more volts on ram = less heat??!!

[Mastakilla]

im testing this too atm,

but id like to hear it too from someone who knows...

i have my pc at 10x250 atm
with 1.55v x 104% on cpu (less wont do 10x250)
and my corsair ram tweaked to the max at 2.6v

now this setup only works when i have my EXTREMELY LOUD 120mm fan aimed at the ram, or more precise, a lil to the right of the ram... (first it was right on the ram, but it still wasnt 100% stable then when looping memtest 8 long time with the 120mm fan @ 7v... after turning the fan a lil to the right it was)

at 7v that fan is still louder and more cmf then my coolermaster fans at 12v, so it still is quite alot airflow when its pointed straight on the board...
so i think something must be getting really hot then

i am thinking now that pwm ic or one of the mofsets (or whatever those things that get hot are called) at the right of the mem are the cause (since mem is only running 2.6v it should get sooooo hot that it needs sooooo much airflow)

now normally i would think 2.6v stresses the mobo and stuff less then fe 3.0v, but then i thought that those 2.6v are derived from the 3.3v line... so can it be possible that running the ram @ 2.6v makes something on my mobo hotter then running it at 3.0v??

edit:
3.0v on ram just failed even earlier then 2.6v... (3.2v didnt boot in windows anymore, trying 2.7v now) so i guess it must be something else thats getting hot... but what?

[Mastakilla]

hmz

think im wrong somehow

after testing it seems 2.6v is the best for me (occt survives the longest with it with fan at 7v, 30min)

something else is getting too hot i think
pwm ic doesnt get above 38°C while occt with fan at 7v, so i dont think its pwm ic
i also dont think its mem, since earlier i pointed my fan a lil away from it (more to the right, around pwm ic i guess) to get better results

anyone have an idea what?

[conrad.maranan]

More voltage will always equate to more heat. Likewise, higher frequencies will also generate more heat. Another tidbit of food for thought is that with your typical fan, the centerpoint is your deadspot. Air is actually being pushed or pulled to or from the area away from the center. This is why fan shrouds are used in automotive applications (radiator). It's use has also been adopted by the PC watercooling guys.

Check your CPU temps. If your CPU gets too hot, so does your memory controller. If your memory controller gets too hot, you'll start to notice the errors.

[EMC2]

--- killa ---

Be sure to go over and read this post by AndyOCZ regarding VCore settings.

Regarding your Vmem setting issue... you may have found two different limits for your current setup. The first is that your memory may not handle the higher voltage. The second is that your MB may not like the low voltage. The area you describe as cooling with your fan when Vmem is at 2.6V is the area of the regulator that generates the Vmem voltage. It is created from the 3.3V rail in the standard jumper configuration, meaning that the farther you get from 3.3V, the hotter the regulator would get.

Here's a pic of the MB area being discussed for clarity:

[Mastakilla]

More voltage will always equate to more heat. Likewise, higher frequencies will also generate more heat. Another tidbit of food for thought is that with your typical fan, the centerpoint is your deadspot. Air is actually being pushed or pulled to or from the area away from the center. This is why fan shrouds are used in automotive applications (radiator). It's use has also been adopted by the PC watercooling guys.

Check your CPU temps. If your CPU gets too hot, so does your memory controller. If your memory controller gets too hot, you'll start to notice the errors.

intresting point, so it might actually still be the ram that is getting hot u think? (even only @ 250mhz and 2.6v?)

damn i really need to find a way to find the hot spot.....

--- killa ---

Be sure to go over and read this post by AndyOCZ regarding VCore settings.

Regarding your Vmem setting issue... you may have found two different limits for your current setup. The first is that your memory may not handle the higher voltage. The second is that your MB may not like the low voltage. The area you describe as cooling with your fan when Vmem is at 2.6V is the area of the regulator that generates the Vmem voltage. It is created from the 3.3V rail in the standard jumper configuration, meaning that the farther you get from 3.3V, the hotter the regulator would get.

Here's a pic of the MB area being discussed for clarity:

aha

so my idea wasnt so bad afterall :p

which one is it on that pic?
the one with the heatsink on? (that one i still might try to touch, the others look a lil dangerous to touch...)

after lots of testing yesterday i couldnt get my pc stable at 10x250 with the big fan on 7v...

occt would fail after 30 mins
the ram does seem to be able to handle it, since memtest looped test 8 10 times and test 5 16 times...

