The default heat/cool function has its own temperature function that differs from that in the equationmanager and uses a different convention. I have a different version with the "proper" convention, but that seems to break some of the test problems (i.e. supernova thin shell formation). I will investigate this more. If someone else is running into this issue, feel free to bother me and I will make it a priority!
This likely effects the fancier conduction models too!