Transcranial direct current stimulation (tDCS) is a non‐invasive brain stimulation method that is frequently used to study cortical excitability changes and their impact on cognitive functions in humans. While most stimulators are capable of operating in double‐blind mode, the amount of discomfort experienced during tDCS may break blinding. Therefore, specifically designed sham stimulation protocols are being used. The “fade‐in, short‐stimulation, fade‐out” (FSF) protocol has been used in hundreds of studies and is commonly believed to be indistinguishable from real stimulation applied at 1 mA for 20 min. We analysed subjective reports of 192 volunteers, who either received real tDCS (n = 96) or FSF tDCS (n = 96). Participants reported more discomfort for real tDCS and correctly guessed the condition above chance‐level. These findings indicate that FSF does not ensure complete blinding and that better active sham protocols are needed. We performed a secondary analysis of our high‐powered (N = 192), pre‐registered, multi‐site replication attempt to study the blinding efficacy of the “fade‐in, short‐stimulation, fade‐out” active sham protocol of transcranial direct current stimulation (tDCS). Bayesian data analysis of subjective ratings revealed that blinding is compromised for 20 min of 1 mA tDCS in healthy, young adults. Our findings emphasize that better active sham protocols are needed for this commonly used protocol.