RX/TX over Power (LIN) Question

[pigman]

I'm messing around with the idea of transmitting serial between two Microcontrollers using power as the means of transmitting. After some research I've come up with potentially the MCP201 LIN Transceiver however I have a question for you Guru's.
https://www.farnell.com/datasheets/47858.pdf

On the Master MCP201 it talks about using a 1K pullup resistor on the LIN Bus. I was under the impression this solution would communicate over the GND wire but a 1K pulllup resistor to LIN (assuming LIN was GND) in say a 24V setup would end up pulling an unreasonable 24mA. So am I right in thinking the LIN Bus works on the + power cable?
And if it is the + side what would a 1K pullup resistor do, especially after a diode which would loose ~0.4v before the resistor and the LIN Bus would be up at + anyway?

Thanks in advance

 

[pigman]

I fear I may have this wrong is LIN looking for a standalone LIN Bus cable? What protocol/chip would I possibly use to provide RX/TX over Power?

 

[ronsimpson]

I can't find the parts I used last time. Philips??? 9600 over power line, 110/220, 50/60 no problem. 120-150khz carrier. Only works in the house will not talk to the neighbors. (will not talk through a transformer)

 

[pigman]

ronsimpson, thanks for posting mate. I'm working with DC power, will this chip do that also?

 

[Mr RB]

You want to send data over 24v DC power lines (2 wires; pwr and gnd)? What distance and what current is normally being carried on the 24v power lines? What data rate do you need to achieve?

 

[ronsimpson]

I'm working with DC power, will this chip do that also?

I did not understand DC!

Yes it will work the idea is the same. With DC you don't have to worry about high voltage, 60hz, rules on power line use, etc.
You could use the old LM1893. Probably not use the transformer.
1. Typically you have a power source with large capacitors across it.
2. Long wire.
3. Loads that might have large capacitors across it.
What I have done is isolate each device on the power wire with a inductor. The coil will be about 0 ohms at DC but have 1 to 10 ohms at you transmitting frequency. 100 to 300khz. Could use MHz. This way your transmitter can drive 300khz on the power wire and only see 10 ohms load. The power also flows because it see 0 ohms in the inductors.

 

[pigman]

Mr RB, the voltage will vary sometimes 12V sometimes 24V and yes just power and gnd distributed from the "master" The distance will be up to 25m but typically 10m and the current varies up to a few amps for short bursts ~20-80ms peaks or so. Data rate does not have to be great but 19200 baud would be nice I'd settle for 9600 or even lower if it was a super simple application.

ronsimpson, at the moment I use a 0r01 resistor infront of each output that will hopefully have this data element to sense current using a opamp, Are you suggesting after this current sense resistor I should setup the LM1893 and then install a 100uH inductor or something before the power goes out to the device to communicate with? What about the slave end? I also don't really need to daisy chain slaves if the implementation isn't too difficult there will be no need to have more than one slave on a line.
Each Device will have about 2200uF capacitance also.

 

[ronsimpson]

Lets say the master and slave have 2200uF caps. As the hot wire leaves each box add the inductor (that can handle the DC current).

The transmitter, through a DC blocking cap, drives a signal into the wire.

On the receiver, the inductor keeps the signal from being shorted out by the 2200uF cap. The receiver input, through a DC blocking cap, can see the signal on the wire.

 

[pigman]

ronsimpson, sounds easy enough. Any chance you could sketch a rough schematic to help me get my head around the implementation? I'm guessing the inductance isn't critical more the ability to handle the load. And this is to work with a LM1893?