Powering up a 5 volt tolerant external pull up.

[DJC]

Does anyone have experience with 5V pull ups? do you have to power on the 5V source after the MCU is started?

I'm running a STM32L431CBT6  with external 5 volt 10k pull-ups on the pins PB11,12,13,14,15 and PA1. (all flagged as FT pins)

I've been able to plug it in and power it on and even program the MCU. All of the pins above are in "reset_state". So, as far as I can tell, we are not using any internal pull-ups/downs.

After some time and, perhaps, consecutive plug-ins, the chip seems to burn out.

I've since cut the chip off the board (since its burnt) and replaced with with a 47ohm resistive load and taken a measurement of the startup voltages (see attached image).

The blue line is the 5V rail, which is powering the pull ups.

The red line is the 3.3v LDO powering the MCU.

There is clearly a short time where the GPIO pin goes above the MCU power, and there is also about 1.6ms between the MCU powering on and the GPIO pull-ups getting over 4V.

According to the documentation DS11453 (Rev3) page 86, one should be able to put 2V on a GPIO pin to an unpowered MCU, you just cant put VDD+4V.

In previous designs, we have had an MCU activated a switch that allows the 5V to these pins after the MCU is powered up. In this design we have taken the switch out. However, there are other changes in the design that I'm also looking into, so I'm not convinced that this is the actual problem.

Is anyone aware of a reference design that demonstrates 5volt tolerance.

[S.Ma]

Any pin as output ? Are those open drain mode?

[DJC]

Pin configurations are described in the original question.

As far as i know, open drain is what is needed to operate a 5vt pin. However, it should also work in reset_state as this equates to an open circuit.

would be good to know where you are going with this question.

[DJC]

I think that, based on you comments and my ongoing investigations this FT pin stuff was not an issue.

I have also since not be able to replicate it. Thanks for all your help.

Further to this case, I was also looking using a Zener diode to further protect this form happening, in this question, but I'm starting to think that it is not necessary since this LDO starts up so fast the up-stream buck converter never has time to get to the VDD+4V level. As such I can assume that I'm always safe.

But still, I'm very interested to hear what you think.