Round or Square duct ?

Control: management of fans, temp/rpm monitoring via soft/hardware

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
Friend of SPCR
Posts: 1346
Joined: Tue Jul 08, 2008 4:53 am
Location: CT

Round or Square duct ?

Post by frenchie » Wed Sep 17, 2008 9:14 am

Hi all,
I've been playing around with ducts (and finally getting rid of them... anyways...) and I'm wondering if having a round-section duct or a square-section duct makes any difference in either airflow or noise ???
I tend to think that a square section leading to a round propeler is a waste of space and it creates more turbulence and therefore noise. On the other hand, reviews of round shaped commercial ducts indicate that the airflow is also impeded...So is a round tube-like shape really better ?

Any thoughts or direct experience on that ?

Friend of SPCR
Posts: 22
Joined: Mon Aug 20, 2007 1:07 pm

Post by Dazrin » Wed Sep 17, 2008 3:11 pm

If the round duct diameter is the same as the height/width of a square duct, there will be more pressure drop through the round duct just because of the extra area in the corners of the square duct. But you are right, there will be more turbulence (and maybe noise) because of the transition into the fan.

I am feeling nerdy today (and I am bored at work) so here goes...

I don't have a fan to measure here, so the dimensions are nominal. Assuming 50 CFM of air, here are some estimates...

Square - 120 mm X 120 mm (4.72"x4.72")
Velocity = 323 FPM
Pressure Drop = 0.043" w.c./100 ft

Round - 120 mm dia (4.72" dia)
Velocity = 411 FPM
Pressure Drop = 0.078" w.c./100 ft

So, for just the duct, those extra corners of area really add up.

But, since in a computer we are dealing with very short distances, the fittings will account for most of the pressure drop.

If we compare a 1 ft section of duct (pressure drop = 0.00043) with a transition from square to round (with a 1/8" radius for the fan inlet, pressure drop = 0.019), the transition has about 45 times more pressure drop across it than the duct does.

So, I guess that means that unless you have a LOT of duct, the round duct would be better.

Posts: 7681
Joined: Mon Dec 09, 2002 7:11 pm
Location: Maynard, MA, Eaarth

Post by NeilBlanchard » Wed Sep 17, 2008 3:43 pm


Round is more efficient, because there is a lot of "friction" in the corners.

Friend of SPCR
Posts: 1346
Joined: Tue Jul 08, 2008 4:53 am
Location: CT

Post by frenchie » Wed Sep 17, 2008 11:45 pm

thanks for your input :)

Darzin, thanks a lot for those numbers, it's impressive :shock: !!!! you got me convinced :D
And that such a reknown member of SPCR (NeilBlanchard) gives the same answer means a lot :)


Posts: 2831
Joined: Wed Feb 04, 2004 2:12 am
Location: USA

Post by m0002a » Thu Sep 18, 2008 12:07 am

Whether or not the duct surface is smooth or rippled (as in a flexible duct) probably makes more difference than the shape in terms of friction. But I doubt that it is anything to worry about.

Posts: 91
Joined: Thu Jul 22, 2004 9:00 am
Location: Southern California

Post by TD22057 » Thu Sep 18, 2008 9:33 am

Using some lessons from woodworking dust collection...

I'd guess you can get some improvement (in airflow - maybe not silence) by working on the transitions. A flat surface like the outside of the computer with a hard edge that is connected to a duct will impede the airflow. A much better design is to have a gradual transition w/ a large as radius as possible into the duct.

Does it matter in a computer? No idea and I don't have the equipment to test it. With power tools, you can possibly see a 5-10% airflow change when switching to a gradual entry into the duct.

Is it practical in a computer? Probably not. It would hard to enlarge all the fan entry holes enough to get any kind of gradual entry into the duct.

Posts: 1705
Joined: Fri Jan 23, 2004 7:50 am
Location: Baton Rouge, Louisiana

Post by IsaacKuo » Thu Sep 18, 2008 11:04 am

My preference is to avoid ducts altogether. Instead, I use partitions to make airflow go where I want. A duct is like a tunnel, but a partition is like a wall with an open doorway.

A partition splits up the case into two halves; so long as all of the "intakes" are on one side and all of the "exhausts" are on the other side, then airflow will be forced through the open doorway. That doorway is where my CPU heatsink is.

I don't have any quantitative measurements, but I found switching to partitions instead of ducts improved airflow and reduced airflow noise.

Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz

Post by ~El~Jefe~ » Thu Sep 18, 2008 8:14 pm

round ducts are massively more efficient than square. square is only used in things like buildings that try to maximize volume of air for a cramped space. the space is square so the duct is. In a free space, circular with the least bends is best.

for anything.

Post Reply