| ??? 05/16/03 22:39 Read: times |
#45711 - RE: people try to understand Responding to: ???'s previous message |
Hal,
You are very correct, particularly in the case of the type of design that Waqar is trying to do where in he is hooking the phones up locally. In the case of the phone exchange things are somewhat different because the miles of wire to the exchange, the interface protection components at the entry to the phone company line cards and the types of sensing equipment at the phone company are all basically low pass filtering the signals. One also must deal with the large overshoot spike that can occur on a phone line when opening and closing the line that is casued by the inductance of older style dial phones or can be affected by the large ring coupling capacitors in some types of phones. Sampling the line at a periodic rate allows an easy filter to be made. Instead of just keeping track of current and previous states for the input one can shift the states of the detected line through a byte, this allows up to 8 poll times of filter information to be gathered. If one should choose a poll time period of 2 milliseconds then this shift type filter would hold 16 milliseconds worth of states. One would decide that the line was stable if say that the 6 most recent bits in this register were all zero's or one's. Michael Karas |



