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.

RS-485 Compatibility Between 3.3V-supply and 5V-supply Devices

Expert 3240 points

Replies: 15

Views: 18372

Recently, a designer asked, "Do you have an app note that explains why 3.3V and 5V 485 chips can operate on the same system?" 

Here is a short explanation of why you can mix and match RS-485 devices that use different supplies.

            The RS-485 standard (TIA-485-A) specifies the voltages and currents needed on the differential bus lines to ensure any compliant transceivers can communicate with each other.  The drivers must be capable of sourcing a 1.5V signal (either A > B +1.5V, or B > A +1.5V) into a standard load, and the receivers must be able to discriminate a differential signal as small as 200 mV (+ or -).  All of the original RS-485 transceivers used 5V supplies, which then allows for 3.5V of “headroom” for the driver circuits.  Newer designs for transceivers which operate with 3.3V supplies require more efficient driver stages, in order to develop 1.5V differential output signals.  

 

            On the receiver side, the allowable common-mode voltage range is from -7V to +12V according to the standard, which comes from adding 7V on either side of the original 0V and 5V power rails.  In order to discriminate a 200 mV differential signal on top of a 12V offset, the differential receivers first attenuate the input signals in order to bring the range within the linear region of the input stages.  For example, with 5V rails, a 4:1 attenuation through resistor dividers would “shrink” the 12V common-mode offset to 3V, which should be within the linear range of the input stage.  At the same time, the differential signal is also attenuated, so internally the receiver’s comparator stage would need a sensitivity of 50 mV in order to detect an external differential signal of 200 mV.  For devices with 3.3V supply, the attenuation must be slightly higher, for example a 6:1 attenuation would reduce a 12V common-mode offset to 2V; this will also require more sensitivity (about 33 mV) at the receiver comparator.

 

            These differences between devices with 5V supply and 3.3V supply account for the typically higher price for 3.3V transceivers compared to 5V transceivers; the output stages require larger transistors and the receiver stages require more accurate and sensitive circuit designs.  Texas Instruments used a variety of advanced process technologies and innovative circuit designs to develop RS-485 transceivers that fully comply with the RS-485 standard, whether the supply voltage is 3.3V or 5V.

-- Clark Kinnaird

15 Replies

  • Dear Clark:

      It seems that the major difference between 5V and 3.3V transceiver in price. Can we use 3.3V RS485 transceiver as 5V RS485 transceiver in old 5V system? So we can keep one 3.3V transceiver in stock for 5V and 3.3V system.

     

    Best Regards

    Mike Wang

    1534 Taipei 24/Feb/2012

  • In reply to Mike Wang70975:

    Mike, 

    In general, devices that are specified for 3.3V-supply operation are not intended for use with 5V supplies.  The concerns are (1) will the power supply damage the device (2) will the logic levels be compatible (3) standard compliance.

    (1) Check the ABS.MAX.Ratings to make sure the 3.3V device will not be damaged by applying the 5V supply to the Vcc pins.  Make sure this includes the highest voltage that the supply might reach, such as 5.25V for a supply with 5% regulation.

    (2) The logic levels on the R, DE, RE, and D pins for 3.3V devices will be designed to work with logic levels based on a 3.3V supply.  If used with a 5V supply, the output level from R may be higher than the inputs to a controller (using 3.3V supply) will accept.  This would only be a problem if the controller has a 3.3V supply while the transceiver has a 5V supply.

    (3) RS-485 devices designed for 3.3V supply may not be fully compliant to the TIA-485 standard when used with a 5V supply.  This typically won't cause any problems with interoperability, but you should consider any potential issues in the system.

    TI does offer devices that can be used with 3.3V or 5V supplies, which are specified and tested for operation over a supply range of 3V to 5.5V.  Check out the SN65HVD08, and the SN65HVD1780, '1781, and '1782.  The parameters for these parts are specified in the datasheet for both 3.3V and 5V supplies.

    -- Clark Kinnaird

  • If we decide to split the biasing at the bus, could it be a problem if the 3.3V transceiver and the 5V tranceiver try to bring the idle-state voltage to different voltage levels?

  • In reply to Ana Rodriguez:

    Split the biasing on the bus ?

    Are you talking CAN or RS-485?

    - regards, Thomas

    Best Regards,

    Thomas

  • In reply to Thomas Kugelstadt:

    RS-485

    Thanks for the quick answer!

  • In reply to Ana Rodriguez:

    If you're referring to failsafe biasing, then No, there should be no problems. WIthin the specified common-mode range of -7V to +12V the differential receivers in RS-485 only look at the differential input signal and NOT the absolute voltage referenced to ground.

    regards, Thomas

    Best Regards,

    Thomas

  • In reply to Thomas Kugelstadt:

    But what if the transceivers are trying to bring the bus to different voltage values? For example a 3.3V device with 680 Ohm biasing impedances (680 Ohm beween 3.3V and B+ and 680 Ohm between A- and GND) and a device with 47 KOhm to 5V. Both devices using as well 120 Ohm termination impedance.

    Wouldn't this produce an undesired current flow at the line?

  • In reply to Ana Rodriguez:

    Your 47k ohms at the 5V-node are so high-impedance compared to the 680 ohms on the 3.3V node that there is hardly any difference if you could remove the 47k entirely. Your bus is already biased through the 680 ohms anyway.

     

    Best Regards,

    Thomas

  • In reply to Thomas Kugelstadt:

     

    You can see at the attached simulation that there is a change 30uA between the circuit with and without the 47k resistors.

    regards, Thomas

    Best Regards,

    Thomas

    5V and 3V Biasing.pdf
  • In reply to Thomas Kugelstadt:

    Hi Clark and Thomas,

    Based your inputs, it looks OK to directly connect two RS485 trasceivers, one powered by 3.3V-supply and another one powered by 5V-supply. Please confirm this, or any concerns?

     

    Regards

    Sheng

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.