Blow vs Suck - my thoughts

The forum for non-component-related silent pc discussions.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
CoolColJ
Posts: 303
Joined: Thu Jul 10, 2003 11:58 pm
Location: Australia

Blow vs Suck - my thoughts

Post by CoolColJ » Thu Oct 09, 2003 1:28 am

Ok I was thinking

Say your in an enclosed room, there is a heat source inside, so the room is hot.

Which would cool you better, a fan blowing cool air from outside into you
or a fan sucking cold air into the room?

Common sense the first option would be the case.
Plus there is less air restriction for a fan blowing air into the room than fan trying to suck the hot air out and hence cool air into the room.

So why doesn't this apply to PC cases?

Semm
Posts: 179
Joined: Mon Sep 08, 2003 2:06 am
Location: SoCal, USA

Post by Semm » Thu Oct 09, 2003 1:33 am

I would say it doesn't apply because in the case of the room, you aren't trying to cool a specific area, so the exact airflow doesn't matter much. Whereas in a case, if I put a big honkin' fan blowing in, and it creates an airflow "short circuit" which doesn't allow any fresh air to reach the CPU, I will have no effective cooling. Plus in the window analogy there's only one inlet/outlet, whereas your case has multiple inlets/outlets. And if you're the one measuring the cooling, rather than a thermometer, the evaporative cooling of your sweat will make something blowing across your skin feel cooler, giving the impression that the fan blowing in (onto you) makes it cooler.

Someone'll correct me if I'm wrong :)

Semm

Rusty075
SPCR Reviewer
Posts: 4000
Joined: Sun Aug 11, 2002 3:26 pm
Location: Phoenix, AZ
Contact:

Re: Blow vs Suck - my thoughts

Post by Rusty075 » Thu Oct 09, 2003 5:39 am

CoolColJ wrote:Which would cool you better, a fan blowing cool air from outside into you
or a fan sucking cold air into the room?
I think you're description is a little confused there, but I know what you meant. :lol:

The fan blowing air into the room will make it feel cooler to a person, but the temp will be the same. The reason being that the inblowing fan will produce a more concentrated air stream to blow across your skin, increasing the evaporative cooling, and making you feel colder. (It's also called "Wind Chill") But to a thermometer it doesn't matter.

CoolColJ wrote:Plus there is less air restriction for a fan blowing air into the room than fan trying to suck the hot air out and hence cool air into the room.
I don't think so. The restriction is really about the same either way. (If it's blowing out of somewhere its blowing into somewhere else) Now some fans do react differently to having a pressure diferential on one side or the other, but usually the differences are small.

puff
Posts: 38
Joined: Fri Aug 15, 2003 1:56 pm

Post by puff » Thu Oct 09, 2003 6:27 am

This is not really answering your question, but I've tested numerous ducts/airflow schemes and I've concluded that although it may seem advantageous to have one fan do both exhaust and CPU cooling, it isn't worth it.

Modern heatsinks are designed for downward airflow, and ducts add to the resistance if only a bit. You have to increase the fan speed to compensate for the rise in CPU temp, thus negating any benefits. A 80L on CPU and 80L on exhaust at ~7V is much quieter imho.

wumpus
Patron of SPCR
Posts: 946
Joined: Sat Sep 06, 2003 9:57 pm
Location: Berkeley, CA, USA
Contact:

Post by wumpus » Thu Oct 09, 2003 1:35 pm

Modern heatsinks are designed for downward airflow
Not the Alpha PAL series.
You have to increase the fan speed to compensate for the rise in CPU temp, thus negating any benefits. A 80L on CPU and 80L on exhaust at ~7V is much quieter imho.
Well, you bring up a good point -- interior fans vs. exterior fans. Some natural attenuation will occur with the interior fan.

What about mounting both fans in parallel, one on the cpu, one on the back of the case, both blowing outward through a duct? could you reduce both fan speeds accordingly to balance? I assume they'd need to run at approximately the same CFM, or maybe slightly more at the rear, to avoid backpressure.

puff
Posts: 38
Joined: Fri Aug 15, 2003 1:56 pm

Post by puff » Thu Oct 09, 2003 3:41 pm

wumpus wrote:Not the Alpha PAL series.
A PAL isn't that structurally different from other heatsinks, doesn't it perform a few degrees cooler with the fan blowing downwards? I don't own one so I wouldn't know, but logic dictates blowing against the fins will provide more turbulence around them than just sucking air. That's especially the case when there isn't much airflow, because I'm sure the delta they use in the old PAL reviews performs just as well sucking or blowing. Dan's data gets better perfs sucking, but they do suck through that metal jacket around the HS. Doing something similar with a duct would provide way too much resistance/wooshing for quiet cooling imho, unless you have a really low-dissipation proc.
wumpus wrote:Well, you bring up a good point -- interior fans vs. exterior fans. Some natural attenuation will occur with the interior fan.

