H733: MCO output active without GPIO init

[LCE]

Heyho,

so I was just checking the MCO1 output of my STM32H733ZGT6 and found some super jittery 64 MHz with full swing between 0V and 3.3V.

Then checked my code and found that I had not set the RCC->CFGR and not even initialized the GPIO.

When I actually set it up it does what it should, but why is there such an output signal without the GPIO having been configured?
This could be some serious EMI source or cause some other trouble if the GPIO is not yet configured, but connected elsewhere (I only have a serial R and a testpoint on my PCB).

[LCE]

Now I put the GPIO init before any RCC / clock setup, setting the MCO pin to input with pull-up without any alternate function.

After power-up I still see this crappy 64 MHz (HSI) on that pin shortly - which makes sense, but is still terrible.

What's wrong here?

[LCE]

That a known issue, or what am I doing wrong?

ST please comment! 

@Peter BENSCH maybe?

[LCE]

Nothing from ST? :frowning_face:

[AScha.3]

Hi,

is this mco1 signal there, if holding reset ? (no program started at all)

I have a H733ZI here, so this i could try...if you want.

[LCE]

Hello @AScha.3 ,

thanks for your help.

It's actually occurring here if reset is released, the GPIO is not configured, and has no load.

So if you can find the time you could just grab a scope and check if you see anything on that pin.

It's a STM32H733ZGT6 in LQFP-144, pin 100 = PA8.

I just checked on my H723 Nucleo: no signal / clock on the H723-Nucleo MCO pin. As it should be.

[AScha.3]

Now -- found :  STM32H733ZGT6 , rev. Z ; boot0 : low : nothing on PA8 on power up .

DSO on single shot, just finds a small "pump" , about 1V, at power connection - nothing else.

btw. PA8 is configured as output in program.

Whats your H733 doing, if set it as output, low ?

ed

update : PA8 not configured, unused. power-up : nothing to see.

[waclawek.jan]

> Then checked my code and found that I had not set the RCC->CFGR and not even initialized the GPIO.

And what's the content of those registers, when the oscillations are ongoing? 

JW

[LCE]

It doesn't actually matter how the registers are configured, at least I haven't found some "MCO off" bit, there's only the source selection (default HSI  = 64MHz) and divider settings (1 = default).

At least I would have expected the same behavior on the H723 Nucleo in the same package (and 99% same source code).

@AScha.3 thanks for testing!

unconfigured GPIO: full 64 MHz (HSI)

configured GPIO (no matter in or out): nothing

So it's only a little bit of noise at start-up until the GPIO is configured - which is still something I don't like!

[AScha.3]

So there remains only : (because "my" H733ZG shows nothing strange) chip version ..? rev. Z here.

[waclawek.jan]

> It doesn't actually matter how the registers are configured, at least I haven't found some "MCO off" bit, there's only the source selection (default HSI = 64MHz) and divider settings (1 = default).

 

Just read them out (both RCC_CFGR and GPIO registers) and post.

JW

[LCE]

Hmm, I can only read RCC_CFGR = 0, GPIO clock is not even enabled...

[waclawek.jan]

> GPIO clock is not even enabled...

And if you enable it, what are the GPIO registers' values?

---

Yes, I do have a suspicion, that this is consequence of a bootloader, either default or your own. Either one might have set various registers and then incompletely restored them.

JW

 

[LCE]

I was just about to check the GPIO, when I found one of the most embarrassing mistakes I made (here): 

The bootloader (which is active for only a few ms) had the MCO configured. Ouch.

Sorry that I have wasted your time! :frowning_face: