This paper presents a three-stage model (analysis, reconstruction and performance) for the conservation of interactive electroacoustic works for which the original technology is now obsolete or otherwise unavailable. The first stage, analysis, is a detailed documentation of the electronic processes and effects required for each work in a format that is independent of any specific device or system. The analysis provides a blueprint for future realisations using available technology. The second stage, reconstruction, provides a working performance resource, as well as a test case for the validity of the analysis. Reconstructed systems are then tested and refined through the third stage, performance. With repeated performances, compositions gain wider exposure and may be evaluated by listeners on their musical merits. To date, the author has analysed, reconstructed and performed several works for clarinet and interactive electronics. Each performance has informed the continued development of the newly reconstructed system, and has in some cases led to corrections to the underlying analysis. As a classically trained clarinettist and computer musician, the author's approach to the conservation of electroacoustic repertoire comes from a desire to find performable works and to keep them viable and accessible for as long as possible. Four works for clarinet and interactive electronics (by Musgrave, Pennycook, Kramer, and Lippe) are presented as test cases for this model.