Cooling Solution for PWMIC Area in DFI Boards

Status
Not open for further replies.
@Crazy_Eddy,
yes you are correct.
its the reason i had specifically mentioned pwm Voltage regulators transistors in my reply above.

The actual pwm ic never really gets hot really since a very few amount of current flows through it. it basically senses the output voltage (Actual vcore in this case) through its feedback circuit and controls the switching action of the transistors depending upon the feedback of vcore (taking into consideration any error correction circuit , if any). it tries to keep it as near to the set value of vcore in the bios.
hence this ic does not get very hot. it is the switching transistors and the smoothing coils (those square blocks near the heatsinks) which get quite hot.

ps: by tinkering with the feedback circuit to this pwm ic or by tweaking the error correction circuit one can easily change the vcore to ones liking ;) .....infact most of the vcore mods actually reduce the feedback vcore voltage to this ic ( by introducing resistance ) thus fooling it into believing the vcore is less than actual.:D
 
Some really nice info there by Eddy and Dj... The temps do get high in that area... Will add some fans there and especially the Vdimm mofset as i am using UTT CH5 and it does get hot up there ...
 
Yeah great info Mr.Dj and Eddy. Touched the PWMIC Heatsinked area, its cool but still i wanna reduce the temp that the Smart guardian is showing me. I am trying to source a good fan to cool that area.

@ Switch if u add some fans post the pics here, it will be helpfull for me since we hve Ultra-D+ Super Lanboy:D
 
@ Deejay:
Yes I understood that the PWM controller IC merely sends out the pulses/sets the voltage .. actual power flow is only through the heatsinked VR's. But what I was wondering about was why I seem to find the VR's running relatively cool (using the 'finger test') :)

In addition, the thermistor is not 'in contact' with the heatsinks, indicating that it senses only the radiated heat from the VR's.

Also posting a "thermal scan" from the Big Bruin site of the DFI board showing the heatsinked area as pretty cool (but of course they're using a very simplistic A64 3000+):
http://www.bigbruin.com/reviews05/nf4thermals/imageview.php?file=6.jpg

All these seem to point that the VR should be relatively cool running or atleast the temps reported should be lower. Of course I dont have a dual-core proc on hand to experiment. Perhaps someone with a dual-core can try the 'finger test' and note how hot the heatsinked areas *actually* are ?
 
Thanks for the great info Eddy :D !!

With the 80mm fan over the heatsinks, PWMIC temps are at 36c now..

Will try to place maybe a 40mm fan over the memory heatsink soon...:)

Will keep you guys updated..
 
and what was the temp being reported in bios or temp monitor program when you did the "Wet Fingure" test crazy_eddy ??
 
^ Sorry, I guess I've caused a bit of a misunderstanding! :D

What I meant was, *mine* tend to run relatively cool, but dont show ridiculously high temps either.. mostly around the 40~45C mark. So was wondering if the rest of those seeing the high temps could actually try touching the VR's to see if the temps are really that high, or whether something else is influencing those reported temps?

Placing a fan over the VR's could be showing dramatic drops in Smart Guardian not necessarily cause they are running cooler, but because the thermistor is receiving all that cool airflow ;)

P.s: Switch, a tip for high voltage memory:

- Never use the 4v vdimm jumper if you intend to use a vdimm below 3.2v (even though it is possible to do so). Reason being, when using the 4v jumper the vdimm Voltage Regulator draws power off the 5v line and the voltage drop has to be let off in the form of heat i.e. lower the vdimm you select = more voltage drop from 5v = more heat put out by the VR. The heat is what makes the VR unstable in the voltage it outputs, making the memory unstable too.

- I've seen people leaving the 4v vdimm jumper on for convenience sake and selecting the full range of voltages including lower values. Its always ideal to stay in the higher voltage range if using the 4v jumper, i've noticed higher volts have been more stable than the lower ones (I used a marginally high 3.4v instead of a <3.2v value when trying out Darky's BH-5 mainly because of this reason and found much of this to hold true). If using anything upto 3.2v, switch back to the default jumper position :)
 
Thanks for the tip ;)... And great timing too... As you said i had left the 4V jumper on for convience sake but next thing i am gonna do as soon i hit submit button is shut the PC down and putting the jumper back to 1-2 Position instead of 2-3 position :)...

And btw Reps ofcourse :)

@Edit

You must spread some Reputation around before giving it to Crazy_Eddy again.
Click to expand...

Reps Due :P
 
Nops... The 4V jumper is off by default... The best way to check it out is to go into the bios and see whats the max Vdimm it shows... If it shows 3.2 as max then its off if it goes above that then its on...

@Eddy Done :)
 
Status
Not open for further replies.
Top Bottom