STM32U575 chip short when JTAG Connected

[southernGB]

Hi all

I am trying to program an STM32U575RIT6 chip using a TAG connect and ST link V3, on plugging in the tag connect, the chip shorts and becomes unrecoverable. I make sure to ground the linker before inserting, doing it all on a grounded matt, with the power to the board off before plugging it in.

This seems to happen at random, as sometimes the programming will work and program fine, other times it shorts. I replace it with a new chip and it works for a while and then shorts again. When the chip is programmed, it works fine and as expected, until the point I come to reprogram it and it shorts.

I've tried other ST links, and cables, to verify its not just my linker and still get the same results. I've tried keeping the board powered on when I plug the STM link in, and still get the issue.

I've moved from using an STM32L452 and STM32L496 which I never had any problems with programming, only ever this U5 chip.

Any advice on what I am doing wrong, or how to stop this from happening.

I've added the schematic diagram for the ST link to my chip for reference.

Thanks in advance :)

[Sarra.S]

Hello @southernGB, 

The 10µF capacitor (C71) on the SWD lines might be causing issues, try reducing the capacitance or removing it temporarily to see if the issue persists? 

[southernGB]

Hi @Sarra.S ,

I've tried this method and the STM32U5 has still shorted out when the ST link is connected.

[Tesla DeLorean]

Then you'll need to review the connector, and what it connects too. Make sure you've not mirrored or confused the pin designations.

This isn't a standard 10/20-pin JTAG as exported by the ST-LINK's

Review your netlist. Review all pin assignments.

Review part orientation.

Check analogue supply.

Check voltage on VCAP pin

[Tesla DeLorean]

PC13 probably isn't suitable for driving things, it's typically in a low-power domain

Do not put an operating board on a conductive surface or foam.

Check unpopulated board vs net-list.

[TDK]

I suspect a ground loop might be the fault. How is the board powered, how are grounds isolated?

10 uF cap on NRST is way too high. Should be 0.1 uF. I'm not sure if this will cause issues. Is VDD decoupled? Is VDDA connected to VDD? Lots of details we can't see in the portion of the schematic you sent. Consistent issues with chips shorting suggest a protection diode is getting blown. Is VDD stable? Generated from an LDO?

[southernGB]

Hi guys, 

Thank you very much for your responses, @Tesla DeLorean I've checked the pin out on the JTAG connector and it matches the pins on the JTAG schematic so I don't suspect its that. With regards to the PC13, that a GPIO logic input of 3V to 0V, so nothing is being driven from it.

The power supply is a stable 9V, which is protected and then passed through a LDO to drop it to 3V3, I've attached the schematic for the protections and LDO implementation, and I have attached the schematic for the VDD and VDDA isolation @TDK. There may be an issue there, but I have followed the recommended layout from the data sheet.

I will try and reduce the cap on NRST to 0.1uF and see what effect that has.

  

[Andrew Neil]

---

@TDK wrote:

I suspect a ground loop might be the fault.

---

@southernGB To test, what happens if you connect your TAGConnect, but with only its GND connected?

Maybe also try with just its GND and VCC...

[southernGB]

To update this post, I have bought a new tag-connect cable and changed the cap C71 to be a 0.1uF to program and the U575 chip has programmed without fault. I've also hooked it up to a scope when programming and didn't see any strange behaviour. If this happens again I will update this post, but all seems well at this point in time. Thanks for your guys help on this.