• State Resolved
  • Locked Locked
  • Replies 1 reply
  • Subscribers 80 subscribers
  • Views 128 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.

TMS320F28033: Problem of External interrupt response time

Green Deng
Expert 4490 points
Part Number: TMS320F28033

Dear team:

When my customer uses 28033 to implement external input triggering XINT1 interrupt (rising edge trigger, other interrupts allow the interrupt to nest), under the condition of the external input signal unchanged, it is found that the delay of entering the interrupt has a maximum of about 300 ns Difference, I don’t know if this time level difference is normal?

He configured a total of 3 interrupts, XINT1 interrupt priority is the highest, and other interrupts allow XINT1 nesting. In the XINT1 interrupt service routine, set GPIO high when entering the interrupt, and pull GPIO low when going out of the interrupt. The following is part of the code:

The following is a comparison chart of entering the interruption time:

Best regards

  • +1
    TI__Guru 62975 points

    Green,

    300ns of jitter would not be expected from the perspective of the hardware state machine for interrupts.

    There are a few confounding factors that should be mitigated when taking the response time measurements:

    1. The input signal has a very slow slew rate so it is difficult to infer when the F2803x first recognizes VIH. Please use a faster (square) input signal for latency measurements.
    2. Confirm that the GPIO input qualification is set to SYSCLK (0)
    3. Disable other interrupts or add GPIO profiling to the other interrupts so that you can see if the response time is being affected by higher priority activity
    4. Execute the code out of RAM

    With a clean input signal to measure and no outside interrupt interference, the response time jitter should only be a handful of SYSCLK cycles.

    -Tommy