• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 34 subscribers
  • Views 1139 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

SN65HVD75: RS-485: x16 or x8 oversampling

Borut Cadez
Prodigy 50 points
Part Number: SN65HVD75
Other Parts Discussed in Thread: THVD1550

Hi.

We need to design a RS485 interface with baud rate 921600 bps. We plan to use the SN65HVD75 tranceiver.
CPU that will be used supports x8 and x16 oversampling at this speed on its UART interface. We can cover the x8 oversampling with single oscillator, but need additional for the x16, so the x8 oversampling is preffered. I would like to find out if x8 oversampling is reliable enough. The clock mismatch error is 0.16%. Max number of units on the bus is 32. We are slave on the bus. The masters we will need to work with are unknown at this moment, conditions on the RS-485 bus are unknown also.

Can you provide any insights into x8 vs x16 oversampling dilemma?

Thank you and best regards,

Borut Cadez

  • 0
    TI__Mastermind 25705 points
    Borut,

    Thanks for reaching out to us. It's a very interesting question. But before I can answer, could you give a bit of more details about your system? What kind of cable would you use, twisted pair, with shield? How far could the cable be? Do you plan to put termination resistors at both ends of the network to match the cable impedance? Is there any coding theme for your signaling (what's the maximum number of consecutive identical bits)? All these factors impact the signal integrity. You can refer to this application notes (www.ti.com/.../slla375.pdf) for some data of THVD1550 to get a sense of how the jitter looks like with different cable length and data rate.

    Regards,
    Hao
  • 0 in reply to Hao L
    Prodigy 50 points
    Thank you for your prompt reply.
    The cable is UTP type, CAT5e. It is not shielded. A and B signals are present in twisted pair. There is one more used pair in the cable, +12V DC and GND. We are still waiting for an answer for cable length, but I believe a practical limit would be around 100 feet. There is 120R termination in both ends. Stubs are small, a few inches. Master is fail-safe biased. There is no coding scheme on the bus. The protocol is plain UART. The idle symbol is 0xFF, during data transmission worst case is data change every 9 bits (start bit as 0 or stop bit as 1).
    Does power supply of the RS485 (5V vs 3.3V) plays a role here? We can go with either one. THVD1550 seems like a good choice for our application. Thanks for the provided informations.

    Regards,
    Borut
  • 0 in reply to Borut Cadez
    TI__Mastermind 25705 points
    Borut,
    It looks like that you get covered on all the aspects. Generally 5V transceivers is a better choice since higher supply rail can provide more room for signaling and better signal integrity. THVD1550 is our latest RS485 device that has good speed and protection. I think x8 oversampling should be enough with it. Of course you can confirm with test data. Please let me know if you have any other questions.
    Regards,
    Hao