Skip to main content
Visitor II
April 10, 2025
Solved

STM32F407 VREF+ voltage not 3.3V

  • April 10, 2025
  • 5 replies
  • 836 views

Hi, I have a custom board in which VREF+ is supposed to be 3.3V. But the actual VREF+ (pin32) is 2.xV, with drifts between power ups. VDDA (pin 33) is 3.3V.

What's wrong with the circuit schematic? Can I just remove the resistor (R22) between VREF+ and VDDA? Thanks very much!

stm32_vref.jpg

    This topic has been closed for replies.
    Best answer by Carl_G

    @TDK noted. That 10K isn't just not helping, it's the opposite of having a cap on the input pin. Yes, you can just remove R22. C47 and C45 should be on AGND not GND. I've seen a lot of negative talk about C63 and splitting AGND from GND in general.

    5 replies

    Graduate II
    April 10, 2025

    Do you use the DAC. At least then you must consider I(VREF+), max 240 uA. 10k in the VREF+ line is no good idea.

    Graduate II
    April 10, 2025

    How have you configured the VREFBUF function so that Vref+ is an input? As an input I would expect a 100n to be on the pin side of that 10K resistor.

    Super User
    April 11, 2025

    The 10 kOhm isn't helping anything here and the best solution would be to short it.

    Per the datasheet, VREF+ should have two caps on it. Your design would probably work okay if these caps were present.

    TDK_0-1744334764499.png

     

     

    @Carl_G There is no VREFBUF on the STM32F4 family. It's always an input.

    Carl_GAnswer
    Graduate II
    April 11, 2025

    @TDK noted. That 10K isn't just not helping, it's the opposite of having a cap on the input pin. Yes, you can just remove R22. C47 and C45 should be on AGND not GND. I've seen a lot of negative talk about C63 and splitting AGND from GND in general.

    Graduate II
    April 14, 2025

    I'm thinking the 10uF is there because of the inductor and thus should probably be placed close to it. C47 near L4. C45 should go near the MCU. And as TDK pointed out you need one more cap by the MCU on VREF+. In this case I'm guessing the new cap should be 1uF.