DUCTING = superior CPU temperatures
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
-
- Posts: 580
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: USA (Phoenix, AZ)
DUCTING = superior CPU temperatures
Well, I've always known ducts rule. I had one on my last heatsink. But then I picked up an AX-7 with an 80mm fan, so the hard work I did on my previous duct had to go.
I didn't mind, the AX-7 alone without duct still provided better temperatures. But then I had to try something...
I made a pathetic duct out of duct tape (HAHAHA!) no, actually it was masking tape.
After leaving the machine on a few hours under the same ambient temperature, I'm looking at a solid 7C decrease in idle CPU temperature (Celeron 1.3) according to MBM 5.
I custom made my previous duct out of aluminum - it looked great but it was a job to create it. What if I later change my mind and get a different heatskink or motherboard or whatever, then I have to create a whole new duct. I don't want to put that effort into it again. Plus I have a window and don't want to stuff an ugly tube in there... Maybe if I can find a clear flexible tube somewhere, at least 80mm...
I didn't mind, the AX-7 alone without duct still provided better temperatures. But then I had to try something...
I made a pathetic duct out of duct tape (HAHAHA!) no, actually it was masking tape.
After leaving the machine on a few hours under the same ambient temperature, I'm looking at a solid 7C decrease in idle CPU temperature (Celeron 1.3) according to MBM 5.
I custom made my previous duct out of aluminum - it looked great but it was a job to create it. What if I later change my mind and get a different heatskink or motherboard or whatever, then I have to create a whole new duct. I don't want to put that effort into it again. Plus I have a window and don't want to stuff an ugly tube in there... Maybe if I can find a clear flexible tube somewhere, at least 80mm...
Simple & Effective? Yes
Attractive & Aesthetically Pleasing? Hardly...
I think gaming god is a little closer to the mark, as I doubt poster paper will be much better than a duct tape duct. (Especially considering you've got a window.)
Looking at the furnace in my home gave me an idea that you might be able to incorporate. Where the ductwork connects to my furnace is a flexible section, looks like some sort of vinyl. I assume this is there to allow for inexact placement of the furnace and ductwork, and to allow for either to shift a bit as the house settles without bringing the ductwork down.
If you constructed your duct to get you in the vicinity of the HSF, then build a flexible adapter to mate the duct to the HSF, you could probably adjust for any later HSF mods to your current mobo. If the CPU placement is similar, you might even be able to swing a MOBO upgrade.
Queue
Attractive & Aesthetically Pleasing? Hardly...
I think gaming god is a little closer to the mark, as I doubt poster paper will be much better than a duct tape duct. (Especially considering you've got a window.)
Looking at the furnace in my home gave me an idea that you might be able to incorporate. Where the ductwork connects to my furnace is a flexible section, looks like some sort of vinyl. I assume this is there to allow for inexact placement of the furnace and ductwork, and to allow for either to shift a bit as the house settles without bringing the ductwork down.
If you constructed your duct to get you in the vicinity of the HSF, then build a flexible adapter to mate the duct to the HSF, you could probably adjust for any later HSF mods to your current mobo. If the CPU placement is similar, you might even be able to swing a MOBO upgrade.
Queue
-
- Posts: 137
- Joined: Mon Nov 04, 2002 3:55 pm
The obvious material to me would probably be a clothes dryer duct segment. They're designed to be thermally efficient to avoid waste heat escaping into the home, they're inexpensive, they're flexible, and they're big.
The construction of the mating between the HSF and ducting would seem to be the tricky part.
The construction of the mating between the HSF and ducting would seem to be the tricky part.
-
- Friend of SPCR
- Posts: 228
- Joined: Sun Sep 29, 2002 9:05 pm
- Location: Powell River, BC, Canada
From the waky ideas department: this author hereby disclaims all credit, liability, danger, death, etc., ad nauseum, for said rediculous and hither-to untested ideas
---How to build a Duct---
Materials:
1. peice -o- slinky, or peice -o- bailing wire
2. peice -o- funky spandex or lycra (sumpn' real stretchy like, pick your favorite psycho pattern)
3. <optional> can -o- clear coat, spray on (liquid plastic like)
Assembly:
A. Cut and bend wire frame. Slinky would be spiral, stiff wire could be 'spiralled' in any progressive shape desired.
B. Cut and wrap item #2 around item #1, leaving a 'pinch' of overlap. Note the stretch factor here for getting good fit easily.
