PSU's 3.3v Rail directly connected to/as vdimm

[bachus_anonym]

Originally posted by Rub87
Some time ago I also did the vdimm=3.3v mod on my NF7-s.. with one 256mb dimm installed, vdimm nicely follows +3.3v and vref=vtt/vbt is 0.5*vdimm.. but when 2 big dimms (2*512mb) are used cdimm drops to 3.2v (in bios) while +3.3v is still 3.4 even under load!, you would think, 3.2v is still good but the problem is that vref=vtt/vbt follows the vdimm thats displayed in bios so with big dimms vtt/vgt drops below 1.6 while vdimm is actually 3.4 because its directly connected to +3.3v line...

What could be the cause of this? A to thin wire?

Do you guys think I could fix this by using a thicker wire to every dimm..?

Can't I directly put a wire from +3.3v to the point were bios reads the vdimm so bios would always see the correct vdimm..?

Or build some kind'a regulator that cuts the vdimm always exactly in half to suplly vref and vtt/vbt..?

sorry for my bad english

i have almost identical situation except for i use thicker (gauge 24) with one solid copper strand.
with vio=vdimm and vref=vtt voltmods:

on idle voltages dont fluctuate
vdimm = 3.60v
3.3vio = 3.64v
vref=vtt = 1.87v

under heavy load (memtest+ or sandra mem benchmark)
vdimm drops to 3.55v
3.3vio = 3.64v
vref=vtt drops to 1.80v

before doing vref=vtt volmod vref was always 1.80v. vtt was at 1.87v and behaved same way as it does now. but i haven't seen any difference in mem stability, though.

so actually i'm not worried about vdimm drop coz i would think it's normal but that vref=vtt not being 1/2 vdimm and its fluctuation sort of pisses me off...

oh, and mem in DC config always makes vdimm and vtt drop lower that one stick in SC... no matter what vdimm mod i had...