• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 35 subscribers
  • Views 1571 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.

PCA9306 Bandwidth - Ethernet serial Management interface (MDC/MDIO)

Antony
Intellectual 300 points
Other Parts Discussed in Thread: PCA9306, LSF0102

Hello,

In our current design we are planning to use PCA9306 I2C voltage level translator for Ethernet MDC/MDIO signal.

Hope we can use this part for this application  , from the datasheet we see this device can operate at speeds of >100 MHz .

we also found LSF0102  part , please advice which one works well for our application.

Thanks!!

Regards

Antony 

  • 0
    Intellectual 300 points
    Experts please answer my question :-)

    we need to finalize the design based on your inputs .

    Regards
    Antony
  • 0
    TI__Guru* 92935 points

    I've moved your question to a more specific Forum.  can you/team review this query?

  • 0 in reply to Blake TI Community
    TI__Genius 16540 points
    Hello Antony,

    The PCA9306 and LSF0102 are almost identical devices, and do support higher bandwidth > 100 MHz. However, it is not a buffered device.


    To give you a short answer: Yes, the PCA9306 can be used for translation on the MDIO interface, it is very similar to I2C. My question to you is at what voltages are you translating between, and what pull-up resistors are you tentatively planning to use?



    Longer answer:
    In order to up translate, it requires a pull-up resistor on the higher voltage side to pull the voltage to the rail. This limits the translation frequency.

    The PCA9306 (and the LSF part in question), both are essentially a FET with the gate connected to the EN pin, the drain to one side, and the source to the other. There is no designated drain or source side, it doesn't matter for this case. Translation occurs when the EN pin is biased such that the FET goes high impedance when your VIH has occured on one side, allowing a pull-up resistor to pull the bus up the rest of the way.

    Down translation is similar, but instead of relying on a pull-up resistor, you can bias the EN pin to ~0.7 V above your desired output voltage, and the switch will 'turn off' once the VGS < 0.7V.


    With this, you can see how you will be able to down translate at faster speeds than up translate (Generally speaking). The bandwidth of the devices is purely the bandwidth of the internal switch when EN is at a high voltage (turning the switch "on").
  • 0 in reply to Jonathan Valdez
    Intellectual 300 points

    Hi Jonathan,

    Thanks for the detailed explanation.

    At the processor side we have 1.8V . So , VERF1 is connected to 1.8V and at the Ethernet PHY side we have 3.3V . So, we connected VREF2 and EN pin tied together to 3.3V via 150K pull-up resistor and both the sides we have used 4.7k pull-up resistor for SCL and SDA signals.

    Regards

    Antony 

  • 0 in reply to Antony
    TI__Genius 16540 points
    Antony,

    This sounds perfect.

    I assume when you say VREF2 and EN pin tied together to 3.3, you have VREF2 and EN shorted together, and then that shorted node gets pulled up to 3.3V through a single 150k resistor (I've seen some try to use 2 resistors, 1 for VREF2 and 1 for EN, this won't work).

    Your pull-up resistors sound appropriate for moderate speeds. Depending on your load capacitance, you may require lower resistance to maintain higher speeds (I don't know what speed you're running at).
  • 0 in reply to Jonathan Valdez
    Intellectual 300 points

    Hi Jonathan,

    Yes, we have shorted together VREF2 and EN , and then that shorted node gets pulled up to 3.3V through a single 150k resistor.

    we are running at 25MHz 

    Thanks !!

    Regards

    Antony