LP-EM-CC2340R5: Energy Trace battery life calculation appears incorrect

Hi,

Agreed that something seems off. Based on the current alone (average of 4.8mA), I would not expect those numbers. Can you share the SDK version and IDE version you are using for your testing? Also, are these power numbers being generated by a custom project you have created or an off-the-shelf example from the SDK?

Best Regards,

Jan

Hello Rob,

Please confirm that they XDS110 is only connecting the RST, VCC, and GND lines to your target.  The SWD IO and CK lines should not be connected as an active debug session would draw additional milliamps of current.  Can you also please confirm your CCS Preferences -> Code Composer Studio -> Advanced Tools -> EnergyTrace Technology settings?  Although written for a separate SimpleLink product family (F2 devices), here is an Application Report you can reference: https://www.ti.com/lit/swra478 

Regards,
Ryan

Hi Jan,

Here are the answers to your questions:

Additionally, I'm using simplelink low power f3 SDK version 8.40.00.61.

The numbers are from a custom project that was derived from the data stream example project.

Hi Ryan,

In this capture, the XDS110 is connected directly to the LP-EM-CC2340R5 development board. There's no way for me to run EnergyTrace without debugging. I'm using CCS version 20.0.0.12 and unless I'm in a debug session the EnergyTrace (Battery Icon) is always disabled regardless of XDS110 connection status.

I don't have an Advanced Tools or EnergyTrace CCS preference option.

Hi Rob,

Thank you for the additional details and I apologize for the late reply.  I will loop in the CCS and EnergyTrace developers to further comment on the battery life calculations.  I have noticed that the CC23XX family devices is not listed in CCS v20 EnergyTrace documentation.  I would also warn that since a debug session is required, it may not be feasible to achieve the lowest power consumption possible given the JTAG connection.  You may consider downgrading to CCS v12.8 solely for EnergyTrace purposes, or use bench equipment (power analyzer, DC power supply, or precision multimeter) to measure your current consumption.

Regards,
Ryan

Rob Oppen said:

However, EnergyTrace is estimating the battery life to be 1730 days, 42 hours, and 10 minutes! Why are these two numbers so far off?

This is definitely a bug. ET is incorrectly using µJ vs mJ in the calculation, hence off by 1000x. I will file a bug for this.

Rob Oppen said:

Converting the battery to Energy (E=C*V): 220 *3600*3.3V = 2,613,600mJ

Also not that by spec, the battery is rated for 3V, not 3.3, so:

220 *3600*3V = 2,376,000mJ