[conrad.maranan]

2.6VDIMM isn't anything out of the ordinary. Your RAM is considered warm at that voltage and those clocks. When you touch the memory modules, do they feel so hot that you can't keep your finger on them?

[Mastakilla]

ill try that tonight

but is it safe to touch the ram's heatsinks during memtest?

[conrad.maranan]

You can touch the modules, just make sure you're not statically charged.

[EMC2]

which one is it on that pic?
the one with the heatsink on? (that one i still might try to touch, the others look a lil dangerous to touch...)

after lots of testing yesterday i couldnt get my pc stable at 10x250 with the big fan on 7v...

occt would fail after 30 mins
the ram does seem to be able to handle it, since memtest looped test 8 10 times and test 5 16 times...

The main Vmem hot spot is the large aluminum heatsink in the pic where the FETs reside. It can get hot, hot. The only danger on the MB anywhere is the issue of static damage to components (and if you have slooooow reflexes maybe burns to your fingers from heat).

[Mastakilla]

just chked it

and with the loud ass 120mm fan off (but all other silent 120mm fans on) i feeled the following during memtest test 8 loop:

10x250 (not 100% stable, error at pass 3, touched at pass 2 i think)
ram = very hot, but still possible to keep ur finger on it
regulator = very hot, but still possible to keep ur finger on it

10x244 (stable i think, no errors till pass 13, touched at pass 13)
ram = very very hot, possible to keep ur finger on it, but its getting really close (hotter then 10x250!!)
regulator = hot, less hot then at 10x250, but still hot

so i THINK its that regulator that prevents me from doin doin 10x250 in memtest...
if anyone thinks different plzzz convince me

[conrad.maranan]

Weird. I find it strange that everything is hotter with the lower clocks. I wish I could help you out, but your problem has got me feeling dumb.

[EMC2]

Weird. I find it strange that everything is hotter with the lower clocks. I wish I could help you out, but your problem has got me feeling dumb.

Look at the pass count he did the touch test at (13th pass for 244, 2nd pass for 250)

--- killa ---

What VMem were those 2 tests done at? (if at same, I would say both need the cooling, but ram more since it got "hot hot")

[conrad.maranan]

Haha. Well, there you have it.

[Skratch]

If your ram is getting hot your putting to much voltage threw them.Im running 3.27 as I type on my bh5 and there cold to the touch.That heatsink next to them is warm (I have the same board)

My problem is my bh5 is power hungry and my board/psu can't keep up.I can run 240 2-2-2-5 at 3.27 v and I tried to up it to 3.35 and The psu goes into safe boot every once and a while.I benched them at 255 2.2.2.5 and they were cold.

It could also be the board im still reading up on the 4v jumper on the board.

[Mastakilla]

its all on 2.6v...

2.5v is too low to do anything decent

and higher then 2.6v is allways a lil worse...

still find it very strange though it gets so hot at such low speeds/volts...

i also noticed that my pwm ic gets REALLY hot during memtesting

after hours of prime / occt pwm ic would only get 42°C max

but when looping memtest 8 for awhile, even 2-3 mins after it (still had to disable memtest and reboot in windows to see temp), it is still 45°C!!
so i suppose it was like 48°C while memtesting...

[Skratch]

its all on 2.6v...

2.5v is too low to do anything decent

and higher then 2.6v is allways a lil worse...

still find it very strange though it gets so hot at such low speeds/volts...

i also noticed that my pwm ic gets REALLY hot during memtesting

after hours of prime / occt pwm ic would only get 42°C max

but when looping memtest 8 for awhile, even 2-3 mins after it (still had to disable memtest and reboot in windows to see temp), it is still 45°C!!
so i suppose it was like 48°C while memtesting...

Wow I thought you were running high volts.That stuff should all run cold with those volts.

whats your voltages on the board.REcheck the board and make shure its not shorting something out.check fans to.

12v
5v
3v
chipset v

[Mastakilla]

i dont have a multimeter

in smartguardian they seem ok (not that this can be trusted)

im quite sure nothing is shorting too

else i wouldnt be runnin 10x244 stable i think (8 hours prime 8 hours memtest and more)