• State TI Thinks Resolved
  • Locked Locked
  • Replies 2 replies
  • Answers 2 answers
  • Subscribers 35 subscribers
  • Views 234 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.

TCA9535: H/L Threshold

Jeff YANG
Mastermind 7990 points
Part Number: TCA9535


Tool/software:

Hi,

Problem Description: We are using TI Expander (TCA9535) in our production line, with a VCC of 1.8V. Due to a new project design with weaker Pull down driving, we found that after voltage division with the expander PD0/PD1 pulled high, there is a voltage level of 0.7V. However, we discovered that the TCA9535 always reports low to the MCU via I2C (in the @ VIH/VIL floating region, it does not intermittently report high or low).

Questions:

  1. In the design of TCA9535 GPIO, is there a significant offset for each IC, resulting in large differences in the H/L threshold?
  2. Regarding the first question, what is the rough threshold? (Ex. 0.7/0.8/0.9V)
  3. If the output is 0.7V to PD1, and we use I2C to get the address: 0x23 for bit1, is it expected to report H or L normally?

Jeff

  • 0
    Guru 318050 points

    The threshold is expected to be near VCC/2. But for voltages between VIL and VIH, the resulting logic voltage is not guaranteed, and there are other reasons to avoid them; see [FAQ] How does a slow or floating input affect a CMOS device?

    To lower the voltage, use a weaker pull-up resistor.

  • 0
    TI__Guru 51598 points

    Hi Jeff,

    Jeff YANG said:
    In the design of TCA9535 GPIO, is there a significant offset for each IC, resulting in large differences in the H/L threshold?
    Jeff YANG said:
    Regarding the first question, what is the rough threshold? (Ex. 0.7/0.8/0.9V)

    I agree with Clemens here. The expected threshold somewhere in the middle VCC/2. There is no exact threshold because I cannot guarantee anything outside the datasheet specifications. We must abide by the VIH and VIL for the device. 

    As Clemens also mentioned and with the attached FAQ, there are other reasons on why having a CMOS input at mid-rail is not a good idea. 

    Jeff YANG said:
    If the output is 0.7V to PD1, and we use I2C to get the address: 0x23 for bit1, is it expected to report H or L normally?

    A VCC = 1.8V means that VIL = 30% of VCC = 0.3 x 1.8 = 0.54V. 

    You are guaranteed a VIL when the voltage is <0.54V. 

    If the output to PD1 is 0.7V then I would classify this as indeterminate. It could be either H/L. My intuition would say that is would register as a LOW input. Now if we were to decide if 0.7V was a LOW input for millions of TCA9535's then I would have to side with the datasheet to be sure. 

    Regards,

    Tyler