If geeks love it, we’re on it

PC airflow and heat – a cooling guide

PC airflow and heat – a cooling guide

Low temperatures are what we want. PC cases become like Swiss cheese riddled with fan holes in the quest for cooling. Is there an optimum configuration? Is one cooling fan enough or is it the combination of fans? One PC case was set aside and filled with fans to find out just what combination works best for airflow’s effect on heat.

We’ve been boring holes in PC cases for a long time and have dreamed up some pretty silly cooling ideas. Back in our “ghetto” days of modding we cooked up a duct that surely makes us shudder today. Take one part innovation and two parts cardboard….

fewmorefolds

Mix with one part masking tape and…presto…a fan duct to draw cool air in from the outside to wash over the interior components.

widefinished

The joy of designing and construction was far more fruitful than what the results showed. Our proud achievement of cardboard, tape and fan soon ended up in a strong metal container for safekeeping.

The quest for cooling sparks many a debate in the forums. Each has their own unique perspective on how many fans and where to place them. We’ve placed quite a few fans in our history. One case had over eleven 80mm. fans and two 92mm. fans not counting the heatsink and northbridge fans. We went stark raving nuts with the AMK1708 which generated a whopping 1708 CFM!

productws
productwsreararrows

Eleven 120mm. fans were found in this beast plus six other high-RPM 80mm. fans. It kept the system cool but at a cost of nearly opening up a dimensional rift in the time-space continuum. The question remains if one fan is enough or is it more? We set aside an AMK SX1000 PC case as the victim.

wssx1000

The parts and mod list

  • AMK SX1000 PC Case
  • ABIT KV7 Motherboard
  • AMD 3200+ XP Processor
  • 2 x 256 MB Corsair XMS DDR 400 Memory
  • Enermax 465 PSU
  • ATI 9700 PRO Video Card
  • Zalman CNPS7000A Heatsink
  • 60 GB Maxtor Hard Drive
  • Rounded EIDE cables
  • 2 x 80mm. front intake fans
  • 2 x 80mm. rear exhaust fans
  • 2 x 80mm. side intake fans
  • 1 x 80mm. top blowhole
  • WindowsXP

How the tests were conducted

Each test was conducted with a room temperature of 23.6 +/- 0.4 degrees Celsius. The system idled in its testing configuration for 30 minutes prior to Sisoft Sandra CPU burn-in (CPU Arithmetic Benchmark/CPU Multi-Media Benchmark) run 100% stress for 25 consecutive times. Motherboard Monitor 5.3.4.0 recorded the results at 5 second intervals. “Idle” temperature was determined by the consistent temperature reading for 20 measurements recorded prior to the burn test. “Time to idle” was determined by the amount of time in minutes and seconds from the last peak CPU temperature until the first reading of the “Idle” temperature.

Testing configurations

cleancase

The basic AMK SX1000 case is a very common and popular PC case for enthusiasts under whatever manufacturer name it happens to be.

frontonly

We installed matching 80mm. fans and drilled holes where fans weren’t meant to go. First two fans in the front.

rearonly

Then two fans in the rear.

airtop

Then a top blowhole.

sideonly

And finally two fans in the side.

airallfans

With all fans in place it looks like quite a lot of cooling power.

We’ve always been preaching the golden words of cooling; “in through the front and out through the rear…or top”. This is the golden rule when it comes to all PC cases for air cooling no matter what components are inside.

wsinsidecaseflow

But, having too much time on our hands, we questioned ourselves. After all we are the same people who came up with a 1700 CFM case and a ghetto duct made of cardboard. It only seemed appropriate that we journey down this road. We questioned ourselves as to the effectiveness of so many fans with so many conflicting airflow patterns. Could it be that more fans isn’t better? Could it be that the airflow may become quite a mess thus diminishing the overall effectiveness?

airallmess

So we spent a few days looking at thermometers and watching a PC cook itself. 10 different configurations were tested in our quest to determine how airflow affects heat.

cleancase

No fans

airallfans

All fans

rearonly

Rear only

toponly

Top only

sideonly

Side only

topfront

Top and front

topside

Top and side

toprear

Top and rear

rearside

Rear and side

frontrear

Rear and front

A word to the wise

A PC case is essentially a box. It can be safely said that there is a consistency to ALL PC case configurations; processors, video cards and drives are generally in the same area. However the specific style of PC case and type of components varies greatly from user to user. These differences will affect your specific results. CPU temperatures may be higher or lower depending on type of heatsink and processor. Another extremely important variable for testing is ambient room temperature. The temperature of a room, warmer or colder, can greatly affect personal results. We’ve intended these tests to demonstrate the combination of fans and their effect on cooling. This may not be the rules set in stone for all but they will surprise you and give you food for thought.