C. Staple (std. paper types) along the pinch to fasten. Might want the pinch around the back side for cosmetics.
D. If you opted for item #3, then spray away to prevent any small leakage. Expect also a stiffer/rigid end result.
My guess is that most of our user ducting will always be on a 'case by case' basis, so we might as well get used to it and have some stoopid fun. As long as it works. As for the clear coat idea above, I doubt you would see more than a few % of leakage even on thin fabric without the coating. The pressure is just so low it should hardly count.
---How to build a Duct---
Materials:
1. peice -o- slinky, or peice -o- bailing wire
2. peice -o- funky spandex or lycra (sumpn' real stretchy like, pick your favorite psycho pattern)
3. <optional> can -o- clear coat, spray on (liquid plastic like)
Assembly:
A. Cut and bend wire frame. Slinky would be spiral, stiff wire could be 'spiralled' in any progressive shape desired.
B. Cut and wrap item #2 around item #1, leaving a 'pinch' of overlap. Note the stretch factor here for getting good fit easily.
C. Staple (std. paper types) along the pinch to fasten. Might want the pinch around the back side for cosmetics.
D. If you opted for item #3, then spray away to prevent any small leakage. Expect also a stiffer/rigid end result.
My guess is that most of our user ducting will always be on a 'case by case' basis, so we might as well get used to it and have some stoopid fun. As long as it works. As for the clear coat idea above, I doubt you would see more than a few % of leakage even on thin fabric without the coating. The pressure is just so low it should hardly count.
-
- Posts: 137
- Joined: Mon Nov 04, 2002 3:55 pm
-
- Posts: 137
- Joined: Mon Nov 04, 2002 3:55 pm
Don't know if this URL was posted yet:
http://www.overclockers.com/tips586/
This definitely looks like a worthwhile direction to investigate. Larger, slower fan with directed air.[/url]
http://www.overclockers.com/tips586/
This definitely looks like a worthwhile direction to investigate. Larger, slower fan with directed air.[/url]
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
I agree pretty much 100% with Rick Talbot who did that article for OC. It's not rocket science, pretty simple stuff (as most of all this is, really). But there are issues with implementation. Sure a duct helps things keep cooler -- does it increase noise? I've played with carboard mailing tubes, tunnels built of medite board, add-on plywood and cardboard mufflers for fan in/out vents, and the vinyl ducting above. I don't use them any more because I don't generally find the need to. Undervolting the CPU and judicious use of 5V quiet decouple-mounted fans with completely open vents (no grill) keep my systems quiet enough for me. About the noise level of a single 7V Panaflo a few feet away.
1) tubes are naturally resonant: the air in them will resonate at specific frequencies. Try holding one to your mouth and humming or singing scales or just making noise -- low to high & back. At some freq, you'll actually feel the air resonance of the tube, it will add volume (loudness) at certain frequencies. The stiffer & thinner the material, the more pronounced the effect. A fan is generally a broadband noise maker, therefore, it will excite resonances in tubes. Only way this can be avoided is to keep the fan noise very low.
2) any impedance near / around the fan increases air turbulence. Ducts do, too, in varying degrees. Just try it. Again, this can be avoid with quiet low airflow fans.
3) smooth contours (streamlining) in the tube including in / out ends minimizes turbulence & noise.
4) a thin layer of light smooth foam applied to the inside surface of the duct can also help reduce noise (if the airflow is significant -- maybe 20 cfm & higher?)
Still not sure about one slow big fan vs a few smaller slow fans. I have tried both & think both can work; the former almost requires a duct.
ENuf said. I bore myself going on & on &....
