A model of valvetrain equipped with a direct acting hydraulic lash adjuster (HLA) is presented. As it is the cause of many problems, a possible aeration is taken into account. The main model is made of three submodels: 1. The first one is a mechanical one with masses, springs and dampers and needs the cam laws and the pressure of the chamber of the HLA as data. 2. The second submodel is an hydraulic one. It studies the HLA chamber through the check valve flows, the leakage flow and the mechanical strains on the chamber oil. This submodel gives the chamber pressure to the first submodel. 3. The third submodel is a physical one and evaluates the instantaneous amount of air present in the chamber oil. It takes into account the arrival of new bubbles through the check valve and the mass exchanges (diffusion, dissolution) between the bubbles and the oil bulk. This amount of air has a direct influence on the chamber pressure calculation made in the second submodel. Experimental validation of the model was conducted on an engine head. This engine head was feeded by an oil circuit in which the aeration could be controlled. Air was injected under pressure and oil aeration was measured in the HLA oil supply by an original sensor, the SMAC. Three sensors were placed in the valvetrain: 1. Acceleration on the valve head 2. Proximity on the valve seat 3. Force between valve and HLA The model correctly predicted measurements on the valve train. The mechanical modelling and the experiments show and explain the problems of valvetrains when air is present in the HLA. The hydraulic modelling allows a better understanding of the behaviour of the HLA and especially of the check valve. The physical modelling and the experiments prove that the aeration in the HLA chamber is not only influenced by the aeration in the circuit and indicated some ways of controlling aeration.