When the system is not powered on, the power supply is still generating Standby current to the board. Thus there is some power there. How this might be routed to the internal speaker is the question.
My initial thought is that the DP cable may be responsible; perhaps causing a short of some kind. Before you ask, it is possible that the differences in board routing could cause this to manifest on one and not another - yet neither is responsible for the issue - or should be considered broken. I would try replacing it with a better quality cable to see if this makes a difference.
Hope this helps,
Wow, that was fast! Yes, i tried another cable sadly with no change. Oddly enough, the same cable with the boards onboard DisplayPort does not produce the sound. Only when the DisplayPort of the graficscard is used the noise is generated.
The other board i tried was an Asus H87M-PLUS - this did not produce the noise.
Hhmmm, not sure what else to suggest for eliminating. I will ask some of the design engineers. In the meantime, I can suggest a workaround (or make you smile anyway)...
While we were testing and validating the board design, we needed to avoid (often generated on purpose) speaker noise bothering the other developers around us (let me tell you, there's nothing worse than a cranky BIOS engineer!). What we found worked was to put a piece of electrical tape on the top of the piezo speaker.