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.

OMAP-L138 Peripheral power

Other Parts Discussed in Thread: OMAP-L138

Hi,

On my board I have quite a few peripherals connected to GIO pins of OMAP-L138.  The processor is subjected to power-up sequence and +3.3V GIO power is the last in that sequence. At the same time my peripherals can get power (+3.3V) earlier than GIOs of the processor.

I wonder whether such situation is critical for the processor operation and is it necessary to postpone peripherals powering until OMAP-L138 completely passes its power-up sequence?

Thanks

 

  • Victor,

    Powering up external components before the GPIO can have the following issues:

    1. If GPIO is supposed to drive data out to the external component, the external component may see glitches during power-up and reset.
    2. If GPIO is receiving data from the external component, any voltage coming into the GPIO pin before power-up will be shunted to the 3.3V power supply through the ESD protection diodes on the GPIO pin -- this can affect the power up supply sequencing and stress the ESD diodes if excessive current is applied to a prolonged period of time.

    -Tommy

  • Tommy,

    Thank you for reply.

    1. I do not mind about glitches - this is not a problem in my case.

    2. To my regret by design requirements I can not power-up my externals after the processor. In such situation I found two possible solutions:

    a) Kind of a "bulletproof" solution - to use level translators powered from one side from Vdd_IO and from other side from external components Vdd. This solution is hardly acceptable cost wise and especially PCB area wise.

    b) To use series resistors in GIOs nets.  Those resistors can have value that is big enough to limit the ESD diodes current on the safe level (200 ohms for example). I could also pull the Vdd_IO rail down to the ground to avoid its early turn up because of input currents.  Don't you think this can help?

    What other approaches can be used?

    Thanks

    --

    Victor

  • Victor Kviat said:
    b) To use series resistors in GIOs nets.  Those resistors can have value that is big enough to limit the ESD diodes current on the safe level (200 ohms for example). I could also pull the Vdd_IO rail down to the ground to avoid its early turn up because of input currents.  Don't you think this can help?

     

    As I understand it, you are proposing pulling the VDD_IO rail (which is tied to the OMAP's IO pins) down to ground?   That will not help - all the internal circuitry in the OMAP peripheral is already tied to ground, and that is precisely the problem.  When the external peripheral sends output signals into the OMAP, the current will be shorting to ground, more or less.  This is stressful on both the OMAP and the external peripheral that is driving the line.

    The series resistors might help with this.  I am running into the same issue with my design, and I don't like it one bit.

    Note to TI: THIS REQUIREMENT IS PAINFUL!

    Our solution is to run all of 3v3 through a "Load Switch" IC, which I am not terribly fond of.  All the external peripheral ICs and the OMAP's IO rails recieve 3v3 at the same time.

  • Victor,

    Given your restrictions, I think the series resistor approach would work ok.  You can always modify the resistor values later on if needed.  Is there any way to have the external components drive logic lows only until the L138 has powered up?

    -Tommy

  • Tommi,

    My problem is that the L138 board will work as a universal module in several already designed (probably complex) systems. I can hardly control the behavior of the external devices in such situation. There is a good chance that during power-up sequence the external components will be not operating yet (not powered or 3-state) but this is not guaranteed at all.

    I will go with resistors - it seems the best price/quality wise solution.

    Thanks

    --

    Victor

  • Victor,

    My two cents to this generic problem. 

    Value of the series resistor is critical. If it is too low (a few hundred ohms) then the current that will flow from the protection diodes

    at the inputs will not be limited. If it is value is too high, such as kiloohms, then the input will be sensitive to the electrical noise. 

    Try to keep a balance between diode current and the noise sensitivity.

    Regards,

    Cengiz

    METRA Electronics