Page 2 of 2

Posted: Sat Jul 14, 2007 3:30 am
by Das_Saunamies
It is simple. Paper is quoted as to catch fire at above 200 degrees centigrade under standard pressure and air mixture conditions. If that limit is reached, paper catches fire and you're in trouble. The means are irrelevant.

If you don't want to risk a fire, electrical or otherwise, don't use flammable materials! :lol:

Posted: Sat Jul 14, 2007 3:56 am
by nick705
Das_Saunamies wrote:It is simple. Paper is quoted as to catch fire at above 200 degrees centigrade under standard pressure and air mixture conditions. If that limit is reached, paper catches fire and you're in trouble. The means are irrelevant.
No, I was actually agreeing with you... the counter-argument that's usually made is that standard PC components don't normally get anywhere near Fahrenheit 451, so there's no danger of using paper/cardboard/whatever.

It's the "normally" bit that's the sticking point though... there's a fair bit of anecdotal evidence that PCs can and do occasionally catch fire for whatever reason, so personally I'd rather not risk it...

Posted: Sat Jul 14, 2007 4:04 am
by Das_Saunamies
And I'm making my own arguments. Presented with the facts, it's up to common sense whether a person wants to risk it or not. Thanks for bringing in the exceptions, as this conversation was indeed only operating on standard grounds... although my generalized "so no worries" was the only comment here describing it as safe, and left unquoted.

Cardboard is great for ducting, or you can even live in it, but it's hardly the best choice for either purpose.

And hey, nothing for granted, paranoia rules. I'm paranoid myself, so I can never live with less than three HDDs and permanent, separate backups... :wink:

Posted: Sat Jul 14, 2007 8:01 pm
by cmthomson
On extremely rare occasions a CPU (or north bridge) that suffers thermal runaway will get extremely hot and more-or-less explode. This is technically not a fire (which would be something you could sue for), but a "rapid thermal event". No, I am not making this up...

But that is very very rare. It is more common for a power supply component to do this, since they generally run much hotter than any other chips in a PC.

That said, the probability of any component in a PC getting hot enough to ignite paper or cardboard is vanishingly small, and even if it should happen, the volume of paper in a duct is so small that it would burn up in a few seconds.

In other words, don't worry about it.

Posted: Sun Jul 15, 2007 8:08 pm
by rickster

Posted: Mon Jul 16, 2007 1:47 am
by russsss
WR304 wrote:
BIONIC_EARS wrote:Hehe, a "BIONIC_EARS style vent". I based my duct off the one by WR304 in this thread (middle of first page):

viewtopic.php?t=40686&postdays=0&postorder=asc&start=0
The main point of that thread was that underclocking the card for 2D mode is well worth doing, along with reseating the GPU core's thermal paste.

At 660mhz core your card is probably putting out a lot of heat. An 8800GTS card should run quite happily with the core clock speeds set at 100mhz for 2D mode usage.

If you're sticking with the stock cooler a single slot PCI blanking plate would be less restrictive of airflow. You'd either cut up the included 2 slot plate or fit a one slot PCI plate from an older graphics card. :)
I'll have a go at reseating the card - from memory it is running a lot slower for 2D anyway (somewhere around 300MHz) - but the RAM still runs full tilt in 2D (2.1GHz).

I experimented with a vent from the front fan directing 90% of the air up to the card cooler fan and whilst this had good results on the graphics card (3-4 degrees C drop in GPU temps) it had zero effect on CPU temps, which was interesting and slightly puzzling - theory up until now was the graphics card dumping heat into the case pushing up overall case temps, which doesn't seem as clear-cut now. All I could think is that by redirecting the front fan to the GPU there's little/no airflow to the Ultra 120 now so I've created a new problem in solving the first one! :?

I'm thinking of adding a fan to the front of the top chamber now (by top I mean the top half of the P180 top chamber!) to see if this improves the cooling (somewhere around the top of the 5 1/4" bays) - I also hope that a couple of Noctuas and the Ultra 120 Extreme will help...

Posted: Wed Jul 18, 2007 2:57 am
by yungsters
A friend and I are planning on making the GF8800 ducts out of acryllic.

Why do the ducts usually stick out over the graphics card? (i.e. in the blueprint by BIONIC_EARS, the side that is extra long) Doesn't this allow some of the hot air to vent up above the graphics card?

Wouldn't it be better if the duct was completely underneath the graphics card?

Posted: Wed Jul 18, 2007 4:13 am
by Das_Saunamies
If the side were below the card edge, it would vent out even more and lose positive pressure. Sealed ducting would provide the least overflow but be more prone to stressing the fan with backpressure as well.

You need a pressure balance(venting room) for it to work, and it's not an exact science this.

Posted: Wed Jul 18, 2007 10:22 am
by BIONIC_EARS
I was also planning to make an acrylic duct, but then I thought that I might change the cooler at some point, and it wouldn't be worth the extra work, so I went with this little plastic duct.

I made one modification to it since. I added tape (only sticking to the duct) to help block the gap between the case slot opening and the side wall of the duct.

An acrylic duct would have a different design. I might very well have kept it under the card. I haven't really thought about it though.

The reason I made the duct extend on the side is that it simplifies the mounting, and, being quite snug against the cooler, there is an adequate seal. I could always add a skirt out of tape to help seal it further without actually sealing it by taping the two surfaces together, but I wanted to leave a little breathing room.

