Can RCC_LSCO be pulled to ground without issue?

[EC.3]

Hello,

We have the RCC_LSCO of the U5 connected to a peripheral that warns that it will pull this line to ground when the peripheral itself is reset. At the moment, the pull strength of this peripheral is unknown, but let's say it's fairly strong.

Does the U5 circuitry buffer the source oscillator that drives the RCC_LSCO output? I'd like to confirm that the peripheral's behavior will not impact the source oscillator, as it could be used for supporting timed wakeup of the U5.

Thanks 

[AScha.3]

Hi,

Why not doing it without any risk to damage the cpu ?

Use a buffer or most simple : a resistor in line, maybe 560 ohm , limiting the "short" current to save levels ?

btw:

[EC.3]

Hi @AScha.3,

Certainly we want to take the necessary steps to protect the CPU, but it'd be nice if that came for free as part of the U5 without having to pay additional cost and space for any extra components. Is it the case then that the CPU doesn't itself have protection for this scenario?

[AScha.3 · Accepted solution]

Hi @EC.3 ,

Come on - you take a U5 - STM32U535RET6 is cheapest at mouser, about 6$ - and worry about adding a ~ 1 ct part to be sure, it will not be damaged ?

And cpu pins are usually not designed for shorting them to VCC or GND . Except some automotive parts i seen, that state in ds : protected against short to adjacent pin, vcc or gnd.

Absolute maximum rating of U5 (from U575 ds ) is:

More will damage it.

[EC.3]

Hi @AScha.3 ,

I'd like to wrap back to my initial concern about impact to the oscillator. So let's say we have the 560ohm resistor in line, the peripheral pulls down the line to ground, and the CPU is protected. Does that also prevent any impact to the source oscillator's behavior?

[AScha.3]

You connect to LSCO, a io-pin driven from the mux - "far away" from the source oscillator , so no problem for the LSI or LSE oscillator.

But you should use the resistor, to keep the load to the pin in allowed range; so you create no "problem" for the U5xx cpu. So if other chip pulls the 560 r to gnd, the cpu pin just gets to drive about 6mA and this is no problem at all.