On Thu, 16 Oct 1997 11:09:13 +0930 Peter Wintulich writes: >It looks like you could use a similar method to the radio stations. >They send > broadcast quality audio over a two wire ballanced line, they have >center tap > the transformers at each end. Then they use earth as one wire and the >center > taps as the second wire. The result is the audio signal is sent differentially between the two wires in the pair, and the DC is sent common-mode. Good transformers (for minimum degradation of the audio signal, for recording-studio quality) are likely to be expensive. If you just need PA or broadcast quality then this could work well. > >You could use the same idea. >1) use the phantom power (about 48V) to power the Micros, I don't know why 48V would be required, use about 12V and pass it through 7805 regulators to power the PIC. >2) super impose the data on the power as a pulse per bit time. > > >I have a rough untested design that uses for the :- >Transmitter == A tristate o/p pin, dc isolation cap, two resistors as >voltage > divider. > Design the driver so if the line gets shorted (sudden change in voltage will cause the coupling cap to discharge through the driver), reversed, or overvolted, the PIC ouput is safe. This probably means using an external driver, not the PIC output directly. At high frequences, the transformer may distort attempts to apply a common-mode signal via the center tap. If it were an ideal transformer it wouldn't, but there will be some leakage in any real transformer. Connecting small capacitors from the audio lines (ends of the winding) to the center tap should solve this. >Reciver == An opamp, four resistors and a cap. > and maybe a 555 ic Yes, just couple the power line signal from before the voltage regulator into an amplifier and clean it up to logic levels. The voltage regulator may not like having a bunch of high-frequency junk at its input. Use a LC filter so the L is a choke to remove data from the power and the C gives the regulator the large input capacitance that they like. If the L in the filter is one side of a transformer then it would be easy to use the other side for data in-out using a differential line driver/receiver chip. The data channel won't have response to DC like a wired RS-232 connection does. For low bit rates, consider FSK modulation / demodulation using 4046 or similar chips. This keeps the data signal frequency out of the audio range. At high bit rates (likely higher than a PIC with something else to do can generate easily directly), use a Manchester code or similar so there is no DC content. The Manchester code will also limit the low-frequency component of the data signal to about half the bit rate. > >Phantom power is often taped off the ballanced line with a couple of >resistors. This could work for a low power device (like a PIC circuit). Not using a transformer means the transducer (microphone, etc) operates with the DC and data still common-mode.