Problems with Voice Systems

Could explain in a little more detail? Do you mean that you have a voice command system for your factory?

Or do you mean that you have a voice generator that says something like "Warning! Machine Guard Disabled!"?

I can interpret what you said in several different ways and be right only once. Give us some more information and maybe we can help you.

Thanks a lot for the discussion, but I guess the discussions are going a bit offtrack. Just a bit of brief info.

This system is a voice based fool proffing system wherein two push buttons are provided to the operator on his workstation.When he presses the first push button there is a control circuit which has a pre-recorded voice in OTP (One time programmable) chip in the PCB with adequate noise filtering capacitors.After he hears the voice he presses the second push button which releases a stopper which is pnematically connected so that the operator can pass on the product to the next station.

Here sometimes the voice from system doesn't come.The probability is 0.7 Ok.What might be the probable reasons?

Well now, see how a good description can make things a little more clearer!

You can still do better, however. I don't know how your circuits are connected (you didn't provide that bit of info).

Does the 1st button trigger something else other than the voice generator? If the other thing is triggered but not the voice gen, then there might be something wrong with the connections. It can be a pain to search for, speaking from experience.

If the button only triggers the voice gen, is the signal reaching the voice gen's input? You mentioned in your original post that you've tested it and it works fine on the bench but not in the shop floor. If the connection is there, it might be that you have a loose (read, intermittent) connection. This can be very frustrating to search for. Much easier to find a complete disconnection than an intermittent one.

Another possibility is that your signal cable is too small. Small wires have more resistance and can drop the input voltage if the cable is too long. In that case, use a thicker cable.

If you're using 5 volts to send the signal to the board that is 200ft or more away, you're voltage might be too small. You can use 12volts or 24volts for the trigger and just reduce it at the board's input.

Next time, give us some more information at the beginning so that we can give proper answers. Your original post was too short and lacking in detail. Your second post was better but still lacking.

Come back and tell us how things went.

Oh, and welcome to CR4!

Hendrik14 makes some very good points.

I would start looking at the microphone and immediate-related, as you likely tested it in a lab environment. If the mic. or associated circuitry (amplifier?) has an AGC (aka AVC) (automatic gain control) and given that the floor may be noisy (even a motor whine in the background may be the source), the audio detection may be maxed out. If it's an external amplifier (from the mic) look at changing its changing its higher frequency component (rolloff response). That not being the case, then address the microphone itself and the way you are using it.

It would be helpful to know just what is intended with the application.