What about mounting both fans in parallel, one on the cpu, one on the back of the case, both blowing outward through a duct? could you reduce both fan speeds accordingly to balance? I assume they'd need to run at approximately the same CFM, or maybe slightly more at the rear, to avoid backpressure.
A fan inside your case (ie CPU fan) is attenuated, yes, but a duct doesn't help with attenuation that much. You have to pad the insides otherwise it resonates (not much at all, but it can increase the pitch of the fans a bit). For a padded duct to be any useful it'd have to be 20-30cm. Not practical with current ATX cases. Also, don't forget panaflow L fans make most of their noise by airflow wooshing. A duct doesn't help much with that.

Again, you will have to increase fan speed (comparing to a freeflow 80L) to get similar cooling by sucking on the CPU heatsink. Since you also have one ducted exhaust at that point, you end up worse than just having 1 CPU/1 exhaust in freeflow. I invite you to try for yourself like I did :)

A 80L running without backpressure is so quiet if you set it right, that I think for low-airflow situations a duct isn't worth it. This is just my experience with ducts, obviously some think differently (considering ducts' popularity).

wumpus
Patron of SPCR
Posts: 946
Joined: Sat Sep 06, 2003 9:57 pm
Location: Berkeley, CA, USA
Contact:

Post by wumpus » Thu Oct 09, 2003 7:42 pm

FWIW I have personally owned two Alphas, and both performed better in the "blow" configuration. I can't remember any review where they didn't perform better in this configuration, but I don't have a photographic memory or anything ;)

EDIT: Damn it, I meant "suck" or sucking air away and out from the HS.
Last edited by wumpus on Fri Oct 10, 2003 1:41 pm, edited 1 time in total.

silvervarg
Posts: 1283
Joined: Wed Sep 03, 2003 1:35 am
Location: Sweden, Linkoping

Post by silvervarg » Thu Oct 09, 2003 11:20 pm

With conventionally designs heatsinks blowing on them will create higher air pressure (=higher density). Higher density means that more heat can be transfered, so it will give slightly better cooling.

Both blowing and sucking on a conventional heatsink will create lots of turbulent air in the case. This means some air goes around in a loop and reenters the CPU fan. The exact paths for the turbulent air depends both on your heatsink and the rest of your case, including the amount of airflow through the case.
By ducting air it to the CPU (blowing fan) or from the cpu and out of the case (sucking fan) will drasticly reduce turbulence. The ducting will affect cooling a lot more than chosing between blowing and sucking fan.
A ducted blowing fan will lower CPU temps more since cool air will always reach the CPU.
A ducted sucking fan will mainly lower case temps since heat from the CPU is blow out of the case. Lower case temps will indirectly lower CPU temps.
Best of both worlds would ofcourse be ducting air both in to the CPU and then out of the case. This is my plan for next computer build project.

pangit
Patron of SPCR
Posts: 445
Joined: Wed Feb 05, 2003 1:48 pm
Location: Sydney, Australia

Post by pangit » Fri Oct 10, 2003 1:07 am

silvervarg wrote:Best of both worlds would ofcourse be ducting air both in to the CPU and then out of the case. This is my plan for next computer build project.
Unfortunately that's something that's difficult with the ATX case design, given the CPU's proximity to the PSU and output ports. It's something I've looked at myself (I currently have an intake duct directly onto the CPU fan from the side of the case, and that helps quite a lot). The next thing I want to try is to put a cardboard shroud around the heatsink to direct the air into the bottom of the PSU, which should keep the case temps down. It's not ideal as the PSU is already pretty hot, and I don't know if it will actually help the CPU temps/fan noise at all.
[size=75]pangit

[url=http://www.mypage.tsn.cc/pcsilencing]Pangit's PC silencing project "on the (not so) cheap"[/url]
[/size]

puff
Posts: 38
Joined: Fri Aug 15, 2003 1:56 pm

Post by puff » Fri Oct 10, 2003 1:55 am

Let us know how your ducting project turns out. Hope you don't fail pathetically like myself ;)

I know that ducting theoretically "arranges" the flow to be straighter and less turbulence-prone, but that doesn't really show in real-life listening tests. I'd still be pleasantly surprised if you end up with less noise or even equal noise (for a given CPU temp) than using freeflow ~6V 80Ls.

