Is 1.4v on the NorthBridge too high for a P45?

Ok scratch that. I got an error with the Windows Memory Diagnostics at 1.4v. By my notes, every time I've seen this error it is at the exact same memory address (1dc2e2f to be precise). I would guess that to be a memory problem, but I have yet to see this memory fail in single channel. I'm going to run a few passes on each memory module and see what happens.

I guess I can rule out voltage seeing as how the address is so consistent. I would imagine a voltage issue would be a bit more sporadic.

RAM is set to OCZ specification of 2.1V as stated on the RAM chip. CAS Timings are also set in accordance to the OCZ specifications.

As for the NB, Intel places their memory controllers on the northbridge. The memory controller is responsible for sending electrical current through to the RAM thereby writing 0's and 1's to the chip itself. RAM is not persistent and the chip capacitors will lose their charge if the memory controller is broken or not doing its job for some reason. The theory was that if the memory controller (i.e. the northbridge on an Intel chipset) is not receiving enough juice and it isn't running adequate current through to the RAM then there would be scenarios where an expected result was not written to the RAM or was leaked from the RAM.

It isn't uncommon for people who overclock their PCs and toss 8GB of ram across 4 slots to have to increase the voltage being delivered to the memory controller.

As for the methods of testing. I tested each chip in single channel to determine if there was a faulty chip. In dual channel mode the memory controller reads data in parallel from both chips, so an error could be with either one. If each chip tests ok, then the problem may be something other than the chip. Possibly the bus itself, the memory controller, dog saliva on the motherboard, whatever.