Data with dc power

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

ruffrecords

Well-known member
Joined
Nov 10, 2006
Messages
16,155
Location
Norfolk - UK
Is there a simple/standard way to send data down say a 12V dc supply. I am thinking in terms of a small uC board per channel of a mixer. It connects across its own dedicated bused supply. They talk to each other over I2C but I am short of pins so I was wandering if I could somehow send the data over the 12V supply.

Cheers

Ian
 
You could probably use FSK modulation (cap coupled in) with a high enough frequency to decouple from the power line with inductors at each end, but you would probably need to shield the power line to keep noise out of stuff.

Probably cheaper to just run another conductor, with less unintended consequences, certainly easier.

JR
 
I agree totally with JR, but, just out of curiosity, you may want to look at what we did at RTS.
Data was voice and signalling (20k and 24k tones)
Power supply used an LM317, active impedance via positive FB.
http://www.manualbirds.com/manuals/download-rts-ps8-user-manual-106013
The beltpacks used a basic Hewland current pump, powered via an LM317
http://manualzz.com/doc/1173661/rts-bp300-service-manual
Indeed there are issues: the voltage from the supply must be significantly higher than the working voltage of the circuit it powers, and since it's current driven, there may be ground currents.
 
Thanks Abbey - fascinating info. It does confirm that the added complexity is not worth the effort. I need to make the effort to find a spare pin.

Cheers

Ian
 
This sort of thing is done all the time in industrial and automotive systems. It's called a can bus and shouldn't be that hard to implement.
 
gemini86 said:
This sort of thing is done all the time in industrial and automotive systems. It's called a can bus and shouldn't be that hard to implement.

Interesting. I have one some on line searching but I cannot find anything that says CAN bus is carried on the dc power lines. Can you point me in the right direction?

Cheers

Ian
 
ruffrecords said:
gemini86 said:
This sort of thing is done all the time in industrial and automotive systems. It's called a can bus and shouldn't be that hard to implement.

Interesting. I have one some on line searching but I cannot find anything that says CAN bus is carried on the dc power lines. Can you point me in the right direction?

Cheers

Ian
+1  there is a DC-BUS https://en.wikipedia.org/wiki/DC-BUS

CAN bus is built into some micro's I use (i never used CAN). Looks like it takes 2 wires by itself.

Thinking out loud... If you designed your own communication protocol that was always sitting positive 12V at rest, with momentary negative (0V) pulses to frame and transmit data. A diode and capacitor after the logic receiver could insure continuous power to your circuit.

A little odd, and perhaps not very fast, not to mention a huge noise source, but you could pass power and data over single wire and ground pair. 

Sounds like more work effort than it is worth.

JR
 
gemini86 said:
This sort of thing is done all the time in industrial and automotive systems. It's called a can bus and shouldn't be that hard to implement.

The CAN bus requires a separate power rail apart from the data pair.

-a
 
JohnRoberts said:
ruffrecords said:
gemini86 said:
This sort of thing is done all the time in industrial and automotive systems. It's called a can bus and shouldn't be that hard to implement.

Interesting. I have one some on line searching but I cannot find anything that says CAN bus is carried on the dc power lines. Can you point me in the right direction?

Cheers

Ian
+1  there is a DC-BUS https://en.wikipedia.org/wiki/DC-BUS

CAN bus is built into some micro's I use (i never used CAN). Looks like it takes 2 wires by itself.

Thinking out loud... If you designed your own communication protocol that was always sitting positive 12V at rest, with momentary negative (0V) pulses to frame and transmit data. A diode and capacitor after the logic receiver could insure continuous power to your circuit.

A little odd, and perhaps not very fast, not to mention a huge noise source, but you could pass power and data over single wire and ground pair. 

Sounds like more work effort than it is worth.

JR

Yes, I found aTI app note for something like this. It uses series inductors to isolate data from  from power and caps to couple data to/from bust transceivers. Data is transmitted differentially  between ground and power.

https://drive.google.com/file/d/0B_n67A1hN3qtSUxWOE1rR3hocTQ/view?usp=sharing

But as you say, probably more trouble that it's worth. Right now I am toying with using a single wire bidirectional asynchronous serial data bus.

Cheers

Ian
 
ruffrecords said:
Right now I am toying with using a single wire bidirectional asynchronous serial data bus.

This will work.

If I get your concept correct, you have a bunch of channels in your mixer, and each has a small micro. You didn't state this as such, but I assume that your mixer's master section has a master controller micro which lords over the channels. You daisy chain one wire from the master across to each slave.

The idea is exactly like USB and I2C, where you have a half-duplex bus with a master and a bunch of slaves, and the slaves only speak in response to requests from the master. 

The trick with any half-duplex interface is that you must avoid bus contention. Your master owns the bus and writes out a data packet to a slave, presumably including slave address in the packet. Once the packet transmission is complete, the master releases the bus and waits for a response from the addressed slave. Only once a slave has decoded the address will it be allowed to drive the bus. The master needs a time-out to ensure that it doesn't wait forever for a non-existent slave to respond.

The usual micro UARTs have separate transmit and receive lines, and really aren't meant for half-duplex operation. You can either figure out how to tristate the transmit pin when not sending, so you can tie the TX and RX pins together, or add an external little-logic tristate buffer and use another micro pin as the transmit enable. And you'll have to disable the receiver while transmitting, otherwise you'll receive what you wrote out and you'll have to deal with it.

If you need help, ping me. Except if you're using a PIC. I hate those things.  ;)
 
The single-pair digital protocol I've seen was the HART system used in industrial instrumentation to put bi-directional digital on top of 4-20mA current loops.

Not exactly the same, but it's an FSK protocol and some of the available parts may be useful if trying to implement FSK-over-DC-Power of some sort if you don't go with a more standard approach.

This sort of thing:
http://www.analog.com/en/products/interface-isolation/hart-4-20ma-loop-interface-devices/hart-modem.html
I believe we used a Maxim part to generate the FSK and mixed it in via analog to the power circuitry on the board I worked with.
 
I think you've already been presented with better options, but food for thought:  model railroad Digital Command-Control protocols (https://en.wikipedia.org/wiki/Digital_Command_Control) send data and power down a metal rail by alternating between +12V and -12V logic levels, which looks loosely like a frequency-modulated scheme (long pulses for 0 and short for 1).  It's immensely easy

You can tap from this wire via a diode clamp and recover the original digital signal, and you can also send the same signal through a full-wave rectifier to generate clean +12V and -12V rails from the same wire, with minimal cap decoupling.

Biggest disadvantage:  the data carrier is specified at 8kHz...but I think it could be run much higher since you don't have to worry about transmitting down 100 meters of large gauge track. :)
 
If you are still looking look up "bias tee" it's very common to do this in RF. The only issue you might have is if your data rate is low needing too large inductors and caps.
 

Latest posts

Back
Top