Power line share

[S. Petrukhin]

Hi, friends!

Have you ever tested or designed a power and information transmission system over a power line?

The task is as follows: there is a power line for a 3-phase motor, 4 wires, 3 phases + PE, 380V.

It is necessary to power a very low-power low-voltage device from this line when the motor is turned off.
I think it is necessary to somehow inject low voltage into the line, and without fear of supplying high voltage to the line.
The minimal distance is 2-3m.

It is necessary to add own device to the rotation device powered through the rotary coupling. It is impossible to replace the coupling with another one with a large number of contacts.
Battery power is not suitable because access is difficult to replace the batteries for a long time.

[mikeselectricstuff]

If you only need to power the device when the motor is off, and "off" means it is isolated from the supply at the supply end (i.e. not going though a variable speed drive or similar) , then I'd be looking at injecting a high frequency ( maybe 10-20khz) signal between a phase and PE.
Couple this at each end with a fairly low value cap ( ? 1-10nf?), rated at the full supply voltage, so when the motor is energised, you won't get much current at mains frequency, so a simple zener could be used for protection. The low cap value should also avoid any issues with earth-leakage tripping etc.

You could send low-speed data back by modulating the power draw of your device, so you only need one cap at each end.
If you need to send data to the device, you could use Manchester code on the driving frequency.

To drive the line, maybe use a small audio power amplifier, and detect the modulation by measuring changes in the DC power draw of this amp. 

[S. Petrukhin]

If you only need to power the device when the motor is off, and "off" means it is isolated from the supply at the supply end (i.e. not going though a variable speed drive or similar) , then I'd be looking at injecting a high frequency ( maybe 10-20khz) signal between a phase and PE.
Couple this at each end with a fairly low value cap ( ? 1-10nf?), rated at the full supply voltage, so when the motor is energised, you won't get much current at mains frequency, so a simple zener could be used for protection. The low cap value should also avoid any issues with earth-leakage tripping etc.

You could send low-speed data back by modulating the power draw of your device, so you only need one cap at each end.
If you need to send data to the device, you could use Manchester code on the driving frequency.

To drive the line, maybe use a small audio power amplifier, and detect the modulation by measuring changes in the DC power draw of this amp.

I am confused by the attenuation of the high-frequency signal in the power cable. The cable capacity can be significantly more than 10pF.

[mikeselectricstuff]

The cable and motor windings  could easily be 1nF or more, but if your coupling caps are 10nf each end, that would only attenuate the signal by 20%

Example : 10nF at 20kHz, impedance is about 800 ohms. at 50hz it's over 300K, so at 380VAC your protection devices only need to handle a milliamp or so, much less than 1 watt

You don't say how much power you need but "very low power" would imply less than 1mA, so 800 ohm impedance would drop less then a volt.
The way I would do this is to send maybe 9-12VAC, then at the receiving end rectify this and regulate it down to 5V or whatever is needed. 

 

[NiHaoMike]

What about charge up some supercaps when the motor is running? Might be possible to just mount it on the rotating part with some coils acting as a generator with the help of magnets mounted to a stationary part.

[Doctorandus_P]

You can use something like DTMF tone generators and detectors. Maybe simpler is to use a single modulator and do FM modulation of for example serial port data or go old school and use similar modulation as used for computer data on old audio cassette data storage mechanisms.

High (er) frequency signals go easily through capacitors so indeed a high voltage capacitor may give enough isolation, or use it in combination with a small signal transformer.

X10 is (was?) a commercial system and also duplicated by lots of hobbyists.

You also need to think about your isolation requirements.
For example if the "remote node" is in a solid enclosure, then you may not need isolation at all, and you can even use a capacitive dropper to make it's power supply. Such circuits are a nuisance to design and debug in a safe manner though.