Yes, that is what got me thinking. THAT potentially causes rail tieing issue right there if power is taken from both the ATX and the HDD power as those ARE different rails in a split rail system.
Also, different HDD leads (on different wires) can be on different rails also. So if that is OK somehow (avoids the differential regulation issue) , then why is another tieing situation NOT ok? I am not convinced this has been thought through all the way and either tieing IS ok, or there needs to be tighter restrictions on the ATX used to AVOID tieing and the "bad things"!! If tieing is bad, then split rail power supplies cannot be used, but if that is ok then tieing should be OK other ways as well. I'm hoping the later is correct but right now the former looks to be the rule and some people may have the wrong PS and are risking issues by unintentionally tieing the 12V rails via the HDD power cables.
RJ am I wrong and we need to warn people away from split rail supplies, or was I right and you are tieing via the HDD connectors but did not realize it?
<EDIT: I am wrong as pointed out below (it is in the WIKI now) but we do need to be sure everyone avoids split 12v rail ATX power supplies- I am thinking of taking pictures of labels of some of the many ATX power supplies to which I have access as a guide (this is goood, this is baaaad) if that is OK with everyone.>