Recently I started a new personal project, where I'm trying to develop my own SMD-to-female-header PCB for the STM32F207VCT6 MCU chip. After ordering all the parts, doing some reflow soldering, checking that there are no solder bridges, and verifying that each pin properly reaches its corresponding header, I proceeded to connect the following circuit to it. Most of it comes from the AN3320 application note by ST for this chip. enter image description here

In order to program it, I got an FTDI FT232R board and connected it to the chip's USART1 (PA9 and PA10). Additionally, I set pins Boot0 and Boot1 to HIGH and LOW respectively to boot from system memory. From what I've seen from other blogs, a connection between STM32 chips and the STM32CubeProgrammer software can be done via USART. However, I'm getting the following error.

enter image description here

I tried setting the Baudrate and parity settings to every combination available, I tried resetting the chip multiple times, checked the appropriate voltages for VCC and VCap, and tried other USART ports on the chip. Am I doing something wrong?

  • \$\begingroup\$ VDDA floating. Nowhere near enough decoupling caps (should have at least one for every supply pin). And is there a reason you're driving VCAP with an opamp instead of letting the STM's internal voltage regulator do the job? \$\endgroup\$
    – brhans
    Commented Sep 25, 2023 at 3:14

1 Answer 1


You have 3 unconnected supply pins and you are using Vcap incorrectly. The AN3320 is quite clear on this how pins must be connected to be correct, and this does not look at all similar.

All supply pins must be properly connected. VDDA is (usually) responsible for generating reset, undervoltage monitoring, and powering other analogue parts such as crystal oscillators and PLLs, so with all those unpowered, it can't work.

You are also not disabling the internal core supply regulator, so the Vcap is an output from the internal regulator for external bypass caps. Even if you wanted to disable internal regulator and feed in core supply through Vcap pins, you need a regulator, as it is unlikely that an opamp can provide the amount of required current.

So with all the supplies unconnected and Vcap used incorrectly, don't expect that the MCU is undamaged and will work after fixing the surrounding circuit.

  • \$\begingroup\$ I do not think that he destroyed the MCU. Unconnecting VDDA is of course significant violation, but I do not think it destroy the chip. He need to fix it by soldering the tiny pin to VDD somehow. Opamp circuit he can just desolder and use C5 and C6 as a propper capacitors which should be connected to VCAP pins. He has lot of decoupling capacitors. Definitely do not match full best pracitces but still may work. I remember era when we were using much more power hungry 8-bits (running at similar frequencies as STM32) on breadboards without any decoupling caps and they worked. \$\endgroup\$
    – Misaz
    Commented Sep 25, 2023 at 22:55
  • \$\begingroup\$ @Misaz Why are you so sure the MCU is not damaged, do you know something that overrides the datasheet statements about all supply pins must be connected and there must be no more than 50mA difference betwee e.g. different power pins? I simply warned that if it does not work any more, it won't be a surprise, as it has been outside absolute maximum ratings. It might as well work just fine or appear to work fine even if it is mildly damaged. \$\endgroup\$
    – Justme
    Commented Sep 26, 2023 at 5:33
  • \$\begingroup\$ (1/2) It is not 50 mA but mV. Stress above absolute maximum ratings MAY cause permanent damage to the chip but also may not. Also permanent damage does not automaticaly mean nonfunctional chip. Some other fails are for example permanently increased power consumption (but otherwise working), changes of working voltage range (it can for example stop working at the lower boundary of VDD) and so on. Nobody can remotely say it is or it is not totaly dead. Similarly, if you connect 4.1V to the VDD(A) it is violation of absolute maximum ratings by 0.1V, but most chips survive such violation. \$\endgroup\$
    – Misaz
    Commented Sep 26, 2023 at 15:24
  • \$\begingroup\$ (2/2) His device currently do not work because of bad board design (unconnected VDDA), but after connecting VDDA it may work I think. \$\endgroup\$
    – Misaz
    Commented Sep 26, 2023 at 15:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.