• State TI Thinks Resolved
  • Locked Locked
  • Replies 2 replies
  • Answers 1 answer
  • Subscribers 82 subscribers
  • Views 465 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.

Original question:

HRCAP timestamping

TMS320F28069: HRCAP timestamping

John Westmoreland43
Guru 13945 points

Part Number: TMS320F28069

To finally get back to Jim on this - there are a few practical issues - 

namely, I'm using the HRCAP IRQ's - and if those are continuously enabled; it's very easy to have your application to become 'interrupt bound' wrt HRCAP IRQ's.

Also in my application, I'm not running SYSCLKOUT and PLL2CLK at the same frequency since I want to run the HRCAP module at 120MHz.  I'm taking a 30 MHz

MEMs - multiplying that up by 4 for PLL2CLK and by 3 for SYSCLK (120MHz and 90MHz, respectively).

But, I want to think about this a little more to see if there's a way to do it and get an actual time stamp.

I am using the SYNC capability of the eCAP module with the HRCAP module; so there may be a way to scale the eCap counter to get a timestamp that is accurate.

Regards,
John W.

  • 0
    TI__Mastermind 42195 points

    John/Jim,

    Are you trying to measure an absolute value a long time from some event?

    I would be also be concerned about stack errors.

    1. From each measurement; this could behave in a random nature and would potentially average out, some values larger than ideal and some smaller than ideal.
    2. From the Calibration; Even if the calibration is off by a small amount this will error will stack for each measurement. It will always be high or low(until you recall the calibration) and your error will diverge in that direction.

    Regards,
    Cody 

  • 0 in reply to Cody Watkins
    Guru 13945 points
    Cody,

    I can't speak for Jim, but just having a timestamp would be useful for me.

    Thanks,
    John