1) tubes are naturally resonant: the air in them will resonate at specific frequencies. Try holding one to your mouth and humming or singing scales or just making noise -- low to high & back. At some freq, you'll actually feel the air resonance of the tube, it will add volume (loudness) at certain frequencies. The stiffer & thinner the material, the more pronounced the effect. A fan is generally a broadband noise maker, therefore, it will excite resonances in tubes. Only way this can be avoided is to keep the fan noise very low.
2) any impedance near / around the fan increases air turbulence. Ducts do, too, in varying degrees. Just try it. Again, this can be avoid with quiet low airflow fans.
3) smooth contours (streamlining) in the tube including in / out ends minimizes turbulence & noise.
4) a thin layer of light smooth foam applied to the inside surface of the duct can also help reduce noise (if the airflow is significant -- maybe 20 cfm & higher?)
Still not sure about one slow big fan vs a few smaller slow fans. I have tried both & think both can work; the former almost requires a duct.
ENuf said. I bore myself going on & on &....
-
- Patron of SPCR
- Posts: 543
- Joined: Mon Oct 21, 2002 1:01 am
- Location: Vancouver, BC, Canada
- Contact:
Put yer sewing skills to the test!
I'm in the process of making a duct for my dually, and I came up with the idea of sewing together a duct from clear vinyl. I was thinking I could get a grade of vinyl that was soft enough to sew and bend around, but stiff enough to hold its shape. At each end of the duct, I wanted to be able to sinch it so I thought of sewing a channel for a drawstring. The drawstring would be clamped together with a button clamp - like the ones at the bottom on ski jackets.
Crazy idea eh?
Crazy idea eh?
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
here's some other duct ideas... http://www.muffledcomputing.com/
-
- Posts: 137
- Joined: Mon Nov 04, 2002 3:55 pm
Mike,
As usual, you bring up some good points. I just noticed from looking at my Dell Precision 220 that it has a shroud, not a tube. The shroud is open on the bottom (it covers two Slot 1 P3 933 processors), and I'm wondering if the fan is supposed to be blowing air onto the heatsinks, because right now I have it blowing air out of the case (which is what I thought the original fan was doing when I replaced that one -- the original was dying a horrifying squeaky death).
Anyone know if MBM works with the older Intel VC820 dually mobos?
But I digress...point being that a shroud probably addresses all your concerns about noise and resonance. I think the foam lining idea isn't a bad one either.
As usual, you bring up some good points. I just noticed from looking at my Dell Precision 220 that it has a shroud, not a tube. The shroud is open on the bottom (it covers two Slot 1 P3 933 processors), and I'm wondering if the fan is supposed to be blowing air onto the heatsinks, because right now I have it blowing air out of the case (which is what I thought the original fan was doing when I replaced that one -- the original was dying a horrifying squeaky death).
Anyone know if MBM works with the older Intel VC820 dually mobos?
But I digress...point being that a shroud probably addresses all your concerns about noise and resonance. I think the foam lining idea isn't a bad one either.
Actually if we should believe John Carmack, even actual rocket science "isn't rocket science"MikeC wrote:It's not rocket science, pretty simple stuff (as most of all this is, really).
Pretty intresting stuff:
http://www.armadilloaerospace.com/n.x/Armadillo/Home
(Totally unrelated but I immediately associated to this when i read the above statement )
Turns out that the slinky-duct might not work.
Well, I went out to my local Home Depot and picked up a bunch of flexible tubing (dryer tubing, semi-rigid aluminum, plastic). It turns out that tubes are, in fact, very good at resonating . Also, as the interiors of flexible tubes are rarely smooth, they create some very turbulent flow. I made a more open duct out of cardboard, and it works great.
I should have remembered the first rule of engine porting:
Make it easier for the air to go where it wants to go, not where you think it should go.
After heat-forming/cutting acrylic to make a hard drive suspension cradle, I am pretty confident that I will be able to build a duct that's tough and quiet. We'll see.
I should have remembered the first rule of engine porting:
Make it easier for the air to go where it wants to go, not where you think it should go.
After heat-forming/cutting acrylic to make a hard drive suspension cradle, I am pretty confident that I will be able to build a duct that's tough and quiet. We'll see.
-
- Posts: 137
- Joined: Mon Nov 04, 2002 3:55 pm