Fan direction

You're underthinking it. Fans are all about getting air where you want it. You need to control the airflow in and out to make sure the hot things get cooled. A general rule of thumb#1 is "plugged filters don't cool nuthin'". Rule of thumb#2 is "All filters get plugged".

If possible, always preferred positive pressure.

The only time I might vary from that would be some physical reason the hottest device can be located behind some vents but the fan must be located somewhere else. Then, negative pressure will draw external air onto the heat source.

When one has a supply in a case that is to be mounted within a case, sometimes one will see the other circumstance.

Typically, however, one wants positive pressure to keep the dust build up down and more controllably located.

Dust coated heat sinks and other components don't colvery well either. One filter is easier to clean than the entire innards a PC or test gear.

Sorry. No rule. But in addition to the filter logic, think about turbulence. Air flow is air flow and logically should cool just as well going in either direction. But that flow might be relatively laminar coming from one direction and quite turbulent when leaking in through a bunch of small openings going the other way. But which way is which one can't say without inspecting the equipment.

This advice is only applicable for equipment that has been carefully designed from a thermal point of view when someone decides that reversing the fan direction isn't significant.

I agree 100% with YOUR STATEMENT.

But, if you live in the real world, outside an environment with full-time maintenance staff to clean filters, it just don't happen.

Certainly depends...but for many systems, an internal environment compromised by some dust is far better than a fan FULLY BLOCKED by that same amount of dust.

IF vendors thought differently, every home-user PC would have a fan filter.

Just for fun, I looked around the house. I found exactly three dust filters...count-em-3. One in the furnace. One in the vacuum cleaner. One in a 40 year old Tektronix oscilloscope. All should have been cleaned long before now.

Found more in the garage in IC engines.

While looking around the house, I found dust bunnies in most of the corners. Since the full-time maintenance staff got a divorce 35 years ago, there are lots of things around the house that just don't get the attention they deserve.

That was fun...

Dust is democratic. It doesn't much care what direction the air is blowing.

There is a big difference in the air flow at the inlet and outlet sides of a fan. At the inlet, air drifts into the fan gently from all directions. At the outlet, there's a focused stream of air, very directional, turbulent, often with a hole in the middle. In some situations, the focused air flow will whack one hot gadget and miss others. Sorry, it's not simple.

I usually have the fan exhaust an enclosure, and arrange air inlets to meter various amounts of air to the things that need it. But then, I don't use filters.

ftp://jjlarkin.lmi.net/Chimera.JPG

Air is the most peverse substance on the planet [1]. Air flow is usually nothing like what one expects. Things near a fan can be in still air. Flows are often in the opposite direction from the obvious. Heatsinking in an air flow is often bizarre. What we do is build physical mockups and analyze them with air flow meters and incense sticks.

John

Push the air in, and as Paul pointed out, Laminar air flow for best cooling. Or Impingement cooling for pin fin heat sinks, see

Also, inlet area should be 1/2 the outlet area.

Cheers

Em 27/12/2010 23:21, Martin Riddle escreveu:

Where in the Earth does this really happen? In order to decrease the air to heat sink resistance *turbulent* air flow must be insured.

Absolutely. Laminar flow is great for lift, not so good for cooling.

I stand corrected, 'Turbulent air flow' is correct. ( its what I was thinking, but not typing ;( )

Cheers

Years ago, we would do our own shake 'n' bake: thermal imaging analysis, temperature stressing, vibration stress, etc. But, those things were more power hungry (a few kilowatts in a box was "no big deal") and had to operate in much more hostile environments (nowadays, I want things that run *inherently* cool -- close to hermetically sealed enclosures -- and, if anything, I'll just test for drop and impact).

You underREAD my original post! :> In particular: "I'm adding a couple of (tiny) fans to some external disk enclosures and found myself facing this question, yet again."

If you think about it:

- the enclosures are designed for disk drives

- the enclosures *don't* have fans currently (so they rely on other methods to get heat out)

- the fact that I will be installing "tiny" fans suggests that the cases don't have room for bigger fans

Agreed. I find it much easier to drag each box out and vacuum the filter (usually without even opening the case) than it is to drag them out, uncable them, disassemble, vacuum interiors, reassemble, recable and slide back into place. In addition to the time involved (downtime), there's also more risk "disturbing something" if you have to open a box to clean it out.

But that's because "every home USER" is typically naive and ignorant about how to take care of (most everything!). It's the same sort of reasoning that gives them "idiot lights" instead of meaningful error messages...

Every box in my office that has a fan has a filter. (hint: I think there are currently 40 power cords "plugged in" -- though not all are *on*, of course)

I keep most of my equipment *under* my work surfaces. There are far too many things and cables under there to "vacuum regularly" (without screwing up something!). So, the machines end up acting like vacuum cleaners -- grabbing any "wild dustbunnies" that happen across their way.

Ah, seems like the problem isn't the filters but, rather, the quality of your "maintenance staff"! :>

I learned long ago that you can keep things working a LOT longer if you can keep them cool(er). So, I don't shy away from fans (I think my servers have six or more in each machine). But, with fans comes "concentrated dust" (as well as audible noise). So, either plan on cleaning out the machines "regularly" or just add filters (and wear headphones to ignore the noise).

That's actually a *great* idea! I should just go around and stick my hand behind each device and tabulate air flow direction. See if there is a pattern.

Thanks!