zero crossing ac noise

hello
i´m doing a zero crossing detector for ac line....both 110 and 220
i´m using diodes to rectifie the line and using a 44k resistor in series with the optocoupler.....however sometimes there is some "false" zeros.... due to the noise;...how can i eliminate or reduce this noise?

You can follow the optocoupler with a retriggerable one-shot. Configure the one-shot to produce a single output pulse that lasts longer than the noise.

Another method is to pass your optocoupler output to a microcontroller that is running a debounce algorithm. The microcontroller would then output the debounced signal.

Hi Veioloko,

A one shot won't cure the problem because "the noise" you are referring
to is no noise at all. Electricity companies superimpose command signals
on the line voltage for all kind of purposes, like on and off switching of
street lights, switching of electricity meters from day to night regime and
so on,... So you have to find out which frequency they are using and
make a tuned circuit to suppress that frequency and to minimise the
effect on the circuit. Bare in mind that your circuit already has a delay
of 250 µS to the actual zero-crossing of the line voltage.

on1aag.

As on1aag pointed out, there may be communication signals superimposed on your power. The amplitude of such communication signals, however, is well below that of the primary power signal amplitude. As a result, you CAN use a one-shot to completely eliminate this type of noise problem! Simply set the time duration of the one-shot so that its output pulse will persist until the power amplitude, plus the worst-case communication signal amplitude, can't cross back over zero the "wrong" way.

If, on the other hand, your problem is due to "dirty" power, a one shot may not be able to completely eliminate the effect of the noise on zero crossing detection. This can happen if the power is either noisy at the source or if there are significant transients that are caused by switched loads, etc.

Your best bet might be to use either a phase lock loop, or a microcontroller that employs a phase locking algorithm. This will compensate for almost any noise that you might encounter, and it will reduce the latency of your zero crossing detection to a minimum.

What opto are you using?

dfullmer

hello
thanks everybody

i´m using 6n139 .....because of the current....so i can detect with low currents....

I´m sure that i have dirty ac signal......because the oscilation on the output of opto only ocorrurs when someone in my company turn on the inductive machine....

I will try to "correct" this on the microcontroler.... if the time is upper than 8,3 ms so it´s incorrect......