SMT32F042K6T7 temperature issue

[ArturIn90]

Hello, 

I am using STM32F042K6T7 microcontroller for doing some start-up configuration for some Ethernet PHYs by SMI, it is working well at room temperature, and cold (-40°C), but when I increase temperature max (+74°C) the microcontroller it seems is not able to start-up because the SMI signals (MDIO and MDC) seems not to work, I did some test with a Nucleo-F042K6 which has the STM32F042K6T6 heating up the microcontroller up to +74°C and it seems not having the same issue then I would like to make sure that my understanding is correct.

It suppose STM32F042K6T6 has a operating temperature of -40°C to 85°C and

STM32F042K6T7 has an operating temperature of -40°C to 105°C, then in my understanding this is more robust to high temperature, do you have any idea of what could be happening?

[AScha.3]

Right, the xxT7  should work up to 105° ;  but what about all other components at this ambient ?

Clock, crystal, PHY chip - all get drift or even problem at hi temp. Its not only the cpu there .

You have to check all of them , to find out ... i would heat up them separate (hot air), to get closer.

[ArturIn90]

Hi AScha.3, 

I isolate the MCU and I heating up with a hot gun at around 74°C (I also heating up PHYs separately to figure out where the problem come from) and the MCU stop sending messages at around 64°C in start-up condition. I tried with two scenarios and in two boards (besides Nucleo board), where I used a 16MHz crystal clock ECS-160-12-33Q-JES-TR and the internal RC oscillator, with both the behavior is the same, it just seem working good in nucleo board this is the schematic I am using in my PCB, if you get some idea or test I could make:

[AScha.3]

I ...would short the 1003 , 1k . Maybe osc has lower gain at hi temp.  Try..

[ArturIn90]

Comparing with the Nucleo-F042K6 in my schematic I didn't have the boot pin resistor, I just added a 10k resistors from PB8 to ground and it is working good, not sure why