@M11, this looks like the trial ends before the sound is played. As your sound is so short (less than a screen refresh), is synced with the screen refresh, and it attempts to play at the end of the trial, the trial ends before the sound is played, perhaps due to latency issues. The easiest solution is to move the sounds to the beginning of the routine instead of the end. Or, if you want the sounds at the end of the routine, add a keyboard component that ensures that the trial duration is set to allow both sounds to play (keyboard with a duration of 1.1 seconds), you should get your desired effect e.g., two short sounds in quick succession before the end of the trial.
If you want superfast sound latency, then I would await the new PsychToolBox sound backend. In the meantime, we performed some sound latency tests with current sound libraries and reported them here, if you want to get an idea of how they perform currently. See PsychoPy 2 Sound Latency - an update.