<p>In this paper, I argue that none of three different ways one can cash out the brain time theory provides a satisfactory general theory of how we temporally bind sensory features into perceptions of coherent perceptions of synchrony. In general, brain time theories state that the time at which sensory features are perceived as occurring relative to each other is fully determined by the time at which sensory features finish processing, i.e., reach their perceptual endpoints. I describe three different versions of brain time theories that differ in terms of how perceptual endpoints are defined. One defines perceptual endpoints in terms of local linear processing, another in terms of local linear and non-linear processing, and a third adds the possibility of perceptual endpoints determined by global non-linear processing. I evaluate each of these relative to how well they can account for two problems that I argue any theory of temporal binding must account for. These are the problems of neural delay and desynchronisation. The first problem concerns accounting for how our perceptual system overcomes delays in neural processing, such that we can perceive events in the world in a timely manner to guide judgments and actions. The second problem concerns accounting for how our perceptual system manages to temporally bind features, such that we can perceptually experience the synchrony relations and temporal order of multimodal events when our brain processes different features at different speeds. I argue that no version of the brain time theory provides a satisfying account of both problems. I end by proposing a way forward.</p>