Good luck on your project!

silvervarg
Posts: 1283
Joined: Wed Sep 03, 2003 1:35 am
Location: Sweden, Linkoping

Post by silvervarg » Fri Oct 10, 2003 5:28 am

Puff:
Let us know how your ducting project turns out.
I'd still be pleasantly surprised if you end up with less noise or even equal noise (for a given CPU temp) than using freeflow ~6V 80Ls.
Well, I aim much more silent than a few 80L@6V. I aim at no fans moving.
The hardest part to find to get started seems to be the Thermal transtech NPH-2 HSF.
Has anyone seen a shop selling this cooler? I would prefere to buy from an EU shop (keeps shipping+VAT down), but I am open for any suggestions.

Oh, one more thing. I will certainly need a dremel for this project. And quite a lot of time also. If/when I get started I will defenatly post (in the Cooling section).

Rory B.
Posts: 186
Joined: Wed Jun 04, 2003 3:09 pm
Contact:

Post by Rory B. » Fri Oct 10, 2003 7:53 am

pangit wrote:Unfortunately that's something that's difficult with the ATX case design, given the CPU's proximity to the PSU and output ports. It's something I've looked at myself (I currently have an intake duct directly onto the CPU fan from the side of the case, and that helps quite a lot). The next thing I want to try is to put a cardboard shroud around the heatsink to direct the air into the bottom of the PSU, which should keep the case temps down. It's not ideal as the PSU is already pretty hot, and I don't know if it will actually help the CPU temps/fan noise at all.
Sounds like the new BTX form factor may be for you. It's got a 92mm intake fan that blows directly across the processor.

wumpus
Patron of SPCR
Posts: 946
Joined: Sat Sep 06, 2003 9:57 pm
Location: Berkeley, CA, USA
Contact:

Post by wumpus » Fri Oct 10, 2003 1:42 pm

FWIW I have personally owned two Alphas, and both performed better in the "blow" configuration.
Yes, I am a moron. I meant SUCK, or drawing air away from the heatsink. I guess I picked the wrong day to stop sniffing glue..

Gooserider
Posts: 587
Joined: Fri Aug 01, 2003 10:45 pm
Location: North Billerica, MA, USA
Contact:

Post by Gooserider » Fri Oct 10, 2003 9:08 pm

Assuming proper ducting, etc., given any fixed resistance, any given fan will move more air volume when sucking than it will while blowing.

Thus if one wants to move a certain set CFM, the sucking fan should be able to run at a lower voltage.

However whether a fan will COOL more effectively is more a function of the heat load and the design of the cooling fins, ducting, etc.

Gooserider
Building Dual Athlon MP system, Tyan mobo, U160 drives, Server Cube case, Linux ONLY, lots of other goodies. Will water cool, attempting to make as silent as possible.

splice
Posts: 9
Joined: Sat Oct 11, 2003 11:37 am
Location: westwood, california

sucking, blowing

Post by splice » Sat Oct 11, 2003 9:52 pm

geez by now this thread must be indexed on every porno search engine in existence



can anyone recommend a SIDE-mounted low-noise cpu fan?

silvervarg
Posts: 1283
Joined: Wed Sep 03, 2003 1:35 am
Location: Sweden, Linkoping

Post by silvervarg » Sun Oct 12, 2003 7:32 am

Splice:
can anyone recommend a SIDE-mounted low-noise cpu fan?
Thermal transtec NPH-2 (or NPH-1 for P4 verson).
However if seems to be difficult to find a place that sells them.

Rusty075
SPCR Reviewer
Posts: 4000
Joined: Sun Aug 11, 2002 3:26 pm
Location: Phoenix, AZ
Contact:

Post by Rusty075 » Sun Oct 12, 2003 7:46 am

The aerocool DP-101 is supposedly fairly quiet...no firsthand experience with one yet though.
[size=75][b]Senior Contributing Writer, SPCR[/b][/size]

pangit
Patron of SPCR
Posts: 445
Joined: Wed Feb 05, 2003 1:48 pm
Location: Sydney, Australia

Post by pangit » Mon Oct 13, 2003 4:08 am

Rory B. wrote:Sounds like the new BTX form factor may be for you. It's got a 92mm intake fan that blows directly across the processor.
splice wrote:can anyone recommend a SIDE-mounted low-noise cpu fan?
You would think that the BTX form factor would require side mounted CPU fans, no? It should certainly work better with them, so maybe we'll start seeing them coming from the big heatsink manufacturers?
[size=75]pangit

[url=http://www.mypage.tsn.cc/pcsilencing]Pangit's PC silencing project "on the (not so) cheap"[/url]
[/size]

Post Reply