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.

OMAP3530: OMAP3530 - PMIC

Part Number: OMAP3530
Other Parts Discussed in Thread: TPS65920

I have adesign where an OMAP 3530 is powered from a PMIC TPS65920.

The VMMC output from the PMIC is very low and the part gets overheated. There is no short-circuit on this output.

The rest of power rails to the OMAP are O.K.

Using an oscilloscpe, I observed a periodic signal of about 200uS Low time and a period of about 7mS from the OMAP SYS_NRESWARM to the PMIC NRESWARM input.

In addition, when the system is powered-up, the OMAP UART3 outputs the 40W string to the terminal, but with a slightly lower baud-rate than 115,200 (about 110,000bps).

Please advise.

Thanks -

Koby Kluger

  • Koby,

    What does the 32kHz and UCLK line look like? Are they both running at the correct frequencies?

    Thanks,
    Nick
  • Hello Nick -

    I measured both clocks at the PMIC outputs and they look fine in terms of signal integrity and their frequencies were also accurate, as measured with a frequency counter.

    Thanks.

  • Koby,

    Can you try probing PWRON and NRESPWRON and see when those are going high or low compared to NRESWARM?

    Is VMMC the only rail that isn't going to the correct voltage level? Is there any way you could measure the current through it?

    Thank you,
    Nick
  • Nick - hi

    Following another conversation I found in the E2E, I understood that this phenomenon can happen because the OMAP Flash is empty, so I programmed it through the UART and now it wakes up properly with the correct baud-rate.

    But, now the PMIC over-heats even more than before (probably because the OMAP is active) and the VMMC line is not active.
    I have the OMAP removed on one board and the PMIC there is cold.

    Since there are no series components on the PMIC voltage rails, I cannot measure each current to see where there is an excessive consumption. The OMAP itself doesn't over-heat.

    Do you think there could be a problem with the rails sequencing, although they are controlled by the PMIC?

    Any other idea?

  • I would like to update:

    The OMAP runs its application program perfectly and doesn't overheat.

    All voltage rails supplied from the PMIC to the OMAP are within their tolerances. The OMAP is the only load on the PMIC, except for VIO SMPS, which we use as the board 1.8V main supply, which is connected to a few light loads. Just to remind: when we remove the OMAP from the board, all voltage rails supplied by the PMIC are still active but the PMIC temperature is low.

    Is it possible that somehow the OMAP overloads the PMIC and still functions normally?

    (It is important to note that because of the board structure, there is no possibility to measure each rail current separately).

  • Koby,

    I am not sure if OMAP would be able to function normally if it is overloading the PMIC. I suppose it may be possible.

    How are you able to tell that the PMIC is overheating? You said you were seeing the NRESWARM signal going low occasionally, but this is a signal to the PMIC from the processor. Is it still doing that after fixing the flash memory?

    If you want to send me your schematic I would be happy to look it over. In the meantime you can also try probing the VMCC LDO input to make sure it is being powered properly.
  • Nick - hi

    At fine last we found the problem: The SMPSs of the PMIC chip work at a few MHz and this means that there is a great significance to the serial inductors bandwidth.

    The Vishay inductors that were used, have only 5MHz bandwidth which is insufficient for the switching frequency and as a result, the SMPSs worked in their linear mode, without oscillating, which also caused them to over-heat. We replaced the inductors with Wurth parts which have close to 80MHz bandwidth and the PMIC works fine, and so does the rest of the board.