Measuring cpu supply voltage at mos-fets [pics]

[Lowther]

In my efforts to find out why my XP2500 keeps throwing me out of 3dMark at almost any speed except stock,
(see this earlier thread: http://www.xtremesystems.org/forums/showthread.php?threadid=28363&goto=newpost ),

I wanted to measure the 5V rail (which feeds the cpu, right?) directly at the mos-fets.
I think I got it right but i just wanted to double check with you guys. It´s the middle leg of the mos-fet right?
The short one thats not really connected, at least, that was the only one that appeared close to beeing right.

http://hem.bredband.net/b107319/forum/mos-fet.jpg

And if what I got there is right, my 5V rail is really stable during 3d benching, despite my PSU being a nasty old Aopen 300W one.

http://hem.bredband.net/b107319/forum/DMM.jpg

[kommando]

GEt a new psu, might help :)

[saaya]

just check the vcore comming from the mosfets.

[Lowther]

Originally posted by saaya
just check the vcore comming from the mosfets.

So my actual Vcore is on one of the other 2 legs??

[enzoR]

yup

[saaya]

well even this will not be the actual vcore, if you want to meassure the real vcore you have to meassure the little chips ON the cpu itself. one side of each chip will be ground and the other side will be the real vcore the cpu core gets.

hard to meassure since you have to move the heatsink a little to be able to reach those chips while the cpu is running. a friend was holding the heatsink on the core while i meassured the vcore. its around 0.02v lower than what the mosfets read.

[Lowther]

Well, i think measuring on the legs of the mosfets will be good enough for my purposes.
Now I only wish I had a oscilloscope at home.......!