• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 2 answers
  • Subscribers 63 subscribers
  • Views 294 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.

TPS27SA08: Accuracy of TJ SENSE

Sam O'Connor
Prodigy 10 points
Part Number: TPS27SA08
Other Parts Discussed in Thread: TPS1HB08-Q1

Tool/software:

The TPS27SA08 data-sheet says:

  • " accurately measure: – Load current, – Supply voltage, – Device temperature"
  • dISNST/dT = 0.0112 mA/°C (but neither min/max or accuracy is specified).
  • ISNST = (TJ – 25°C) × dISNST / dT + 0.85

The only application note I have found that mentions TJ sense is SLVAEX0 – OCTOBER 2020.
It mentions that the measurement "becomes more reliable over time", but I'm not sure that I really understand what this means.

In my tests with 4 TPS27SA08 ICs on two boards:

  • the voltage and current measurement functions work as expected,
  • the TJ temperature sense output seems believable at high temperatures (around 90C), but
  • at low temperature there is a lot of variation between the 4 ICs.

At ambient 23C, with switch on, but no load, and after sitting to stabilise for a few minutes, I see readings like 12C, 14C, 18C on different ICs.

Questions:

  • What is the intended accuracy of TJ SENSE?
  • What is the expected part-to-part variation in offset (0.85 mA @25C) and gradient (0.0112 mA/°C)?
  • If TJ SENSE "becomes more reliable over time", how long does it take to become accurate?

Thankyou,

Sam

  • 0
    TI__Mastermind 29240 points

    Hi Sam,

    Welcome to e2e!

    The TJ sense accuracy gets worse as the temperature (SNS output current) gets lower as well. Please use the TJ SNS limits of TPS1HB08-Q1 as a design guide. The TJ and part temperature take some time to settle due to the part self-heating, increased package temperature, etc as well as the SNS settling time itself (though this is very quick) - Figure 2.1 in SLVAEX0 gives an idea on how long this takes.

    The part-to-part variation should be captured relatively well in the TPS1HB08-Q1 TJ SNS specs/

    Thanks,

    Patrick

  • 0 in reply to Patrick Shalton
    Prodigy 10 points

    Thanks Patrick,

    From the TPS1HB08-Q1 numbers it seems that a 1.8 mA SNS current could indicate anywhere between 86C and 150C (based on the "max" number for 85C and the "min" number for 150C).

    (table from page 7, SLVSE16C –MAY 2019–REVISED JANUARY 2020)

    It seems that that software will have to be calibrated for each part to get useful measurements.

    Does TI have a recommended method of creating a known TJ condition in order to calibrate the SNS output?

    Maybe increasing load until thermal cutout and calibrating the last SNS reading before thermal cutout as 150C?

    Regards,


    Sam

  • 0 in reply to Sam O'Connor
    TI__Mastermind 29240 points

    Hi Sam,

    What purpose will you be using the temperature sense for? I would not recommend calibrating by when the device enters thermal shutdown unless you are using the temperature sense to warn of impending thermal shutdown, as the minimum thermal shutdown temperature is 160C, so it might be higher.

    I would instead recommend calibrating the temperature sense by calculating what the SNS current should be at each ambient temperature during your production test (or whenever you will be calibrating this), then correlating those values with the SNS current with no load in your calibration.

    Does that sound like it would work for you?

    Thanks,

    Patrick