21
\$\begingroup\$

For a fan, should air be sucked in, or should it blow air out?

I'm talking about an enclosure mounted fan.

\$\endgroup\$
  • 1
    \$\begingroup\$ I thought you were supposed to have fans blow air inward, because otherwise they are blowing hot air over themselves, which reduces their lifetime, but I can't find any references for this. \$\endgroup\$ – endolith Nov 6 '14 at 18:51
  • \$\begingroup\$ Dust is probably the significant factor in electronics, but in other cases it is not. People prefer to have fans blowing on them, not away from them. It is generally better to blow cool air in to an enclosure, room or attic, because then the cool air hits things and cools them. If you just suck hot air out, it doesn't cool the objects inside the enclosure (room, attic) as well. Blowing in would be best, but for electronics, you need a dust filter. But I suppose you need one anyway, right? \$\endgroup\$ – user56384 May 12 '16 at 20:02
27
\$\begingroup\$

Airflow is the key. Any direction will do. Just keep in mind where the hot components inside your enclosure is.

However, if you blow into the enclosure, you have the option of putting a dust filter on your fan. Whereas if you have your fan blowing out, air will enter your enclosure through all sorts of holes, and lots of dust may eventually accumulate.

\$\endgroup\$
  • 6
    \$\begingroup\$ All old Class-A computer hardware used to have dust filters and sucking fans. Positive air pressure kept the machinery clean. But you had to service the filter. For consumer gear like the PC, they went with blow, and no filter to forget to change. Chux multicloths are AC filter cloth ( re-purposed in 'NAM IIRC) When the stripes are invisible, change the filter cloth. \$\endgroup\$ – Tim Williscroft Nov 8 '10 at 0:47
  • 1
    \$\begingroup\$ In theory you don't have to clean out a PC, but if you've ever had to deal with a chain smoker's computer... \$\endgroup\$ – Thomas O Nov 9 '10 at 13:27
  • 1
    \$\begingroup\$ Ugh, I had to deal with a chain smoker's computer once - the computer was overheating and turning off due to everything being clogged (the gunk was packed down in the blades of the heatsink too). \$\endgroup\$ – techdude Jun 27 '16 at 19:34
12
\$\begingroup\$

In my experience, the choice of direction determines where the inevitable dust build-up will occur. Dust seems to accumulate wherever the air enters the case.

With fans blowing in, you tend to get the most dust build up right in front of the fan. You wouldn't want this happening on the heat sink of your CPU, as the dust acts like an insulating blanket.

On the other hand, with fans blowing out, you get dust in the case near every possible way for air to get in. One place this can be a problem in a PC is when the air comes in through the openings for removable drives.

Probably the best design I've seen involves having two fans, one blowing in and the other out. The exhaust fan sits near the heat sink of the CPU to ensure good air flow there, w/o the dust problem, and the other fan pulls air into the case to keep the pressure in the case high enough to prevent sucking air (and dust) in through all the other openings.

\$\endgroup\$
4
\$\begingroup\$

When I design something like airflow, there are a few important points.

  1. What components generate heat, and how distributed my heat sources are.
  2. Where I can expect the user to be in reference to the device.
  3. Dust buildup.

In order. If you have single units that generate a bulk of your heat, they need something near them. You have options, a fan on them in extreme cases similar to a high end processor or a fan on the enclosure next to them. If you have generally generated heat, you just need to focus on getting general air flow.

After you know where fans are needed, you need to think about the user. Why would your o-scope have a fan on the front that blows in? It is generating a lot of heat there, but if it blew forward and blew warm air in the user's eyes all the time I doubt they would thank you. Design air to exhaust in a direction the user can be expected not to be.

Now, on dust buildup, you want to limit this. I have two directions for this.

  1. Fan with filter blowing air in. Without a filter the high speed air will pull in dust with the air, but as air slows down it will act like a river slowing down and drop silt.
  2. Exhaust in a few key locations you can pull in over a large area and expect significantly less dust to enter the enclosure due to low air speed entering the device, but you will still have dust issues. I would prefer everyone used filters, but I do not mind blowing out devices once in a while with an air compressor(just kidding).
\$\endgroup\$
3
\$\begingroup\$

If you are talking about a fan on an enclosure, then the whole point of this fan is to cycle hot air out of the enclosure and pull cool air in. You have to have the appropriate vent on the other side of the enclosure to allow this to happen. Some enclosures have fans blowing both directions to help the process along. If you only have one, make it blow out, and near the biggest source of heat.

