MCU Oscillator malfunctions on 0 degree temperature?

[cartman007]

Hey guys.

I recently found that a MCU's built in oscillator failed whenever the device went anything under 0 degrees "Celsius"

The Osc worked in LOW power mode and it was at 20Mhz.
When we divided the clock with anything above 2 it seemed to work fine.
Now this is weird since if you divide the clock it doesn't necessarily mean that the oscillator is going any slower...it just gets divided before being referenced to the CPU?

Might it be the CPU being hampered in some way by the COLD?

Why would that happen?

How does COLD weather affect the electronics or functionality of an Oscillator?

 

[D.A.(Tony)Stewart]

Long ago epoxy plastic seals were not reliable to moisture ingress below 0C. If this product is rated below 0'C and fails it could be defective. Normally CMOS improves bandwidth when cold. ( ie lower delays, faster transition rates)

If there is some DC leakage on the input side of the internal inverter that changes near 0'C or worse yet, delamination or junction damage, then you may have a defect , perhaps from handling if these pins are also external.

 

[FvM]

Are you sure that you are operating the processor inside the specified voltage and clock frequency area?

 

[cartman007]

Are you sure that you are operating the processor inside the specified voltage and clock frequency area?

Yes well its an internal Osc soo overclocking it will be a bit hard .

Is it possible though?

 

[dick_freebird]

It may be that you have a simple decoupling problem
(and/or too much capacitive load on outputs) such that
when the output drivers' current-drive stiffens up at low
temp, you bang the ground and/or supply rail to the
point that inputs are spuriously toggled. See "simultaneous
switching" in wide bus drivers.