RBC Model with Subsistence Consumption - Comparison between Non-Linear and Log-Linearized Model

Good afternoon, I am trying to replicate with a model log-linearized by hand the IRFs I get from a non-linear RBC model with subsistence consumption. The two models should generate the same IRFs but I don’t manage to make them overlap. In particular, the response of consumption to a technology shock is not as expected in the model linearized by hand. I double checked the log-linearization and it seems to be right. I really don’t know where I’m wrong. Attached you can find the codes I’m using. Thanks in advance and have a nice weekend!

RBC_linearized.mod (1.7 KB) RBC_nonlinear.mod (2.1 KB)

Did you cross-check with e.g. https://www3.nd.edu/~esims1/rbc_notes_2016.pdf