\$\endgroup\$
  • \$\begingroup\$ But my oscilloscope's fan doesn't blow out, neither does my computer's fan... \$\endgroup\$ – Thomas O Nov 7 '10 at 22:22
  • 1
    \$\begingroup\$ @Thomas I have always seen fans blowing in on heat sinks and pulling air out of enclosures. \$\endgroup\$ – Kellenjb Nov 7 '10 at 22:37
  • 1
    \$\begingroup\$ Every PC power supply with a fan has it blowing out of the case. If yours doesn't, then it is on backwards. If you have case fans on the front, they should be blowing in, pulling cool air in, where it is exhaust out the back. \$\endgroup\$ – Brad Nov 7 '10 at 22:39
  • \$\begingroup\$ Then why does my scope's fan suck air into its power supply and onto the mainboard (acquisition area), not blow it out? \$\endgroup\$ – Thomas O Nov 7 '10 at 22:39
  • \$\begingroup\$ As Brad said, it's more about the flow of the air than the direction of the fan. Air needs to come in from somewhere if it's being blown out and to go out somewhere if it's being sucked in. \$\endgroup\$ – Mr. Hedgehog Nov 7 '10 at 22:40
2
\$\begingroup\$

All the air that gets in will get out and the air resistance is the same in both direction so it is more or less the same thing. There are however some scenarios where it does make a difference. For example, if you have:

| Low power sensible components | High power components | Fan ---> |

Then you definitely want the fan to take the air out because otherwise, the high power component would heat the sensible component.

Another aspect to keep in consideration is that air speed is faster and more directional at the output of fan and slower and more distributed at the input. If you have a single very high power component that is difficult to cool down, you may want to point airflow directly to it, even if it will cause the rest of the unit to be hotter.

| Heat resistant components | Single very hot component | Fan <--- |

\$\endgroup\$
1
\$\begingroup\$

Generally, airflow isn't reversible in electronics without ramifications.

Complex software like Flowtherm is used to pattern airflow and heat patterns for a specific air direction. Usually parts are placed to maximize cooling with specific direction and speed of air.

Going against this could potentially lead to hot spots and bad things like thermal runaway.

\$\endgroup\$
0
\$\begingroup\$

I do believe in a simple thing about cooling: it's better to cool the components instead of draining the heat. Why? Let me show you a simple reason, that was done on my lab. power supply on pictures below:

enter image description here

This is my lab bench supply with 80x80mm FAN mounted on back. It is blowing air from OUTSIDE to the INSIDE. Why? Look another photo below. It's not the best, but that's all I got on me right now.

enter image description here

If I would be sucking air from inside to the outside, the fan would suck the air from closest gaps to the fan. But instead of this, the air is being blown into the power supply across the components, which will cool them all and go through any hole in the power supply cover as it wants.

\$\endgroup\$
  • \$\begingroup\$ The air should move along the fins, not across. The second heat sink doesn't get air flow from the fan. I suspect that thermal design was an afterthought in this instrument. Let the -1 serve as a warning beacon to others. No hard feelings. \$\endgroup\$ – Nick Alexeev Feb 18 '18 at 8:07
  • \$\begingroup\$ No hard feelings for sure, I'm open for any kind of debate. Well, I see what you mean but the thing is, that the air will cool the fins of the heatsings, eventhough it don't flow directly along the fins. In this case, it's okay as it is, but for large series, I'd design it differently. \$\endgroup\$ – Jakey Feb 18 '18 at 22:27
-1
\$\begingroup\$

I think you should have a negative airflow within the case itself. Dust to me is not a problem. I have 2 exhaust fans (fans sucking out) and air entering the front of the case via the vent holes (with filters) I have never, ever had a heat problem with this configuration. This way not only is the CPU and Ram and North/South bridges kept cool but the GPU will also blow out, therefore keeping itself cool. If you live in a stupidly warm room, this might not work but I like to keep my room at about 18c.

\$\endgroup\$

Not the answer you're looking for? Browse other questions tagged or ask your own question.