If randomization without replacement is more common, then you could just create a set of trials with varying location values for the stimuli and then randomize the order of those trials. Would that work for you?