# Send and "read" a tone over the AC power lines

I would like to use this circuit to send a signal over the a live wire;

Assuming the live wire on the right carry the signal;

simulate this circuit – Schematic created using CircuitLab

This would work? The signal can be between 1 kHz to 12 kHz.

PS. I have an electrician background, installing and maintaining equipment on 127/220/380/440V, so I understand the safety issues. I'm careful.

• I have yet simpler XTAL oscillator (6MHz) and AM radio to detect wires. Commented Aug 27, 2019 at 8:46
• You don't have a specific, concisely answerable question here. Likely what you should be doing is searching for existing solutions. Commented Aug 27, 2019 at 13:34
• Hi Chris, I do think I have one; In a nutshell: If I inject a signal with a uC to the live wire, how can I "read" it with another uC? What is the simplest solution?. I put the links of the protocol and the circuits to intend the idea, I know there is a way to inject a signal, but how I do the opposite and read it? . Thanks. Commented Aug 27, 2019 at 13:54
• @AlejandroSantiago The general search terms you need to look up are "Power-line communication, Power Line Networking". There are many solutions for data communication over lines used in conjunction with mains power. Solutions vary in complexity and design based on needs (are you trying to work just in your house? In across several miles? What data throughput is needed? etc.) Commented Aug 27, 2019 at 22:18
• @EasyOhm, the basic idea, is to send a diferent tone over diferent wires, and use a probe to probe the bundle, I can use pinch wire clamps. The only real need is to send and achive to read diferent tones over live wires. Not for long runs not beyond the 30m. Commented Aug 27, 2019 at 22:44

This is going to take some work. The biggest problem is the 1 kHz is fairly close to 60 Hz, being only a factor of 16 different. A simple filter will (at best) reduce line frequency by a factor of 16 compared to your tone. So, if you inject 1 V rms into the power line, your received signal will (again, this is best case) have 1 V rms at 1 kHz, and (120/16) or 7.5 V rms at 60 Hz. Since your proposed input to a uC is presumably a simple digital 0/1, you're going to have a hard time detecting the smaller signal.

In EE terms, you need a good low-pass filter, and a single cap isn't going to cut it.

Furthermore, transistors don't work "from back to front" as you seem to think, so you can't just take the transmitter circuit and swap inputs and outputs as your figure seems to think.

You'll need to do some research, both on low-pass filters and on comparator circuits. There are lots of comparator ICs on the market, and you would feed the comparator input with the output of your selected low-pass filter, which would only pass frequencies above some frequency like 500 Hz.

We're not a design service, so I'll let you get to work.

• Thanks for the insigth, from a lot of reading, I was able to see the frequency to send the pulse is 100Khz or so, and they are using an Capacitor inductor filter, and yes, will look foward on the transistors side. Commented Sep 29, 2019 at 20:35

I assume, that "reading" the signal works simplified very similar to sending it. I would try and decouple the two rails with a capacitor/highpass and amplify the noise (of which hopefully a good portion is the signal). Just make sure you filter all the frequencies you don't want.