled R and C are the same value.

The 555 generates a square and sort of triangle wave the first op-amp amplifies this, the R and C on the output low pass filters the triangle wave to give a sine wave, and the last amplifies this again. There are simpler ways to do this but I desined this circuit a long time ago, I used a dual sterio pot for R and a gang rotary switch to switch the value of C. I recomend you use 10

[gg4rest]

Thanks for the appl note audioguru. That one is very good.

[Guest Alun]

Great app-note Audioguru!

[audioguru]

Thanks Guys,
When I first saw that new appl note, I built a buffered phase-shift osc and fixed its problems with a 4th opamp. I used 3/4th of a 4-gang pot I got out of a toy. It works great. Using a little low-droppout regulator, its amplitude is rock-solid and its distortion is very low. ;D

[prateeksikka]

HEY GUYS!
WHATS THE PROBLEM IN USING AN INTEGRATOR CIRCUIT WITH FEEBLE AMPLIFICATION USING OPAMPS?
PRATEEK

[audioguru]

Hi Prateek,
A single integrator won't convert a square-wave to a sine-wave. Many integrator stages will attenuate the fundamental frequency as well as its harmonics. That's why I recommended using a multi-pole Butterworth low-pass-filter, because of its sharp cutoff.

Opamps don't have feeble amplification. They have an open-loop gain of hundreds of thousands and with negative feedback have almost any amplification you need.