Leakage current potentially via STLINK pins on demo board?

[bigjosh]

I am trying to make some very low current (<1uA) power measurements on the MCU on the STM32U083C-DK demo board.

I see that jumper `JP7` is provided specifically for this purpose, but I am concerned that connections from the MCU to the STLINK interface might leak current and throw off these measurements- especially when testing with MCU Vcc voltages that differ from the STLINK supply voltage. 

 

Is this a valid concern, or are those connections equivalent to air-gapped when the MCU is normally running?

Thanks!

[Tesla DeLorean]

>>Is this a valid concern...

Probably, perhaps also look at the CAD files, see if there's anything that can export a Net List.

You might be able to mitigate somewhat by switching the pins here into an analogue input / tri-state. I'd suggest initiating that as a console/monitor command to the system, rather than at start-up, such that recovery and reprogramming is kept workable rather than immediately thwarted.

[BarryWhit]

> Is this a valid concern

2.3 "Ensuring accurate power measurement"

 

Trying to measure current consumption in the nA low-power modes of ST is quite a specialized thing. I suggest you open an online support ticket on ST OLS to get help directly from them.

 

Alternatively, ST offers the ST-LINK-V3PWR  specifically for doing these sorts of power measurements. Perhaps that would save you from the thorny job of chasing down multiple and subtle error sources, which would be required for such a difficult measurement.

 

4-wire.

[bigjosh]

I have a Joulescope for making the Actual measurements, the line "Always remove the debug probe and power cycle before measuring current." from the linked document is critical since there does not seem to be a way to disconnect the debugger from the MCU on this demo board. 

[Andrew Neil]

---

@bigjosh wrote:

Is this a valid concern

---

Yes - absolutely!

You really need to scour the schematics for all possible leakage paths - the documentation doesn't necessarily cover them all.

Bitter experience:

https://community.st.com/t5/stm32-mcus-wireless/excess-current-consumption-running-b-l072z-lrwan1-from-battery/td-p/322961

 

Addendum:

This is a reason I dislike such boards crammed with "extras" - too often they just get in the way!

There seem to be a couple of Nucleo boards for this processor:

https://www.st.com/en/microcontrollers-microprocessors/stm32u083mc.html#tools-software 

https://www.st.com/en/evaluation-tools/nucleo-u031r8.html

https://www.st.com/en/evaluation-tools/nucleo-u083rc.html

Unfortunately don't seem to be ones where you can break off the ST-Link, but at least they have less stuff on them!

[Tesla DeLorean]

I remember the LRWAN1 issue, still use those boards, typically ditching the battery holder. Most recently replacing with a charge-and-run Lithium-Ion board from AdaFruit as a UPS, earlier rev of this  https://www.adafruit.com/product/390

 

Yeah, DISCO boards often too busy, and either removing parts, or finding a better proxy for development/testing.

[BarryWhit]

@bigjosh , once you've done your measurement, please post a follow up describing your process. Whether there's really nothing more to it than that, or you encounter any additional pitfalls not already covered, it's valuable knowledge and I'd like to learn from your experience doing this. 

[Andrew Neil]

---

@BarryWhit wrote:

@bigjosh , once you've done your measurement, please post a follow up describing your process.

---

+1 to that! :thumbs_up:

I have had to go back to my own post (linked above) about the LRWAN1 to remind myself what needs to be done!

[BarryWhit]

I don't think anyone really expects the ST-Link section to talk to the mainboard MCU by telepathy.

 

But It is unclear what the point of the slots are. Are customers just supposed to just snap off the embedded ST-LINK? That's definitely one way to air-gap.

[Andrew Neil]

---

@BarryWhit wrote:

Are customers just supposed to just snap off the embedded ST-LINK? That's definitely one way to air-gap.

---

yes, that is exactly the point!

https://community.st.com/t5/stm32-mcus-products/stm32-mcu-choosing-and-programming/m-p/657536/highlight/true#M239953

 

[BarryWhit]

> the path by which they do so should be manifest rather than hidden.

If the ST user manual instructs you to snap off the complementary ST-LINK module and reconnect it with a harness to the board, that's pretty "manifest". it's also easy to "visually verify". etc. You like the way TI does it? you're free to do so. 

But, you suggested ST didn't even consider the need to completely and provably disconnect the embedded ST-LINK from the rest of the system, and I showed you that you are wrong about that. So now, you dislike the way they do address this need - and that's also your prerogative. I suggest we leave it at that.

[Andrew Neil]

---

@BarryWhit wrote:

ST didn't even consider the need to totally and provably disconnect the embedded ST-LINK from the rest of the system

---

It's possible on the Nucleos with the break-off ST-Link, but not on other boards.

The OP's STM32U083C-DK doesn't have the break-off ST-Link.

The LRWAN1 in another without the break-off ST-Link, and it has some obscure & undocumented leakage paths.