I am really not the resident memory expert and this is just a shot in the dark, kinda one person's opinion. As an owner of G-Skill Ripjaws memory it is my thoughts that it is slightly over spec'd for the low voltage of 1.5V (that is their 1333 memory and higher) and then you have people like us looking for work arounds to get it to work reliably at all, let alone at the published specs of 9,9,9,24, 2T 1.5V and then further worry about bumping up the voltage one tenth of one volt DC to 1.6V and trying to decide what might be the long term consequences overall to their equipment.
Personally, If I had any real long term concerns I'd dumb down the frequency to 533Mhz (1066) and set the timings to 7,7,7,20, 2T and leave all voltages everywhere else on the board at default / normal. And see what happens.
I really don't suspect that what they are calling 1600MHz memory is truly that at all and might only ultimately come close to that performance if using either Intel's XMP (Extreme Memory Profile) or AMD's Black Edition equivalent (maybe called AMP) which are actually sort of silent behind the scenes memory overclocking.
You can drive yourself crazy obsessing over
JEDEC specifications and reading endless tech articles but the bottom line for me is I want my memory to work as advertised and I don't really want to have to go through endless contortions to make that happen. I will likely purchase either Kingston or Crucial for my next build and see how that works. Not saying that their is anything particularly bad about Gskill, I suspect some of the blame may go towards the MoBo and the cryptic references to supported memory and O/C'd notations, which I suspect goes ignored by many until times like these, I just want my life to be simpler and my computer not Blue Screening for no apparent reason, which turns out to be, not bad memory actually, but a very subtle and pain in the butt to find, one tenth of one volt D.C.