cooling the cpu from behind

f00t · Feb 7, 2003

Would cutting a hole in the mobo mount behind the cpu, attaching a fan to it,and cuttig a hole in the side of the case lower the temps of the cpu at all.
The fan would suck cool air in from the outside and cool the socket/ cpu from the back. Is it worth the time to do this mod?

setl4nothing · Feb 7, 2003

i actually thought about doing this as well. im having a hard enough time cutting out a door window though. I think that that would probably bring in a lot of dust behind your board...pain in the butt to clean out i would imagine (filter would be a simple solution) but, if its bringing in cold air, it would probably help out a lot. ever felt the backside of an over-clocked m/b? i know mine gets a little warm. so hell, try it out and let us know!
nolan

myrmidon84 · Feb 7, 2003

yeah that is a good idea...would someone please try it and let me know if it helps...hey i might just try that myself...well ill think about it ths weekend and if i do it i will let you know

-M84

slater3333uk · Feb 7, 2003

yes it works
it could take 2-3C off my cpu temp and 5-6C of my mobo temps

have fun

c0nfusion · Feb 7, 2003

Good idea...I'll have to think about this when i get some extra cash

HiProfile · Feb 7, 2003

Assuming this is a cpu that mounts like a P4/XP, don't nearly all boards have stuff on that part of the pcb? I know that much of it might be part of the temp prob circutry, but...

Maybe someday some crazy nut will make a small WB and watercool BOTH sides of the cpu, such as people did with air and slot-form factor cpus like the early Athlons and P3s. Not a bad idea actually...maybe thats how people will bust the 5ghz barrier in the future - freeze both sides.

Yodums · Feb 7, 2003

Does anyone have correct comparisons using something that is measured by on-die temps? Because if your board is insocket that thermistor can be affected by the air going towards it subtracting more than it actually is.

Yodums

HiProfile · Feb 7, 2003

Is the accuracy even a deal with airflow around it? Ceramic/organic material, such as cpu cases, are know for insulating well, so its not like you really loose out much from skewing the reading from an already inaccurate device.

Krowa 02 · Feb 7, 2003

Just be careful when you do it, you dont want anything to fly onto your board, just a thought

.

Yodums · Feb 7, 2003

HiProfile said:
Is the accuracy even a deal with airflow around it? Ceramic/organic material, such as cpu cases, are know for insulating well, so its not like you really loose out much from skewing the reading from an already inaccurate device.

Well yes considering if you put a fan right over a thermistor it may subtract alot but it isn't really that much of an increase. If it insulates too much I don't understand how it can even read temperatures? That's why the on-die sensors are so dependable.

Da_Beeblebrox · Feb 7, 2003

It might work great. If you had it pulling air through a blow hole in the side of your case it would work great. And you could put the screen on the case instead of the hole in the mobo so you wouldn't have to have the screen so close to the cpu (wonder if it would somehow insulate the cpu or something)

Eheh... $0.02

KlinKx · Apr 4, 2003

Here are my out of box results with a 60MM x15MM, 22CFM fan http://www.svc.com/60x15mm.html

Duron 900@1176 (112*10.5) 1.775Volts, Alpha PEP66, MSI Turbo Ver3, chipset cooled with same fan on lapped stock chipset HS.

Without fan on backside 35idle/39full
With fan 33idle/37full

Hope it helps with the OC ...

KlinKx · Apr 4, 2003

I'd say it helps regardless of any affect on the probe readings. Before the fan, I couldn't boot past the Windows logo at 113*10.5, now I can actually log on the machine (not quite stable though). I double checked this with Knoppix and the results are the same.

cV · Apr 5, 2003

It works. Leaving the door off the right side of the case for me helps, because the backplane has vents. 2-3C drop. That's with no fan.

Tomac · Apr 5, 2003

It works. After buying a window sidepanel w/fan for my new case I replaced the non-fan sidepanel under the mobo tray with the original sidepanel w/92mm fan. Blowing air in under the motherboard tray (which does have several ventilation holes) I saw a 5C drop in cpu surface temp and a 2C drop in cpu internal temp. System temp didn't change.
Tomac

rogerdugans · Apr 5, 2003

I have tried this on a few systems in the past too: it works, but how much its of value is....very system dependent it seems.

And the indicated temps shown by in-socket thermistors will be thrown off some more- as if they weren't skewed enough in some cases

From the systems I tried it on: not worth the effort of doing it, mostly. I did run one system faster that way though.

Gnufsh · Apr 5, 2003

Blowing right on the back of the cpu would yield benefits. You could even put a heatsink there. THere problem is, cutting a hole in the motherboard is probably not a good idea. Most Mobos are 4-layer PCB, so you won't be able to tell for sure that there are no traces there (and I think there probably are). Just blowing on to the back of the mobo probably helps too. There will e secondary heatpaths through the traces and pins into the socket. THey're conductive. Don't know if it's worth the effort, though.

Daemonfly · Apr 5, 2003

Think he's talking about just cutting the mobo mounting tray.

It will deffinately work, especially with air cooling/heatsinks. The circuit board, and especially the copper traces are just another path for heat to travel. Back of the CPU area usually does get a bit warm, and any cooling there would help CPU temps slightly. Probably won't be anything extreme, but might just be enough to stabalize you at an currently unstable speed.

TheMightyBuck · Apr 6, 2003

it's an idea worth tryin'... it really ain't that hard, and it can't hurt. i'll tell ya'll how it impacts my system... after i'm done with my watercooling....

that'll be in about 2 years.!

PalominoBURN · Apr 6, 2003

I GOT AN IDEA!

Ok.. what if.. we epoxied a small heatsink to fit directly in the middle of the pins on the back of the mobo? There is always a dead space correct? Then have a fan blow directly on this. Perhaps with the greater surface area provided by the heatsink we can see a bigger decrease in cpu temps. What do you guys think? I am assuming there is nothing there to conduct, cause any traces that might be there will be a layer or 2 under. Give me some feedback.

cooling the cpu from behind

Member

Member

Registered

Member

Member

Member

Senior Member

Member

Member

Senior Member

Member

Member

Member

Member

Registered

Linux challenged Senior, not that it stops me...

Senior Member

Member

Banned for being a BAD TRADER aka a Troll

Member

Similar threads