Using crystal HSE configuration with external HSE source?

[DMårt]

Hi!

I have a STM32103 processor and I wonder if something bad can occurs if I use an external oscillator to OSC_IN and leave the OSC,_OUT floating. With a ceramic crystal HSE configuration in CubeMX.

I know that the BYPASS clock source is the best HSE configuration for an external source (as I have) because then the OCS_OUT is not oscilating. 

But for me, it won't make any difference because my OSC_OUT is floating and not connected to anything. Or will this break after few years of running?

I know that my OSC_OUT have the same oscillations as OSC_IN due to the "wrong" configuration. 

[mbarg.1]

You can leave any output open and nothing will happen, short term or long term.

BUT

remember that unused outputs with some signal inside can be a source of RF noise and that can limit RF compliance.

As a general law, better not to take out unused signals in new design.

[DMårt]

@mbarg.1 

Thank you!

Understand that this might cause bad RF. Luckily, I don't have any RF requirements, only functional requirements.

[Andrew Neil]

---

@DMårt wrote:

I don't have any RF requirements

---

That's not true!

All electronic products need to comply with the relevant EMC regulations wherever they're used/sold.

[mƎALLEm]

Hello,

---

@DMårt wrote:

I have a STM32103 processor and I wonder if something bad can occurs if I use an external oscillator to OSC_IN and leave the OSC,_OUT floating. With a ceramic crystal HSE configuration in CubeMX.

---

That statement is a bit ambiguous.

Are you using a crystal (XTAL) or a crystal oscillator (XO)?

If you are using a crystal oscillator (XO) that's the hardware configuration you should adopt in Bypass mode. OSC_IN connected to the oscillator output and OSC_OUT should kept floating. 

From the reference manual RM0008:

[DMårt]

Hi @mƎALLEm 

I'm using a XO with a ceramic configuration. But OSC_OUT is floating.

Yes, i know that I should adopt Bypass mode. But in this case it's ceramic/resonator mode.

I hope this won't cause any functional troubles.

[mƎALLEm]

---

@DMårt wrote:

Hi @mƎALLEm 

I hope this won't cause any functional troubles.

---

No this is what it was suggested in the RM as I stated in my previous post.