MSP430FR6922: Intermittent High-Power Consumption During LPM3

Apologies. I made a typo in the initial post. The configuration for the debug in J1 (VCC TARGET), not J2. Power is being applied externally, not from the MSP-FET programmer.

But my main question still applies. Sorry for the confusion.

Thanks,
Brian

Does your current measurement cover just the MSP430 or does it include the other devices attached to it?

Something to watch out for is the clock request system. If one of the MSP430 peripherals is using/requesting SMCLK, then instead of LPM3 you will get LPM1.

Hi David,

Yes, the custom hardware includes other devices in addition to the MSP430. I'm less inclined to think that there is leakage on these devices since I know that the hardware is capable of achieving down to ~0.6uA with the same I/O configuration. (I put the I/O into the same configuration every time the device goes to sleep; theoretically, everything should be disabled except the MSP430.)

Yes, I am also using the SMCLK for a few peripherals: SPI, I2C. Do I need to disable each peripheral and the timer specifically? Or is disabling peripherals and timers handled in this call automatically?

__bis_SR_register(LPM3_bits | GIE);

Getting stuck in LPM1 makes sense to me. I was a bit confused because the EnergyTrace output seemed to indicate that the 430 was in LPM3 and the SMCLK was disabled. I will investigate this a little more.

Thanks,
Brian

Setting a low power mode does not turn off peripherals. That would kind of undo the clock request system.

Serial ports usually power down when not actively clocking serial data.

Be careful about those other devices. It is not uncommon for the clamping diodes on CMOS I/O pins to provide power to a device when Vcc is disconnected. Making sure that any such pins driven by the MSP430 are in a low state would be required.

Brian,

Where you able to resolve this issue? 

Regards,

Evan

Hi Evan,

Sorry, I've been a bit busy and haven't had much time to dedicate to this issue but I'm still working on this same issue. The nets seem to be pulled high or low correctly. I also tried disabling the clock request system (clearing xCLKREGEN) and SMCLK (setting SMCLKOFF) before LPM3 but this is still unsuccessful.

Thanks,
Brian

Hi Brian,

Can you clarify what you mean by "debugger detached" in your original post? If the EEM (debug) module on the MSP430 is engaged in any way it will consume current.

Regards,

Evan

Hi Evan,

I have been using the debugger to check if the MSP430 thinks it is in LPM3. Then I disconnect the debugging header and re-measure the current.

Thanks,
Brian

Hi Brian,

Thanks for the response. I have a couple questions about your test setup:

• You are measuring the current using multimeter not the EnergyTrace GUI, correct? 
• The device initially draws ~0.6uA but after an extend period (i.e. overnight) the device draws 330uA, correct?
  • During the above scenario is the debugger ever connected to the device? If so can you try monitoring the power overnight without the debugger ever connected?

To proceed I think it'd be a good idea figure out if the debugging hardware is contributing to the power draw. I suspect it is not, but it will help to eliminate.

Regards,

Evan

Hi Evan,

• I am using an Otti to measure the current -- Otii Arc (qoitech.com). I double-checked the current draw with a multimeter and the current in the Otii seems to match
• Your understanding of the issue is correct. The debugger is not connected overnight.

I may have a lead on the issue. I want to confirm this and then I can reach out with more information.

Thanks,
Brian

Hi Evan,

This is resolved now. There was a mixup with the boards. The confusion was that there were actually 2 different PCBs with different components populated, rather than one PCB with a change in power draw overnight. Sorry for the confusion. Thanks for all the help.

Thanks,
Brian

No problem, glad it is resolved.

Evan