I use 100Khz oscillator for LSE on STM32H7B0VBT6 but LSERDY cannot work

[kgai]

100khz Oscillator for LSE input. The OSCillator works right bcz I test the waveform by scope.

but stm32h7b0vb cannot get lserdy bit set.

Seems I m the 1st Person meet the situation.

plz help me. Thanks.

[waclawek.jan]

Oscillator or crystal? How is LSEBYP set?

@Uwe Bonnes​ has reported problems with LSEBYP/LSERDY in the past, I don't know whether and how this relates to the 'H7 too.

Please note that this is a primarily user-driven forum, so we don't have insider information. You may want to contact ST directly, through FAE or web contact form. If you do so, please come back with the findings.

JW

[kgai]

Thank u first!

it is sure Oscillator , Not crystal. I set the bypass

I guess it is bcz ST not test LSE by freq other than 32768hz.

I also try​ call the function LL_RCC_LSE_SelectDigitalClock() but no effect

I will check the link u mentioned from now. hope I can find some points.​

[waclawek.jan]

The primary purpose of LSERDY is to make sure the crystal starts to oscillate before proceeding with setting up other things. If you are sure your external oscillator is oscillating, you can simply avoid testing LSERDY.

JW

[kgai]

Thank u anyway!

but things are not that simple. Look at the pic below. I will explain it.

the "ready logic" module inside the red circle is very important, bcz follow it there is an "AND" gate, while the "ready logic" or "LSERDY Bit" output 0 then we will get CKLSE out always 0.

so we cannot ignore the Ready bit.

[kgai]

who can help me? IS this the bug of stm32?

[kgai]

I fix the problem. THANKS

the Reason is I must call LL_RCC_SelectDigitalClock after EnableBypass.

now work well.​

[kgai · Accepted solution]

no, the problem still exist.

new info: I change the Oscillator to 32.768Khz and it works well.

so conclusion is the STM32H7B0 LSE cannot use freq other than 32k768. the datasheet is wrong about it.