STM32H735 custom board power draw

[Neutron235]

Hi,

Ive made a custom board with a STM32H735 micro on it. I have built a board with nothing but a micro and a few of its needed passive components and went to provide it power and test that its working. I supplied power from a benchtop power supply with a current limit of 400ma and it hit this limit at 3v. Dropping to 100ma it hit the limit at 2.2V. Checking the resistance on the power rails it 10.25k and the Vcap rail is 500k. Thermal camera shows a square heat zone expanding from the centre of the IC (LFQP package) towards the outside. Ive now tired 3 different ICs on 3 different boards. The schematic is very similar to the STM32H735G-DK  with changes mostly being things like jumpers. What could be causing this issue and what steps can i take from here?
Thanks in advanced

[mƎALLEm]

Hello @Neutron235 and welcome to the community,

Need to share your schematics even it's similar to the discovery board. So other users can help you efficiently.

Thank you

[Neutron235]

Here is my schematic. Thanks

 

[mƎALLEm]

---

@Neutron235 wrote:

Here is my schematic. Thanks

 

---

Sorry but this is a bad picture resolution. I suggest you to attach a pdf file instead of screenshot.

You can also refer to the AN5419 "Getting started with STM32H723/733, STM32H725/735 and STM32H730 Value
Line hardware development"

[Ozone]

I would check the voltages at the regulator and near the MCU with a scope, and a high temporal resolution - like 5...10us per increment.
In other words, the switch-on behavior.
Perhaps there is a layout issue.

[AScha.3]

Hi,

whats on VCAP : 2x 2u + R to V??  ? What voltage you measure here, on vcap ?

[Ozone]

By the way ...

> ... and went to provide it power and test that its working. 

Does this include trying to get access with a debug pod ?

In other words, does the MCU start up ?

 

[LCE]

Bad resolution... too much guessing.

Are all the SMPS components correctly installed?

Have you set the source code accordingly (SMPS / LDO)?

"benchtop power supply" - is there some on-board voltage regulator in-between?
I'd never do that without, because if one day I forget to switch back the external supply from 24V, then it's nighty night MCU... ;)

[Neutron235]

@Ozone 

With the scope set to 5us probing a cap on VDD im seeing a nice smooth line at about 2.1V (Lab power supply is at 100ma current limited and 2.3V)

I have tried with both a SWD port and holding boot0 low and connecting to a USB. I havent ever programmed this chip so i believe it should show up as a DFU device. The micro is obviously doing something because it dosnt draw much power until ~2v

@mƎALLEm 

Ive attached a PDF of the micro schematic. I have looked through and followed that application note

 

@AScha.3 

There is a zero ohm connected to VFBSMPS, its currently not populated. The datasheet and application note say that the internal SMPS can either power the core directly or via the internal LDO. this zero ohm would allow me to make that connection externally.

 

@LCE 

This board dosnt have any however the last board i had has all of the components and i checked that all were well soldered. Both boards have the same off resistance and the same voltage and current power characteristics 
I havnt been able to connect to the micro to be able to set the SMPS or LDO up and cannot find the default config hence my zero ohms to power things individually.

No there currently isnt, i know this is risky and not really great practice. I accept the risk as im only working on this one board atm, its only me that touches the PSU and as soon as i get some sign of life from the micro ill build the entire board up and that has regulation onboard

 

 

Thanks all for your help

 

[Ozone]

> The micro is obviously doing something because it dosnt draw much power until ~2v

I'm not a hardware guy, but I think this is most probably the brown-out limit.
You would have to check the datasheet.

> I have tried with both a SWD port and holding boot0 low and connecting to a USB. I havent ever programmed this chip so i believe it should show up as a DFU device.

I would try to check if access works at all.
Read out the core ID with the debug pod, or something like that.

[Neutron235]

Yeah 1.62v is min operating technically.

Using STM32CubeProgrammer im getting :

18:29:43 : Error: Unable to get core ID
18:29:43 : Error: No STM32 target found! If your product embeds Debug Authentication, please perform a discovery using Debug Authentication

[LCE]

> There is a zero ohm connected to VFBSMPS, its currently not populated.
> The datasheet and application note say that the internal SMPS can either
> power the core directly or via the internal LDO. this zero ohm would
> allow me to make that connection externally.

 

Not so sure about this, the SMPS needs some feedback signal, and that's VFBSMPS.

So try with the 0R.

That might explain why you measure something close to input voltage, your "VDD" should be something around 1.1V (check datasheet).

This pic from AN5419 about H7 power (red notes by me):

 

[Neutron235]

So my first board tried number 1 and was shorting on the VDDSMPS pins.

The zero ohm there lets me put it into type 2, i have another that lets me go into type 3.

No matter this config the VFBSMPS is connected correctly