Noise in an AC Circuit

You might try installing a capacitor...

http://forum.allaboutcircuits.com/showthread.php?t=49040

electrolytic capacitors

A 5.20 Volts high frequency zener Diode in series with a low value resistance across signal can clamp down noise,

Rachael,

An LC filter should reduce noise. You may be looking at something oscillating rather than noise.

What is the frequency and amplitude of the square wave, and what is the frequency and amplitude of the noise?

Dave

Bypass Caps. are used for reducing noise.

Here is a filter to get rid of the DC component. Use the D74H4B:

http://www.freqdev.com/products/filters/d74.html

Click on this one to see the Freq response.

Is the 45~50 noise or noise is in this range?

DC Current ? Is the output signal Offset some voltage level ? (+) or (-) not symmetrical to zero level?

DC isolation of Sq wave cannot be attained at lower freqs without distortion [deformed waveshape] even at higher freqs.

You did not write whether

1. You need to extract the signal (45 - 50 Hz??) amplitude/frequency/waveform alone.
2. The current value or voltage value is required.
   
3. Or the 3 to 5 V @ 200 mA must drive a load of impedance x + jy?

You write LC filters reduce current excessively, suggesting you need to drive a certain load for which 4 V and 200 mA suggests 4/0.2 = 20 ohms.

Your description sounds something like extracting the AC loudspeaker sound signal from a npn-npn or npn-pnp transistor "totem pole" output, which has a standing DC voltage at zero signal. Since this involves a fluctuating AC signal of 20 - 20000 Hz, closely observed by "high fidelity" enthusiasts, its low signal distortion is well proven.

For this the standard solution [for about 50 years] has been an aluminium electrolytic capacitor [of negligible impedance compared to the load of about 15 ohms] in series with the load, positive terminal to the positive terminal of the source.

3300 microfarads is about 1 ohm at 50 Hz (its voltage rating should equal the DC supply voltage of the 45-50 Hz). You need one of physical size adequate to have an internal loss resistance of much less than 1 ohm.

If you need to have further filtering, the removal of the DC with this capacitor will make that job easier.

It would help a lot to know what your gear is doing and what "noise" frequencies you need to remove - high. low or band-pass filter?

Is the "noise" generated internally or externally? Preventing the signal from ever entering your circuit would be a good idea.

Please define "noise". According to Fourier, a true square signal is itself an infinite sum of signals, multipes of the base frequency (harmonics) that are not noise. They are signal components. S.M.

Thank you this was said earlier and on more testing I have found that I had removed the noise and what left was harmonics in the 3 and 5 cycles this then fed into the microcontroller ATMEL type ATMMEG48 the noise was created by the circuit and the harmonics was external we have fited a Ferite bead on the signal input line and although it hasn't been removed it has reduced it to a degree that it doesn't effect the controler. We reduced the voltage on the zenner diode and the combined has imroved the opperation and stopped the controller resetting.

Rachael

The $1 solution to the $100,000 worth of engineering time spent on it.

And then your boss says: Well! Anybody could've done that!

If you had written initially "the noise is causing spurious reset of a microcontroller", this would have shouted "high amplitude noise spikes" and got the cure much quicker!

I, for one, thought you had a "signal extraction from pulse waveform" problem.

67model

Thank you all for the help and suggestions finally got the circuit working properly

WOULD NOT HAVE SOLVED THE PROBLEM WITHOUT ALL YOU GOOD PEOPLES HELP

Rachael