MSP430FR2311: Profiling MPS430FR2311's Low Power Mode 3.5 (LPM 3.5) with EEVBlog's uCurrent

Hi Vijayenthiran,

That is a very thorough blog post!

For your first point of feedback, the order of some of the events you describe depends on your code implementation. In order to be more certain of the order of events and the timing, you can set GPIO pins to toggle at the beginning and end of the events such as when waiting for the oscillator flag or accessing the RTC ISR. Measuring these additional GPIO on the oscilloscope can verify the order you mentioned.

For the second point of feedback, it appears the numbers you are comparing are not derived in the same way. For Figure 14, you say that the area under the yellow curve is the current over a period of one second. The value 28.4 mVs appears to be the integral of the yellow curve over the whole screen. In order to get average current over the screen, you would divide the area by the amount of time over which the integral is taken or 28.4 mVs/2.4 s => 11.83 uA. Something similar for Figure 15 can be done giving the result 12.9 mVs/1.2 s => 10.75 uA average current over the oscilloscope window frame which isn’t too far off from Figure 14. This difference can be attributed to the proportions of time the device is shown in active and low power modes. Figure 14 shows the device in an active mode for a larger proportion of the time, causing the average current to increase.  

If you want to find the average current for a smaller section of time, you can use the oscilloscope cursors to find (BY-AY)/(BX-AX) of the violet line, but make sure that the horizontal cursor intersects the violet line at the same point as the vertical cursor.

To compare these results with the average shown by the multimeter, you should be careful that you’re measuring current for the device while it is only in LPM3.5 for both the multimeter and oscilloscope, and it might help to have the uCurrent GOLD set to 1mV/1nA for both measurements.

Texas Instruments has specialized equipment and procedures to test a large sampling of devices to set the values shown in datasheets to make sure all devices fall within the ranges shown. Using an oscilloscope and the hardware you have shown in your blog post is a good way to measure timing and current :)

You can also compare your results to simulated values generated in EnergyTrace in Code Composer Studio.

Regards,

Ryan

Hello Ryan,

Thanks a lot for the feedback.

> For Figure 14, you say that the area under the yellow curve is the current over a period of one second.

I said this based on my understanding from the EEVblog video where he said that the unit uVs displayed on the screen is the accumulated uV if the curve is spread over a period of one second(area under the curve spread over a second). But it seems you are correct as the area under the screen shown in Fig 14 is indeed over the entire span of the time on the screen and not just 1 sec. Few pointed out the mistake in the EEVBlog Fourm. 

> If you want to find the average current for a smaller section of time, you can use the oscilloscope cursors to find (BY-AY)/(BX-AX) of the violet line, but make sure that the horizontal cursor intersects the violet line at the same point as the vertical cursor.

I did this in Fig 15. which showed 11.6mV which matches with your calculated value of 10.75 uA.

> Texas Instruments has specialized equipment and procedures to test

Asking out of curiosity, Is there any more detail available on this?

> You can also compare your results to simulated values generated in EnergyTrace in Code Composer Studio.

I have used EnergyTrace extensively before assembling my uCurrent GOLD. The only downside of EnegyTrace is it cannot measure current down to the range of nA which is required while the MSP430 is in LPM. Otherwise, I depend upon EnergyTrace whenever I needed to measure the power consumption.

One feedback for EnegyTrace: It would be great if there is a way to disable/disconnect the debug pins from EnegryTrace software so that I don't have to disconnect and connect the debug pins when measuring LPM 3.5 or LPM 4. 

I will try to repeat the experiment again based on your feedback and update the post over the weekend.

Regards,

Vijay