Are heatpipes effective when not oriented vertically?

KMorley

Distinguished
Apr 4, 2009
2
0
18,510
I've read "CPU Cooler Charts 2008" and think that Tom's has done a great job reviewing different coolers in real-world environments. But I think the subject of heatpipe orientation was overlooked and possibly misunderstood.

Heatpipes depend upon convection and gravity for effective operation. The fluid should be in contact with the heat source (the CPU) so it can absord heat, boil off, rise as vapor, be cooled by the heatsink/fan, condense to liquid and then trickle back down to the heat source to begin the cycle again.

In my experience, the majority of system boards are mounted vertically in tower cases, which means that virtually all heatpipes are running horizontally through the heatsinks.

Regardless of orientation, the heatpipe fluid is going to collect at the lowest point. With every heatpipe cooler Tom's reviewed, that lowest point is at the CPU ONLY if the system board is mounted horizontally. With the vertical system board orientation, none of these coolers
collects the fluid at the CPU and therefore the heatpipe is nowhere near as effective as it could be.

It would be interesting to see a few of those tests repeated with the system board mounted horizontally. There's no question that the coolers will be much more effective in that orientation, but it might be surprising to see just how much of a difference it makes.

Tom's may find that some of the coolers rated poorly in the review actually perform acceptably well when oriented as designed.

Also, why doesn't someone design a heatpipe CPU cooler specifically for vertically-mounted system boards?

 

Fruity

Distinguished
Feb 20, 2009
200
0
18,680
....."To get around these problems, the vast majority of heat pipes, including all of the pipes used for PC cooling, have a wick around the inside of the tube that draws the condensed fluid back to the hot end by capillary action. Wicked tubes don't work quite as well when they're hot-side-down as unwicked tubes, all other things being equal, and they also still work marginally worse when they're upside down. But that difference is nothing to worry about; it's likely to be quite difficult to even measure, if all you've got to do it with is motherboard temperature sensors......"


(from http://www.dansdata.com/askdan00004.htm)

Assuming the article is correct, there should only be a marginal performance gain from mounting vertical.

Would be interesting to see the tests you've mentioned though.
 

Raviolissimo

Distinguished
Apr 29, 2006
357
0
18,780
definitely. i have an E6850, 3 GHz 65 nM second generation Core2Duo (if the 6600 & 6700 are first generation & the E8500 is 3rd generation.)

theoretically it should generate some heat. but with a Noctua NH-U12, it may not even need a fan. the case is vertical & the heat pipes are horizontal.
 
The wicking action of the tubes negates its orientation being a problem.

Tested that very thing about 2yrs ago using an Artic Freezer 64 Pro cooler, there was no difference in the temps idle or load, between the M/B being in a vertical or horizontal position.

If you're concerned about it run your own tests, then you'll know for sure.
 

vlc

Honorable
Feb 20, 2012
6
0
10,510
Totally agree and remember the graphics cards. They seem to assume horizontal orientation ( which is probably more common now with the popularity of tower cases ). This could explain why water-cooling is used at all. There are now 'semi-flexible' gas pipes used in the building industry. Perhaps that should be considered for PC heat-pipes.
 


I'm glad you revived this thread there is some new info since the thread was created, read question 6 in the link below.

http://www.tomshardware.com/forum/id-2196038/air-cooling-water-cooling-things.html