STM32F417VGT6 - SWD - Cant connect to target

[firmen]

Posted on July 17, 2015 at 17:44

Hello,

After my first steps with the discovery board, I designed my very first PCB with a STM32 MCU (STM32F417VGT6), which I intend to program via SWD (SWDIO, SWCLK, GND, 3.3V). 

Unfortunately I don't get a connection to the MCU using the STM32 ST Link Utility via my SWD programmer. Here is the screenshot of the error:

https://dl.dropboxusercontent.com/u/10448441/STM32_ConnectionError.png

I also attached you my schematic of the basic STM32 and SWD layout, which I made:

https://dl.dropboxusercontent.com/u/10448441/STM32F417_SWD_Connection.png

Can you check my layout? Did I forgot something?

With best regards

Paul

[nesrine]

Posted on July 17, 2015 at 19:00

Hi,

I suggest you to connect PDR_ON to VDD if your VDD supply = > 1.8V and to VSS if your VDD <1.8V

-Syrine-

[Tesla DeLorean]

Posted on July 17, 2015 at 19:52

Put a pull-down on BOOT0

Check the voltage on the VCAP pins.

Check the state of NRST.

Ideally have NRST on you debug header.

Double check part orientation.

The other signs-of-life method is via USART1 (PA9/PA10) or USART3, with BOOT0 pulled high. Send an 0x7F character at 9600 8E1, look to see an 0x79 response.

[firmen]

Posted on July 18, 2015 at 13:02

Hi,

thanks for you quick replies.

PDR_ON:

I should have mentioned, that I use the LQFP 100 Package. In the datasheet is mentioned, that PDR_ON is not available since Sept-2011... Here is the extract:

https://dl.dropboxusercontent.com/u/10448441/PDR_ON.png

Boot 0:

I did connect a cable to the Boot0 pin and pull down (10k), and tried again, but I get still the same error.

I think nRst is not needed, when I connect my programmer with the discovery/nucleo board, I did not connect the nrst line of the programmer to the board either.

So next thing to do is to check the voltage on VCAP?

Do you know if a basic standard layout/schematic exists (maybe an application note) for the STM32 family with just power supply + SWD programming?

[Tesla DeLorean]

Posted on July 18, 2015 at 14:30

On paper it looks Ok, going to depend on your implementation and if you got all the pin definitions correct, and the parts soldered down and orientated correctly.

The Reference and Data Manuals do have all the required info. I can see the appeal of a minimal schematic, but as there are dozens of parts (package/family) it's something that gets hard to manage, be uniform/correct, and people will still cherry pick what the want to implement. Then there's minimal vs easy to debug/diagnose. It's quite easy to build something that's lifeless and dysfunctional.

[raptorhal2]

Posted on July 19, 2015 at 16:01

I think nRst is not needed, when I connect my programmer with the discovery/nucleo board, I did not connect the nrst line of the programmer to the board either.

Check your discovery/nucleo (whichever it is you have) board schematic in the user manual. The processor has NRST pulled high on the board through a 100K resister. The board's Reset pushbutton then pulls NRST low to reset the processor.

At the very least, you need to add a 100k pullup to NRST. The Discovery and Nucleo boards have a Reset button for tutorial purposes, your choice if you need to implement that also.

Cheers, Hal