Problem in modeling a simple heat loss through convection

Hi mates,
I'm trying to make a very simple model for a pin losing heat through free convection to an environment. To account to for the buoyancy force, I have used the term "alpha*g*rho_air*(T-Tamb)" (taken from a sample by FEATool).
For boundary conditions, I have used Temperature and Pressure for the vertical wall against warm plate.
For the top part I have used "Natural outflow" with "convective heat flux" and for the bottom part I have tried "convective heat flux" and "Insulation" with "Wall no slip" and "Natural out flow".
Neither of these conditions give me a stable answer. Anybody has any idea what's going wrong here? I have uploaded the file here as well. pin_2D.fea

Thanks for your help beforehand.

PS: In the samples of FEATool for convection inside a box, I replaced the above-mentioned term with Pr.Ra for the Fy and it totally changed the results and velocity profile, but at least the model could converge.

Unless you are absolutely sure, the solution should (easily) converge to a stable solution, the answer usually is that it is a difficult non-linear problem, possibly also that no stable solution exists. The workflow for highly non-linear problems is usually:

- Increase the non-linear relaxation parameters and maximum number of non-linear steps in the solver settings
- Perform multiple restarted solver steps, while in each step restarting using the old/previous solution and increasing the parameters leading to the non-linearity little by little
- Switch to a time dependent solver