The results

timetoidle
cputemp
systemtemp

Surprised?

Conclusions

We would be lying if we weren’t. It was our assumption that the tests with ALL the fans in operation would produce the best results but it didn’t. Time to idle represents how effectively the configuration removes heat from the PC case. The shorter the time the better. CPU peak and idle as well as System peak and idle are easy to interpret. We would like to think that System temperature represents an average of how cool every component in a PC is.

  • The top and rear exhaust produced the best CPU and System results but nearly placed last for time to idle. We tested three times for this result as we didn’t believe the first two.
  • A single rear exhaust fan produces the best results overall. This flushes the theory of more is better right out the door.
  • A top only or top and front combination places in the middle of the pack for CPU and System peak cooling BUT does whisk away the heat in a very short amount of time.
  • Even with no cooling fans besides the heatink…heatsink size and type of fan can deliver good results.

So there you have it. A few theories dashed upon the rock perhaps. At least for this type of PC case and components. What should be taken away from this? Quite simply that more may not necessarily be better but, for us enthusiasts, more may be cool…for looks.

Comments

  1. Silva
    Silva Excelent review, guys, much appreciated
  2. Winfrey
    Winfrey Wow for a second i almost thought MM had come back to do an article O_O
  3. Tushon
    Tushon An epic necro
  4. midga
    midga
    An epic necro
    Truth. Glad for it, though, cause now I've read it.
  5. Hashim Elhadi
    Hashim Elhadi I took it as areference each time I assemble a pc very good article
  6. Xaevryn
    Xaevryn I guess great minds think alike. My setup is similar (though not exactly the same) on my PC. I set it up under the guidelines of what would be the most efficient way to cool it and maintain airflow. I think I'll use what's in the link to modify my setup a little bit.
  7. fatcat
    fatcat this.

    image
    image

    end game
  8. Ryder
    Ryder That is the most ridiculous cooler ever.
  9. fatcat
    fatcat
    That is the most ridiculous cooler ever.
    it's an Noctua NH-D14 with a 3rd fan, that's all

    image

    yanno, the best air cooling heatsink to date...
  10. Ryder
    Ryder Best and most ridiculous at the same time... Ok.
  11. doabarrellroll
    doabarrellroll You mean to tell me, I could be the first person to install an intercooler on a PC?

    That picture reminds me of the 3.2L flat six on the Carrera 3.2.
    image
  12. qwerty
    qwerty would love to see a test with bottom + top
  13. Curtis
    Curtis I'm no expert on cooling but wouldn't having top and rear only result in a negatively pressurized case and excessive dust accumulation in the case cracks? If you left all the unused ports open then that would probably not be an issue.

    Also, did you cover the not in use ports? There becomes a much greater number of variations if you leave some uncovered without a fan vs covering all not in use. If you left them uncovered then it wouldn't be suprising that the two exhaust fan option at top and rear would be best, air would be flowing more quickly through the case due to lack of back pressure due to the large number of inlet ports. Since your top fan is so far forward (as compared to a bottom mount PSU) it's not suprising it took longer to stabilize temperature. Any flow from the front or bottom intakes would get distributed along a path between them and the top port, which pretty much doesn't intersect any of the components your trying to cool but pulls accumulated hot air from the top of the case.

    Also on a bottom mount PSU I bet you'd get much better flow across the components since both the back and top would be closer to the corner opposite the front and bottom intakes.
  14. Jon
    Jon The test needs to explain how much cfm intake and exhaust in each example. Positive vs negative pressure is an important point not covered. Thank you very much for doing this experiment, it's helps answer some questions, leaves some unanswered. Great start though, I wish I had a bunch of fans and some tools to drill holes in my case.
  15. primesuspect
    primesuspect Ten years and still going strong!

Howdy, ! Got something to say?