Skip to main content
Visitor II
June 27, 2025
Solved

Calculating Rext after increasing LSE Drive Level on STM32H7

  • June 27, 2025
  • 2 replies
  • 651 views

Hello,

A design I am working on utilizes a LSE crystal with the following parameters:

f = 32.768kHz
CL = 12.5pF
Co = 0.9pF
ESR = 50kR Max
DL = 1uW MAX

I am calculating gmcrit ~ 1.52uA/V.  According to table 5 in AN2867, the drive level should be set to Medium High, will REXT required?  The design is already produced without REXT so adding it is not advised if avoidable.

I don't have the instrumentation to measure the actual drive level as it requires ~0.1pF probes for proper measurement.  There is no indication of LSE max drive current for the H7, the datasheet only specifies 550nA IDD at medium high.

Should I be concerned that Medium High may violate 1uW?

Thanks

    This topic has been closed for replies.
    Best answer by STOne-32

    Dear @jmh299 ,

    You do not need to have Rext in your design and I do not recommend for all of our LSE oscillator on all STM32 series alike . It is an optional for High speed HSE oscillators . I never saw a design able to exceed the drive level for our 32,768KHz over billions of MCUs in full production since Years . Indeed doing that measurements is tricky and should have a very special Probe to not compromise the oscillation .

    Hope it helps you.

    Ciao

    STOne-32

    2 replies

    Technical Moderator
    June 27, 2025

    Hello @jmh299 and welcome to the ST community,

    Rext is required only if If DL > DLcrystal.

    Please refer to the application note AN2867 "Guidelines for oscillator design on STM8AF/AL/S and STM32 MCUs/MPUs":

    mALLEm_0-1751035017495.png

    Please also refer to the section 3.5 Drive level and external resistor calculation in the same application note

    Hope that helps.

     

    jmh299Author
    Visitor II
    June 27, 2025

    I have read AN2867, unfortunately I do not have the instrumentation to perform these measurements accurately.  I was not sure if there was any rule of thumb regarding the drive level setting and the drive current.

    Thanks

    Technical Moderator
    June 27, 2025

    You need at least to measure VPP with an oscilloscope, otherwise you cannot continue:

    mALLEm_1-1751035925182.png

     

     

    jmh299Author
    Visitor II
    June 27, 2025

    Just as an update, I measured with my lab equipment albeit not 0.1pF-1pF per suggested as we don't have that kind of instrumentation.  
    Used a standard 18.5-22.5pF passive probe.

    Measured 224mV Pk-Pk on Medium High.

    If using the combined capacitance of the passive probe(18.5-22.5pF) and the oscilloscope (18pF) itself ~0.0373uW seems awfully low and well within spec but not sure how much my measurement method is affecting this.

    Thanks

    STOne-32Answer
    Technical Moderator
    June 27, 2025

    Dear @jmh299 ,

    You do not need to have Rext in your design and I do not recommend for all of our LSE oscillator on all STM32 series alike . It is an optional for High speed HSE oscillators . I never saw a design able to exceed the drive level for our 32,768KHz over billions of MCUs in full production since Years . Indeed doing that measurements is tricky and should have a very special Probe to not compromise the oscillation .

    Hope it helps you.

    Ciao

    STOne-32