Jacky; If I were you (but my French is not that good) I would do this: Use a comparator powered from the 12V side, with a pullup on the output to generate a 5V signal. Zeners are such a crappy way to shape input. Set a threshold on one of the comparator input pins to half of the signal, and put the signal from the alternator into the other. BUT, run thealternator signal through an RC circuit, say 10K and 100nF, to act as a filter. That ought to do the trick, but if not, there are ways to AC couple the signal to compensate for poor average value, as well as more sophisticated ways of filtering. I suggest the comparator because the way you have it with a zener is susceptable to noise, but also, you do not have good control over the threshold of where the signal is converted. You should not prototype with a microprocessor, where you measure the exact voltages of binary threshold. The manufacturer does not guarantee a precise threshold, only a range. With a comparator, you will have a precise threshold, and since the input is high impedance, you also get the ability to filter. Chris Eddy Jacky Joulin wrote: > Thank for yours many reply. > > I have test this circuit for read the "W" output alternator: > > input >------ R 2,2K-------------*-------------- R 10K ------ > to > pin RB6 ( input pic 16F84 ) > I > I > I > Zener 5,1v > I > I > I > 0v > > Results: > Truck at 600T/mn pulse = 2,2 ms F= 227 Hz > Truck at 700T/mn pulse = 2,0 ms F= 250 Hz > Truck at 2550T/mn pulse = 0,7 ms F= 715 Hz > > My problem : > Some parasites on the "W" output alternator degrade the signal end the result > is wrong. > > How it's possible to perfect the electronique circuit ? > > Thank for your help > > Jacky > > Weslay@aol.com