Data on 12V supply wires

What are you sensing? The standard industry method is 4-20mA current loop, can be stretched to 10-50mA. The low end is the circuit quiescent current, the high end is max signal. Supply with unregulated 20v to 40V depending on loop V drop.

At the sensor end, a 5V or 12V zener gives you local supply voltage for electronics, then you modulate the current to indicate the reading,

4mA for zero, 20mA for full scale. Examples in LM10 opamp.reference datasheet and AN201, and other older opamp cookbooks.

Single opamp and a power transistor or MOSFET, plus signal conditioning and a zener for power, quite easy. Depends what you measuring.

Grant.

Yes, that seems like a good way to do it. For data going from the display to the sensor I'm thinking of varying the supply voltage - say by about

0.2V. Then current variation at the sensor for data going the other way.

I was hoping for something ready-made that would use modulated fsk data or similar - like a wireless system, but using the supply cables. I don't want to use a real wireless system because the probe is buried.

James

To transfer some digital information in addition to the analog 4-20 mA measurement, the HART protocol could be used.

Alternatively, take a look what has been used in various power line communication systems are using.

Anyway, the problem is how to inject and extract the data signal to the DC or LF AC power signal. Typically transformers are used, with the DC power flowing through a few turn primary and the signal connected to the multiturn secondary. The signal should not contain any DC components, so direct UART connection is not usable, but Manchester coding could be used.

FSK or On/Off keying could also be used. Keying a 567 PLL on the transmitter side and using an other 567 PLL as the receiver might work. For 1 kbit/s data rate run the PLLs at 50-500 kHz.

Hi James,

maybe an MBUS could help. But you have to double the 12V. Pls see:

Regards

Karl

If the load is not to large you can simply toggle the power line using a mosfet and filter afterwards. That is, a cap supplies temporary power to the device while the power line is off representing a change in state. As long as the data rate is fairly fast then it shouldn't be too much of a problem. The faster the better. This would require a mosfet, a cap, and a diode/mosfet for upstream only. Else it is more complex for bi-dir or downstream. Basically the device sending the data cuts it's power temporarily to signify a change in state. As long as the device has backup power then it will should surive. A long string of zero's of course might cause a problem and there are easy ways around this.

Thanks to all for ideas. I have put in a simple circuit to detect voltage pulsing of the supply line - by about half a volt or prehaps slightly more. Current pulsing at the sensor end. I don't know how it will be for noise immunity. If not good enough I will try adding a PLL later, looking at M-BUS and HART as well.

James

How about rethinking the requirement and use PoE, which has chips and other standard stuff.