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-L137: Powering the L137 D version (456MHz) at a lower frequency

Part Number: OMAP-L137


Hello,

we have two products using OMAP L137. The first was released when this processor began its life when only the 375MHz version was available. We did power it at 1V2 core voltage (recommended nominal). We are currently using the DZKB3 silicon.

The second product uses the 456MHz version (DZKB4) and is powered at 1V3 core voltage (recommended nominal).

The question is, in order to have only one reference, can we safely use the 456MHz version (DZKB4) for our first product and power it at 1V2 voltage, knowing that the core frequency will not go beyond 375MHz ?

In other words, what causes the increase in nominal core voltage requirement : is it only due to higher max frequency operating condition, or the silicon is different for the 456MHz version ?

The Recommended Operating Conditions in the Datasheet gives 1V3 nominal and 1V25 min for the 456MHz version, so it seems that it's not a question of frequency but silicon.

Thank you,

Eric

  • Hi Eric,

    Typically voltage scaling on CVDD is permitted, and a useful approach for power savings if you do not need the performance of the CPU/DSP at higher frequencies.

    I see how you came to have this question and I have sent out a request for clarification.

    The recommended operating conditions table mentions a min CVDD 1.25V for the 456 MHz version (regardless of operating frequency).
    It could have stated the min CVDD requirements for CPU/DSP operating upto 375MHz and 456MHz.
    Reading it directly makes it seem like the 456 MHz version (DZKB4) always requires 1.3V even when operating with CPU frequency less than 375MHz.

    I disagree with that interpretation, but will confirm internally.

    Table 3-1. Characteristics of the OMAP-L137 Processor seems to support voltage scaling on CVDD by stating…
    CPU Frequency MHz
    674x DSP at 375 MHz(1.2V) or 456 MHz (1.3V)
    ARM926 at 375 MHz(1.2V) or 456 MHz (1.3V)

    Also TRM section 10.2 Power Consumption Overview states “Dynamic power can be reduced by scaling the operating voltage, when the performance requirements are not that high and the device can be operated at a corresponding lower frequency.”

    And the power estimation tool allows 1.2V or 1.3V selection (though it does not complain about the combination of 1.2V with 456MHz.

    I will let you know when I have found out with confidence that you can supply 1.2V to the 456 MHz version (DZKB4) if it never exceeds 375MHz CPU/DSP frequency.

    Regards,
    Mark

  • Thank you Mark for your fast answer. I can see we have the same feeling, but that needs confirmation as we don't want to risk erratic functioning. By the way, do you have any news from the Engineers ?

    Regards

    Eric

  • Hi Eric

    Sorry for the delay- I am still working with my product engineering team to confirm that it is ok for you to run the 456 MHz part at 375 MHz @ 1.2V Nominal.

    I hope to confirm early next week if not sooner. 

    Regards

    Mukul 

  • Hi Eric

    Confirming that you should be ok running 375MHz @ 1.2V nominal using the ZKB4 part #. 

    The question is, in order to have only one reference, can we safely use the 456MHz version (DZKB4) for our first product and power it at 1V2 voltage, knowing that the core frequency will not go beyond 375MHz ?

    Yes. 

    In other words, what causes the increase in nominal core voltage requirement : is it only due to higher max frequency operating condition, or the silicon is different for the 456MHz version ?

    Yes. Primarily frequency. 

    Hope that helps.
    Regards

    Mukul

  • Thank you very much !