• State Resolved
  • Locked Locked
  • Replies 12 replies
  • Subscribers 40 subscribers
  • Views 23393 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

data over dc power line using RS485(2 wire)

Knighty
Intellectual 380 points
Other Parts Discussed in Thread: SN65HVD96, SN65HVD1785

Can we use RS485, SN75176 to transmit data over DC power(2-wire)? any application note or reference design from TI?

- data rate 2400 Bd.

- distance 1000 mt.

NT

  • 0
    Prodigy 10 points

    Dear Knighty, I am Valter from Italy. I have the same question. I need to transmit and receive data , based on RS-485 standard, on a 2 wire copper lien where there is a DC Voltage ( +175 and -175 Vdc) . Did you get any response ?

    Thanks

    valter.pileri@ascpoint.com

  • 0 in reply to Valter Pileri
    Intellectual 380 points

    Dear Valter,

    Not yet, I waiting for TI whether they have suggestion or application note.

    NT

  • 0
    TI__Mastermind 28265 points

    Sorry, there is no apps note or reference design. In principle you would couple the signal capacitively onto the power lines (see below). common-mode jokes could be used to provide high isolate the data signal from the 12V supply and from the voltage regulator. A customer experienced difficulties with inductor saturationwhen transmitting NRZ data, so basically the original data without coding.

    To avoid saturation, either Manchetsre coding or FSK-coding should be used to remove DC components. Usually power line communications (PLC) as well as low-volt power communication use commercially available power line modems for example for LinkSprite http://www.linksprite.com/

    we therefore stepped away from competing efforts against these powerline pros. The circuit below however, should give you a pretty good start. and as long as you keep the power lines busy with data transitions, that is during idle you should send some kind of frequency, no major saturation effects will occur.

    Regards, Thomas  

  • 0 in reply to Thomas Kugelstadt
    Prodigy 20 points

    Has anybody tried this circuit, it's so simple it seems too good to be true.

    _______
    Rob

  • 0 in reply to Rob Gray
    TI__Mastermind 28265 points

    yes, quite a few have. With some minor mods of course. After grasping the concept of tranformer saturation, most use an idle frequency instead of letting the bus drift on a DC potential. Good luck in your endeavour. - Tom

  • 0 in reply to Thomas Kugelstadt
    Prodigy 55 points

    I'm replying to an old thread here, but nevertheless, this is a very useful circuit Thomas.  Thank you much.    That makes sense about using an idle frequency to avoid transformer saturation.  

    What would be the nature of the typical 'minor mods' that you mention?
     

    Thomas Kugelstadt said:

    yes, quite a few have. With some minor mods of course. 

  • 0 in reply to Thomas Kugelstadt
    Prodigy 10 points

    Dear Thomas,

    Could you advice me transformers for using that schematic? Thanks in advance.

    Demet

  • 0 in reply to demet aktaş
    Prodigy 20 points

    Thomas mentions using a common mode choke although the schematic shows differential mode chokes and I'm having trouble finding off the shelf parts, even less options for high current.  Anyone else?

    Aren't you also playing games with transmission line impedances?  Shouldn't the capacitance be tuned to the inductance the chokes put on the line?

  • 0 in reply to oxobfe
    TI__Mastermind 28265 points

    True, the inductors for high currents are a problem. Many Chinese OEMs approach manufacturers of magnetics components for custom designs of their chokes. The only way I see is contacting companies like Wuerth-Electronics or Coilcraft and ask their support staff for help.

    With regards to the line impedance, the 12V DC across 120 ohms in parallel (so 60 ohms) causes some major currents to flow. Again some customers don't apply termination at all for that reason. Also quite a few have turned towards the SN65HVD96, which has a CAN-type signaling scheme. A logic Low at the driver input causes the A and B lines to assume Vcc/2, while a logic high at the input causes a differential voltage on the bus. Still the signal waveforms on the bus look bad and distorted.

    One Italien customer simply changed to using CAT-5 where he uses one pair for the 12V supply and another signal pair for data transmission.

    Unfortunately we never had time to conclude our efforts towards a successful solution.

    Sorry for that, Thomas

  • 0 in reply to Thomas Kugelstadt
    Prodigy 10 points

    Hi Thomas,

    we're thinking about using SN65HVD1785 transceivers on an intra-PCB 24V bus. The overall length of that bus would be around 1.5m and we have one master and 4 slaves. The bus would carry 2A on those 24V. Would that be feasible with a design like you showed above?

    Regards

    Joern

  • 0 in reply to Thomas Kugelstadt
    Prodigy 10 points
    how about biasing instead of keeping the power lines busy with data transitions.
  • 0 in reply to dsgf sgfr
    TI__Guru 62920 points
    Do you mean creating a differential bias voltage on the input to the RS-485 receiver? That would be a viable solution if the low-frequency content present in the data being sent is always of one polarity (e.g., a period in which the bus is idle "high" and then transmission occurs in higher-rate bursts beginning with a low bit).

    By the way, we have since built up a similar circuit and had performed some testing. You can find more details here:

    www.ti.com/.../tida-00527

    One thing to note is that this kind of scheme (using a bias-tee network) tends to work best for higher-rate data transmission so that a majority of the frequency content of the transmitted data occurs above the high-pass corner frequency of the bias tee. For very low data rates, it is better to either artificially increase the data rate through modulation or to use a different scheme. One such scheme could involve using the SN65HVD96 device as a receiver to detect edge transitions. The output of the HVD96 chip could then be latched using a toggle flip-flop so that the logic state persists throughout the entire intended bit period (until the next signal transition occurs).

    Regards,
    Max