we are using STM32H755IIT6 MCU we made 5 boards with the same MCU STM32H755IIT6 all the board can be programable with verification but only one board of start to run what we can check on the board to find the problem

[SLeit.2]

..

[SLeit.2]

yes we are using SMPS

[Tesla DeLorean]

Like I said, there's little FAE participation here, they work directly with customer accounts in their own regions. I think we've established they are not going to respond here, so you'll need to use more classical contact methods. Buying from distributors, shake their tree.

Board bring-up tends to be the primary responsibility of the engineering team who designed it. And might require significant due-dillegence work to go over the design again, pin assignment, PCB, PCBA, netlists, etc.

So if understand the current state of things you can program and verify all the boards. This suggest code is running, but not necessarily well, and it likely either running from ROM, or stuck in some while(1) loop. The boards are at least semi functional.

The trick now would be to determine if the code is executing, and how deep it's getting.

The parts will run out of reset, you don't need to set any clocks or plls.

I'd start by adding code at the Reset Handler to set/toggle a GPIO, and then work deeper into the code flow to see how far I can still see signs of life.

Would do it in assembler, and I'd likely try to get a UART up next as a means of reporting status from within. The H7 should run at 64 MHz via HSI, so use that for baud rate computations.

[SLeit.2]

the problem is that we are using the same code for evaluation board and it is running well

we also have one out of 6 boards that working well all other dost work at all we just successes to program

[Tesla DeLorean]

Ok, but then you've got to figure out what's different in the design/implementation.

That the boards are programmable suggests you're not completely dead in the water.

You'll likely want to code something simple starting at the Reset_Handler, and establish some beach-head functionality from which you can inspect from within.

Perhaps some infinite loop of NOPs would allow for the debugger to connect initially ?

[GLASS]

You can also give a look to Option Bytes and check that STM default values (look at Reference Manual) are ​not altered on all boards. You can read Option Bytes with STM Cube Programmer or perhaps St Link utility ( not sure that this 'old' SW know H7)...

[LCE]

... and you have checked and measured every supply voltage and ground pin?

Checked the crystals / oscillators?

Reset?

If one board is running, maybe it's some hardware problem, bad caps, bad soldering?

IIT6 is LQFP-package? If yes, at least you can optically check and measure at each pin.

[LCE]

I would further try:

• connecting crystal / HSE
• using LDO instead of SMPS