• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 34 subscribers
  • Views 697 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.

TCA9509: Logic level compatibility with slave I2C devices causing problems with ACK translation

Mitchell Kary
Intellectual 630 points
Part Number: TCA9509
Other Parts Discussed in Thread: DS92LV0422, , PCA9306

Hi,

We have a board design where the TCA9509 is used as a +3.3V to +1.8V I2C translator with the master device on the B side running at +3.3V and four DS92LV0422 slave TI devices connected to the A side. Ack's from the slave devices are not getting translated across the buffer to the master and the bus is hanging. Looking carefully at the data sheet for the TCA9509, I see now that the DS92LV0422  logic low is specified as .36V max and the TCA9509 has a VILC min of less than .15V. I assume this is why the part is not working. The requirement of a .15V logic level seems very restrictive almost to the point of making the TCA9509 unusable in most applications. Does TI engineering have any suggestions (other than designing the TCA9509 part out) for making the application I have work?

Thanks,

Mitch

  • 0
    TI__Guru 53575 points

    Hi Mitch,

    Your understanding of the TCA9509 requirements are correct. However, the logic low output of the DS92LV0422 (Vol) can be adjusted by changing the value of the pullup resistor on the A side I2C lines. It is likely that a strong pullup resistor is not allowing the voltage to drop low enough for the voltage translator to register a LOW. A weaker pullup will allow the slave device to pull the bus voltage closer to GND potential and under the Vilc of the translator. This is outlined in section 10.2.2.2 of the TCA9509 datasheet "Vilc and Pulldown Strength Requirements".
    More information on Pullup Resistor Calculation can be found in this Application Report:
    www.ti.com/.../slva689.pdf

    I hope this information helps. If you have any questions regarding this or this does not solve the Ack problem, please let me know.

    Regards,
    Eric

  • 0 in reply to Eric Schott1
    Intellectual 630 points

    Hi Eric,

    The DS92LV0422 devices are on the A side of the TCA9509 which is the side with the 1mA current sources hence there are no pull ups. I attached a picture in which the purple trace is the side A 1.8V slave SDA and the blue trace underneath is the B side master SDA.

    -Mitch

  • 0 in reply to Mitchell Kary
    TI__Guru 53575 points

    Hi Mitch,

    I am not able to view the image you attached. Could you try to upload again?

    Apologies for my misunderstanding. In the situation you describe, I cannot think of a solution that does not involve changing the DS92LV0422 device to something that is able to sink more current (pull lower than 0.15V) or swapping the TCA9509 with a device with a higher Vil MAX such as the PCA9306.

    As to the reason the Vilc of the TCA9509 is so low: The pull-down driver on the A-side LOW is set for approximately 0.2V. An input threshold exists on the internal buffer of about 50mV so that the device can distinguish between an output LOW and input LOW. These two must be distinct in order to prevent bus lockup. To have a low output voltage of 0.2V requires the input voltage to be the 0.15V. The pull-up for this device is provided by an internal current source. Any device hoping to supply an input LOW to the TCA9509 would need to be able to sink enough current to drop the voltage of the SDA/SCL lines below 0.15V. Sadly, the DS92LV0422 does not appear to be able to do this.

    Let me know if you would like assistance with a design including PCA9306 or if you would like more options for replacements.

    Regards,
    Eric

  • 0 in reply to Eric Schott1
    Intellectual 630 points

    Hi Eric,

    Unfortunately, I was expecting that was the answer...If TI had made the PCA9306 pin compatible with the 9509 QFN package, then it would have provided me with a seamless solution.

    Thanks for the support.

    -Mitch