• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 97 subscribers
  • Views 674 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.

OMAP L138 power consumption spreadsheet: Inconsistencies

Anonymous
Anonymous

Hi,

I would like to ask a question with I/O power.

I have no experience with 1.8V I/O pin before, but when I am trying to calculate power consumption difference between these two IO voltages, I observed that IO33 consumes approximately 3 times power than IO18.

 

 

So it seems that the difference between 1.8V and 3.3V voltage supply, in terms of power consumption, is based on the well-known basic formula:

P = U2/R

3.32 : 1.82 = 3.361, very close to the ratio revealed by the spreadsheet.

Inaccuracy

There are also cells in the table for entering trace length. For example:

 

 

When trace length doubles, because R∝L, R also doubles. Then

1.    By P = U2/R, P should get down to half

2.    By spreadsheet, P doubles ⟹ suggesting P=I2R∝R ⟹ I(current) constant.

We clearly have contradicting results.

But clearly I cannot be constant, otherwise if trace length triples, quadruple or even longer, a constant current means voltage drop along the trace would also change by these numbers, then the voltage at the receiving end might not be enough to drive the load.

I have very little knowledge with semiconductor, but I guess the actual model might be much more complicated than formulas included in the spreadsheet; and the purpose of this spreadsheet is only for a rough estimation, so the simplistic formulas it used leads to the observed inconsistencies above.

 

 

Zheng

  • 0
    TI__Expert 8785 points

    The power spreadsheet was created from collecting lots of power consumption data points using real silicon samples.  Therefore, the underlying formulas in the power spreadsheet are not just simple equations but have been correlated with silicon.

  • Anonymous
    0 Anonymous in reply to clam

    clam,

    Two questions:

    1. Is the ratio between IO33 and IO18 power consumption, namely, P(IO33) ≈ 3.361 P(IO18), accurate? This means using 1.8V I/O saves at least 2/3 power on I/O pins, quite surprising. Is it true?

    2. Could you try to give a reconciliation between the inconsistencies above (with R↑ , U2/R leads to decreased power, whereas I2R leads to increased power) from physics perspective?

     

    Zheng

  • 0 in reply to Anonymous
    TI__Expert 8785 points

    The power cannot be represented as a simple P=VI formula. 

    The trace length does not only change resistance.  It also changes the capacitance and load that the I/O pin sees.  Therefore, the P=VI formula cannot be used.

  • Anonymous
    0 Anonymous in reply to clam

    clam,

    I see it, my application of the simple formula was incorrect. Thanks.

    Zheng