“Your personal results may vary.”

“Example is not indicative of typical results.”

“Results not guaranteed.”

These and other disclaimers are very common at this time of year, especially with weight loss or fitness advertisements.  However, they should never be present with an integrated circuit’s performance!  All performance data, such as is found in a datasheet, should have a defined test setup that allows anyone with a reasonable amount of training to duplicate the measurement or waveform.  The result almost always shows typical and expected performance, so we should be able to see this same typical result.

So what happens when we go into our lab with the EVM for a device and don’t get the same answer that the datasheet shows us?  Of the many possible answers to this question, test setup differences and nuances are at the top of the list.  Do I have the device in question configured to the same operating conditions as the datasheet?  Am I applying the same voltages and currents to the circuit as the datasheet?  Am I using the correct equipment for the measurement? 

With many new ultra-low quiescent current (IQ) devices being developed these days, efficiency can become difficult to measure.  For instance, when I tried to measure the efficiency of the 360-nA IQ TPS62740, I did it wrong three times before finally understanding a correct test setup!  Luckily for each of you, I detailed my learning in this app note.  Now my results correspond to those in the datasheet: over 90% efficiency for almost the whole load range.

Let me know if these learnings were helpful to you or if there is something else to mention, so that all of us can reproduce and see for ourselves the amazing 90% efficiency at just 15 µA of load current.

 

Additional Resources:

Anonymous