To test if NDS clients after tomorrow will correctly apply the upcoming leap second, I injected a sub-second signal into a DAQ DQ channel today. The H1:PEM-EX_ADC_0_08_OUT_DQ channel is a PEM test channel with ADC bit noise on its input. I ran awgstream to inject a signal which is one second long, and all zero except for 1/16th of a second starting at the 0.5S mark. This 1/16th second block is a step function, with an amplitude of 1000 counts.

The time of injection is chosen to be modulo 64 in GPS time, so it will be the first second of a 64second DAQ frame. The time of injection is

The model runs at 16kHz, this channel is acquired to the DAQ at 2048Hz. The decimation filter is producing the ringing seen at the step transitions.