• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 79 subscribers
  • Views 260 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.

12bit A/D REF in fast Mode

Tomas Grajales
Prodigy 80 points


I have an application where battery life is critical, and want to speed up the A/D conversions to as fast as posible, and go back to LPM3 most of the time.  The recommended values on the A/D REF of 10u and 0.1 u are affecting the reads.  I can measure in the scope 10 msec..  I want to reduce the large cap to 1 u and reduce the rise time to 1 msec.  I tested it and can live with the small ripple( few bits), but want to know if its acceptable and if it would increase with temperature swings.  I have extreme temperature requirements.  Any advice? 

  • 0
    Guru 227245 points

    Tomas Grajales said:
    The recommended values on the A/D REF of 10u and 0.1 u are affecting the reads.

    How? These are required to stabilize the reference for high sampling speed. However, a larger capacitance will significally increase the reference settling time.

    You should calculate the power consumption if using a different setup:

    • set up a timer for startign a smaplign cycle
    • go into LPM
    • when the timer comes up enable ADC and reference and set the timer for reference settlign delay, then back to LPM
    • next timer interrupt start the conversion, slow enough to not need any external capacitance, set up ADC interupt and go back to LPM
    • once the conversion is done, disable ADC and reference, process teh conversion result, then start over.

    This results is a lower sampling frequency, but you don't seem to need a high sampling frequency anyway. But this procedure reduces the power requirements by only enabling the reference and ADC for the required time, does not need an external capacitance to be charged and maintained.

    Depending on your requirements, this may enlarge battery life significantly.

  • 0 in reply to Jens-Michael Gross
    Prodigy 80 points

    I thank you for the response.

    You are correct, I only need to wake up and sample every minute, but the selected  battery should last  for years.  Having the reference settling time of 10 msec, uses where is using various milliamps is the problem.  Like I said I tested and reduced the cap to 1 u F, and the time is reduced to aprox 1 msec, which i am happy with.   The question is why TI does not show values smaller than 4.7 uF., should I have concerns with parameters like temperature?  I don't think so, but want to check if someone has.  

  • 0 in reply to Tomas Grajales
    Guru 227245 points

    Tomas Grajales said:
    Having the reference settling time of 10 msec, uses where is using various milliamps is the problem

    Using a slower sampling time, so the internal capacitance is sufficient (and maybe using the reduced reference strength for <50kSamples) willr educe teh current draw during reference settlign significantly.
    Various milliamps seem a bit much anyway.

    Tomas Grajales said:
    The question is why TI does not show values smaller than 4.7 uF.

    My guess is that thsi value will ensure that the specified precision is met unde rall conditions. Smaller capacitance may or may not result in the expected performance, based on capacitor quality and ESR, VCC, whatever.
    It may be that 100nF are sufficient, but nobody tested a large-enough number of situations to give any guarantee.

**Attention** This is a public forum