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.

INA219: rise time (tRCL) spec definition

Part Number: INA219

Hi Team,

Customer measured the I2C rise time and fall time.

So far it seems violate our specs. (for both FAST MODE and HIGH-SPEED MODE)

Customer wants to understand how to read our tRCL?

Is it when over 300nS for FAST MODE and over 40nS for HIGH-SPEED MODE means fail and we don't guarantee the I2C traffic working correctly?

Especially for 40nS, it seems very fast.... 

Thanks!

Andrew

  • Hello Andrew,

    This looks like the pull-up resistors on your I2C bus may be too large. These resistors can significantly affect the rise and fall time, so you may want to try using a smaller value. The datasheet has 3.3kΩ as the typical value.

    Also note, as it says in your screenshot above, for the bus timing diagram definitions: "Values based on a statistical analysis of a one-time sample of devices. Minimum and maximum values are not ensured and not production tested."

  • Hi Mitch,

    Thanks.

    Customer is not worried about if communication error, actually it works fine. But customer is more like asking why we put this kind of spec.

    What is meant by 'Rise Time MAX' ? especially 40nS at High-Speed Mode. 

    Is it meant, clock rise time should never be slower than 40nS?

    I feel strange too, I2C should be a slow bus, so why is 40nS rise time max for?

    Could you please help me verify with design team for rise time tRCL definition?

    Thanks!

    Andrew

  • Hello Andrew,

    I talked with the design team, and they said, "Long rise times introduce some uncertainty in the threshold detection which might cause a setup problem for SDA depending on it’s timing relative to SCL such that a false START/STOP might be detected." So, if the clock rise time is too slow, it actually could cause some problems.  They also mentioned that one more way to get faster clock rise and fall times is to set the controller into push/pull mode, if available.

    To answer your question, yes, "Rise time Max" is the longest time that should be allowed for the rise transition.

    Also, as a side note, from you scope shot, it looks like the signal starts low instead of high (SCL should start high).  Maybe this is because we are not seeing the whole transmission, but just wanted to point it out just in case. 

  • Hi Andrew,

    this document might help:

    UM10204-1.pdf

    Kai

  • Thanks a lot,  Mitch and Kai!