SLLA595C march   2023  – june 2023 THVD1424 , THVD1454

 

  1.   Abstract
  2.   2
  3.   Trademarks
  4. 1Typical RS-485 Networks and the Need for Termination
  5. 2Network Length, Data Rate, and Stubs
  6. 3Discrete Design for Switchable Termination and Duplex Switching
  7. 4Discrete Design for Duplex Switching
  8. 5 THVD1424 and THVD1454 Flexible RS-485
  9. 6Application Diagrams with THVD1424
  10. 7Experimental Results from Four Node Testing of THVD1424
  11. 8Conclusion
  12. 9Revision History

Typical RS-485 Networks and the Need for Termination

Figure 1-1 and Figure 1-2 illustrate typical RS-485 networks in either half-duplex or full-duplex configurations, respectively. In these topologies, the participating drivers, receivers, and transceivers connect to a main cable trunk via network stubs. A stub is the electrical distance between a transceiver and cable trunk, and essentially represents a non-terminated piece of bus line.

GUID-20221011-SS0I-CD3X-WX6B-8XDHGTLRHKN7-low.svg Figure 1-1 Full-Duplex Network
GUID-20221011-SS0I-LS6H-LRVQ-9V6M5XL0WRZL-low.svg Figure 1-2 Half-Duplex Network

The full-duplex implementation requires two signal pairs (four wires), and full-duplex transceivers with separate bus access lines for transmitter and receiver. Full-duplex network allows a node to simultaneously transmit data on one pair while receiving data on the other pair- thus achieving high effective throughput. In half-duplex network, only one signal pair is used, requiring the driving and receiving of data to occur at different times. This configuration reduces the network cabling cost (compared to full-duplex network) at the expense of reduced throughput.

Most RS-485 transceivers available in the market are either half-duplex or full-duplex, meaning devices are in different pinout and packages. This is the first problem for system designers to select different devices for their half-duplex and full-duplex design platforms.

Electrical signal travels in the copper cable (physical medium) from the driver to all receivers on the network. While driving the network, the driver (TX) output impedance is low, whereas the receiver’s (RX) input impedance is typically in kiloohms (kΩ). As shown in image below, every time the signal encounters impedance mismatch such as stubs of the middle nodes (at points A and B) or receiver input terminals (at Node n), some amount of signal is reflected back which interferes with the signal on the bus degrading the signal quality. Reflection factor ( r ) is given by Equation 1.

Equation 1. r = (Zt - Zo) / (Zt + Zo)

where Zt is the terminating impedance and Zo is the cable characteristic impedance

As per the transmission line theory, it is vital that the impedance mismatch discontinuities are limited to minimize reflections. To achieve this, recommended design practice is to keep stub lengths to a minimum and terminate the farthest node. If the signal can travel in both directions, both far ends of the network need to be properly terminated.

GUID-20221011-SS0I-4GFQ-QVTK-RGDXDH90CZHX-low.svg Figure 1-3 Signal Reflection Phenomena in an RS-485 Network

Bus termination is an effective method to improve signal quality. As shown Figure 1-1 and Figure 1-2, typically both end nodes are terminated with termination resistors whose value matches to the characteristic impedance of the transmission cable. In certain applications such as in building automation (HVAC, Thermostats, and so on), nodes can be added or removed from the RS-485 network to reconfigure it. This leads to second and third problem for system designers- application boards for end nodes have to be designed differently than the middle nodes, and a technician needs to manually intervene to re-configure the termination in network, which is prone to human errors such as inverting the polarity of the cable, improper termination, and so on.