• State
  • Locked Locked
  • Replies 5 replies
  • Subscribers 33 subscribers
  • Views 564 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.

DP83848H: DP83848H clocking slower for non-Ethernet application

Kevin Jackson
Prodigy 120 points
Part Number: DP83848H
Other Parts Discussed in Thread: DP83822I

We want to use the DP83848H (or other TI 100BaseTX transceiver chips) in a non-Ethernet application as a data transceiver for longer than 100m (over Cat5 cable) distances.

We want to use the 100BaseTX line-coding mode (running at much slower rates).

We would input and output data to the MII that was custom (non-Ethernet data).

Is it possible to clock the DP83848H at a much lower frequency and still have the internal circuitry work i.e. the PLL and clock recovery etc?

Regards,

Kevin Jackson

  • 0
    TI__Mastermind 18425 points
    Hello Kevin,

    DP83848 can operate with >100m cable length. Running low channel utilization is also acceptable at 100BaseTx. Non-Ethernet data should also work. The input clock however needs to be 25MHz. How low were you planning to go? Also, can you provide a brief description about your application and why the input clock needs to be lower than 25MHz?

    -Regards,
    Aniruddha
  • 0 in reply to Aniruddha Khadye
    Prodigy 120 points

    Hi Aniruddha,

    We are trying to go 500m, with a data rate of at least 15Mbps.  The rationale for the slower clock is the high frequency signal attenuation in the Cat5 cable at these distances. We wanted to use the DP83848H at both ends of a custom link.

    The slower data-rate would reduce the high frequency components in the signal.

    Am I correct in assuming that the TX PLL oscillator and or the RX CDR oscillator are tuned to operate at close to 25MHz an cannot be pulled far away from this frequency?

    Bottom line is we are trying to achieve a data-link (not necessarily Ethernet) 15Mbps over 500m of Cat5.  Other techniques than an Ethernet transceiver could be used.  We are also investigating VDSL but would like to use something cheaper an simpler, VDSL seems like overkill and it consumes a fair amount of power.

    Other suggestions are welcome.

    I have investigated RS422 with pre and post emphasis but it seems only 5Mbps could be achievable over these distances, hence I was trying to leverage the more bandwidth efficient encoding of the DP83848H with its MLT-3 encoding in a cheap PHY.

    Is there any way to fool the trick the clocking circuitry within the DP83848H to operate at the 10BaseT rate with 100BaseT line coding? This would at least give use 10Mbps.

    Regards,

    Kevin Jackson

  • 0 in reply to Kevin Jackson
    TI__Mastermind 18425 points
    Hello Kevin,

    I don't think phys can work in 10BaseT rate with 100BaseT encoding since the standards are different. If 10Mbps data rate is ok, would you be able to use 10M mode of DP83848?

    -Regards,
    Aniruddha
  • 0 in reply to Aniruddha Khadye
    Prodigy 120 points

    Hi Aniruddha,

    10BaseT rates might be acceptable (a little on the slow side).

    But I'm concerned that the 10BaseT is Manchester encoded (higher frequency components relative to 100Base T MLT-3), so probably won't be able to go 500m.  The Cat5 cable will probably attenuate the signal too much.

    Do you know how far 10BaseT might go on Cat5?

    Regards,

    Kevin

  • 0 in reply to Kevin Jackson
    TI__Mastermind 23675 points
    Hi Kevin,

    10BASE-T PHYs can reach around 200m over CAT5 for the DP83822I.
    For the DP83848H, you will be able to reach around 137m over CAT5.

    Anything beyond that and data transfer will not be reliable.

    Kind regards,
    Ross