What little heat escapes from this duct rises up to the case exhaust fan, so it's doing its job as a diverter. That's another benefit of having the side extend up when there isn't a complete seal.

Good luck with your acrylic duct. Post pics when you're done!

Posted: Mon Jul 30, 2007 8:17 pm
by christopher3393
yungsters wrote:A friend and I are planning on making the GF8800 ducts out of acrylic
Any development yet?


An acrylic duct would have a different design. I might very well have kept it under the card. I haven't really thought about it though
BIONIC_EARS: Have you given this any more thought? Why would an acrylic duct have a different design?

I'm building a system in a P182 and have an 8800GTX. I also have an HR-03, but thinking that something like this might be sufficient. Looking for the most effective ducting approach. Thanks.

Posted: Mon Jul 30, 2007 8:31 pm
by yungsters
The friend who is making the ducts with me has been focusing on putting together his new computer (in a P182) and two days ago, we were experimenting with hard drive suspension. He's currently waiting for his defective motherboard to be replaced, so the ducts will come... sooner than later, hopefully.

Meanwhile, I've put together a little cardboard/duct tape duct held in place with a paperclip. Honestly, I haven't noticed any noticeable temperature drop, although I do feel the temperature difference right below the graphics card.

Posted: Mon Jul 30, 2007 9:25 pm
by Lensman
I went with the slot exhaust fan idea - picked one where the intake matched up with the exhaust slots of my 8800GTS so I could mount it on the immediately adjacent slot.

I couldn't find one that was quiet enough so I attached it to my fan controller.

Posted: Mon Jul 30, 2007 10:02 pm
by yungsters
How much did the slot exhaust fan change your temperatures? Also, what kind of case do you have and how much empty space is below your graphics card?

On another note... (but still related to this topic) on my P190, I notice that with my side panel on, my temperatures are higher than when it's off, by 3-5*C. Also, with my side fan on, the CPU and GPU are hotter but the (passively cooled) NB is cooler. Any ideas on how to deal with that?

Posted: Mon Jul 30, 2007 11:39 pm
by Das_Saunamies
yungsters wrote:--P190, I notice that with my side panel on, my temperatures are higher than when it's off, by 3-5*C. Also, with my side fan on, the CPU and GPU are hotter but the (passively cooled) NB is cooler. Any ideas on how to deal with that?
Airflow issues. Hotter-with-side-panel-on no surprise as taking it off increases cool intake manyfold. Nothing to worry about if you can live with the temperatures -- increase intake if you can't.

Active cooling getting hot and passive cooling going cool suggests side fan is blowing in rather than out. Reverse it to see if that helps, if not then it's messing up the airflow. Reduce RPM from side fan and power up fans circulating in the proper direction(bottom to top, intake to exhaust).

Posted: Wed Aug 15, 2007 12:32 pm
by v3n
GPU cooling design needs some radical change as the amount of heat created by these Uber cards is getting silly

My GPU idles at 50c and it has a PCI X-fi 1mm below it so I have no idea how it gets any air

I think external heatpipes extending out of the back of the case would be great 8)

Posted: Fri Aug 17, 2007 4:44 am
by MoJo
Here is a PNG version of the image BIONIC_EARS posted. The only difference is that it has correct DPI information, so if you print it at 1:1 scale it should have exactly the right dimensions.

I printed on normal paper, then stuck that over some card and marked the corner points. I then marked the card up with a pencil and cut it to shape. Off to fit it now.

Image

Posted: Fri Aug 17, 2007 6:10 am
by christopher3393
Nice work, Mojo! Thanks.

Posted: Fri Aug 17, 2007 6:11 am
by christopher3393
Has anyone tried constructing this vent from acrylic?

Posted: Sat Aug 18, 2007 3:16 pm
by wumpus
I agree, it looks like a constricted airflow problem on the P182. This would explain why there's such a large delta between temperatures with the side panel on and off.

For the record, I found my P182 / 8800GTX combo would overheat during gameplay because it was basically suffocating. Plenty of air going out, but not enough air coming in. I documented my results here. There's graphs and everything..

viewtopic.php?t=42801

I think you absolutely HAVE to remove the fan cover, fan grill, and dust cover for the front middle fan if you have a high powered system in the P182. Even then you may have to still ..

a) open the front panel of the P182 when you play games to let it "breathe"

b) install a front fan so enough air is forcefully drawn in through the restrictive front panel when closed

Posted: Sat Aug 18, 2007 3:51 pm
by cmthomson
A P180/182 runs significantly cooler with the front door open, especially if you've added extra intake cross section (eg, with a Kama Bay).

You don't need an intake fan, but it is definitely a good idea to remove the swing-out doors that suffocate the inlet filters.

Posted: Tue Apr 15, 2008 3:35 am
by phantom220
Just had to add in my 2 cents' worth. Using Mojo's copy of the BIONIC_EARS duct, my GPU temps dropped a solid 6-7C. I was amazed to say the least. While I can't run the 8800 fan any slower (stupid Vista won't let me install RivaTuner or ATITool), I'd encourage anyone out there with a stock 8800 to make themselves a duct. Pretty easy and the results were awesome :)

Posted: Tue Apr 15, 2008 3:59 am
by MoJo
Replacing the thermal paste can also give you a few degrees improvement. I wanted to open up the grills at the back a bit but didn't want to do irreparable mods to an expensive card.