FMAC control loop issue at lower temperatures

[empps]

Hi,

I am currently working on a synchronous buck converter (24 VDC in and 1-15 VDC, 1-15A out) for which I used the HRTIM of the STM32G474 in combination with the FMAC as a control loop:

I mainly followed AN5496 and AN5305 to get my converter up and running and it has been functioning seemingly fine for a few ten hours, until I started doing some tests in a climate chamber. At an ambient temperature of +70°C is seems to work fine and has been working at full load for a few hours, but in negative temperatures I start to run into an issue. At -10°C to -20°C my converter shuts down a few seconds after starting it up due to the over-current protection. After connecting a DSO it seems like the FMAC control loop outputs some very large values, causing the duty cycle to be 90% which is the maximum.

Ch1/yellow: current clamp

Ch2/green: current sensing circuit on PCB

Ch3/blue: switch node (between high side MOSFET and inductor)

Ch4/pink: 3.3V supply rail (I think)

Sidenote: because I am measuring in a climate chamber I needed to solder on some wires at different points to which I could connect the probes, hence the artefacts.

It seems to be a gradual process, for example at 0°C to -5°C I can see a 90% duty cyle during one or two switching cycles occasionally (which is not enough to set off the OCP) and the lower the temperature the more often switching cycles with 90% duty cycle seem to occur.

For completeness, my OCP consists of a shunt current sense resistor, precision op-amp of which to output is connected to a comparator on the STM32G4 MCU and when it reaches a certain threshold it triggers a hardware fault line of the HRTIM causing it to shut down.