Studies of change detection have shown that changing the task-irrelevant features of remembered objects impairs change detection for task-relevant features, a phenomenon known as the irrelevant change effect. Although this effect is pronounced at short study-test intervals, it is eliminated at longer delays. This has prompted the proposal that although all features of attended objects are initially stored together in visual working memory (VWM), top-down control can be used to suppress task-irrelevant features over time. The present study reports the results of three experiments aimed at testing the top-down suppression hypothesis. Experiments 1 and 2 tested whether the magnitude or time course of the irrelevant change effect was affected by the concurrent performance of a demanding executive load task (counting backwards by threes). Contrary to the top-down suppression view, the decreased availability of executive resources did not prolong the duration of the irrelevant change effect in either experiment, as would be expected if these resources were necessary to actively suppress task-irrelevant features. Experiment 3 showed that a visual pattern mask eliminates the irrelevant change effect and suggests that the source of the effect may lie in the use a high-resolution, sensory memory representation to match the memory and test displays when no task-irrelevant feature changes are present. These results suggest that the dissipation of the irrelevant change effect over time likely does not depend on the use of top-down control and raises questions about what can be inferred about the nature of storage in VWM from studies of this effect.