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.

UCC2946

Guru 13485 points
Other Parts Discussed in Thread: UCC2946, UCC3946, TPL5000

need a watchdog timer which generates RESET when no input signal is asserted in the watchdog timeout (see the figure below).

Is the ucc2946 device – can it be suitable forthis requirements?

In this device the timeout is [25*Cwp], and when the capacitor Cwp is 1uF, for instance, then the timeout is 25 seconds (which the external counter we use will multiply by 12 to get 5 minutes). 

Can we use this device with such a capacitor?

What is the accuracy of this device?

  • Eli,


    Correct me if I'm wrong but by my understanding you want a watchdog timer that asserts a RESET signal if no signal is detected from the microcontroller within the watchdog period. If so, the UCC2946 is capable of doing this if you AND the outputs WDO and RESET together. From the datasheet:

    You can use the device with that large of a capacitor if the leakage current of the capacitor is well below the current source that is used to charge the capacitor (400nA typical). I would suggest that the leakage current should be at most 1/100th of the charging current so around 4nA or less. This will serve to minimize accuracy problems and instability problems of the watchdog timer.


    The accuracy of the device for the timeout periods will depend heavily on the output capacitor used and the layout of the board, and resistance in the path to the capacitor will reduce the accuracy, along with the leakage current of the capacitor. A rough estimate of the accuracy can be found in the datasheet specifications. Given all of the test conditions, it has a timer accuracy of 30%, but again this accuracy will vary based on a number of conditions.

    Regards,

    David

  • Hi

    As for the Cwp capacitor with around 4nA leakage current, can you recommend such a capacitor of 1uF (specific P/N)?

  • Hello Eli,

    For very low leakage capacitors, one generally wants to use a film capacitor. An example of this would be the following acrylic film capacitor, although there are other film types that may be more desirable, for various reasons:

    http://industrial.panasonic.com/www-cgi/jvcr13pz.cgi?E+PZ+3+ABD0012+ECPU1C105MA5+7+WW

    Regards,

    David

  • Hi DAvid,

    Thanks.Do you have its EVM (for UCC2946).

    according to my calculation the leakage current (for 3.3V and the 1uF film cap) is ~11nA.

    If my calculation is correct, will this capacitor be OK for timer accuracy of 30%?

  • Hello Eli,

    We do not have an EVM available for the UCC2946.

    For the timer accuracy, the UCC3946 has a more accurate timer and under listed conditions its accuracy should meet  that specification. I would advise the UCC3946 over the 2946 for tighter accuracy specifications.


    Regards,

    David

  • Hello David

    The UCC3946 is for temperature of 0-70C while the UCC2946is for -40-85C, so that’s probably the reason for the tighter accuracy of UCC3946 .

    I can only guess that the accuracy of the UCC2946 is basically similar to the UCC3946 within the 0-70C temperature range.

    Can you confirm?

  • Eli,


    Our accuracy specifications include a number of factors, including temperature, but there are other variables that do affect the accuracy specification.

    Regards,

    David

  • Hi The questions if their sre diffetences in the acurscy betwern th2946 snd 3946
  • Eli,

    There are other differences outside of temperature.
    Regards,

    David

  • Hello David

    Regarding tolerance, beside the temp. which are the other relevant parameters that differences between the two version which influence on their accuracy??

    Customer need to use -40C to +85C.Is it correct that it has less tolerance?

  • Hello Eli,

    There is less tolerance on the UCC2946.

    There are several factors outside temp. In just the specifications for the device, the operating voltage range is also different for the 2 devices (2V to 5.5V for the 2946, 2.1V to 5.5V for the 3946).

    The major factor that will affect the accuracy will be the capacitors used for the timer, which can improve or degrade the accuracy by a large margin. The tolerance of the capacitance, the leakage current, the temperature coefficient, the lifetime change in capacitance, the capacitance drift, and more will effect the accuracy. As such, the PDS accuracy specifications are only for the listed conditions.

    For looking at the requirements that you are asking for, I would advise that you look at using a timer with a watchdog function, or using a supervisor with a watchdog function that can trigger a timer that can provide the accuracy needed. TI's product line of timers are available here:

    http://www.ti.com/lsds/ti/clock-and-timing/timers-products.page

    Regards,

    David

  • Hi David

    the link dose'nt work

    Can you send the correct one?

    B.regards

    Eli

  • In any case, we did look for watchdog timer in TI’s website. That’s how we found the UCC2946 

    Do you  suggest choosing a different device? Which one?

  • Eli,

    I am sorry to hear that the link didn't work, but this one may:

    http://www.ti.com/lsds/ti/analog/clockandtiming/clock_and_timing.page

    The UCC2946 is meant as a supervisor for a power rail into a micro-controller to ensure that the rail for the MCU is fully on before the MCU starts to function. Supervisors typically have a much shorter delay time than you are looking for.

    TI has a portfolio of dedicated timers that would be more suited for your application. A device that may meet your needs is the TPL5000, but I haven't worked with that device much. The link is below:

    http://www.ti.com/product/tpl5000

    The E2E forum for this device is here:

    http://e2e.ti.com/support/clocks/f/48.aspx

    Regards,

    David