The semiconductor industry is under a lot of pressure due to major chip shortages. Downtime of lithography systems can cost semiconductor companies up to ten thousands of euros per hour (ASML Holding N.V., 2014). To improve their systems' reliability, ASML replaces old or frequently failing components with new and more reliable redesigns. In this research, we develop a tool that is able to determine the best policy to implement these redesigned components in the field. Seven distinct policies are considered that differ in (1) roll-out strategy, (2) whether or not to use old parts that are still on stock, and (3) whether or not to rework failed old parts that are built out of the machines. The roll-out strategy indicates whether parts are preventively replaced through a Field Change Order (FCO) or correctively through a Replace on Failure (ROF). Using a Markov Chain approach and a simulation approach, we test two models: one in which old parts can be reworked into the redesigned version, and one where this is not possible. Generally, the policies where the old parts can be reworked are preferred as they result in lowest costs and failures. ROF is preferred when the aim is to reduce costs, and FCO is preferred when the aim is to reduce the number of failures. When the focus is on reducing waste (circular economy) we recommend to rework parts, but not in combination with an FCO roll-out strategy as the parts on stock are then at high risk of becoming excess or obsolete. i Probability that an old part can be repaired s Base stock level w n Salvage value of new parts w o Salvage value of old parts z n Costs to repair a new part (including repair and normal shipment) z o Costs to rework an old part to the new version x