Water Cooling Tube Size - Does it Matter?
Linus Tech Tips
·Linus Tech Tips
·2019-05-06
·
1,177 words · ~5 min read
0:00
What tubing size should I buy?
0:06
Well, even though water cooling legend Cathar
0:11
addressed this once and for all over a decade ago, it is still a topic of
0:16
debate in PC and liquid cooling groups. So, we are going to find out.
0:32
So, my answer to this question for the last six or seven years has been 3/8 in
0:38
ID. So, that's short for inner diameter, 5/8 in OD, so that's for outer diameter,
0:45
and usually from Primo Chill and
0:48
whatever color your heart desires. Uh, their tubing's pretty good stuff and
0:52
it's been my belief that any bigger than this will make no difference to your
0:58
system temperatures. However, I do have to confess that I have never actually
1:05
grabbed some big old thick old tubing
1:08
and tested this in a controlled experiment. So, it is time to change
1:13
that. First, we're going to need a few things. Intel Extreme Edition 5960X.
1:18
Check. Three different diameters of tubing,
1:23
check. A triple 120 mm radiator, check.
1:27
EK Supremacy EVO CPU block, check.
1:30
Fittings, check. And check, check. And
1:35
the optimal footwear for science. Extra
1:39
check. Yes. Pose.
1:45
All right. The last thing we need, this is a flow meter. This will allow us to
1:50
measure the flow rate of our different loops. And a shout out to the guys over
1:54
at Performance PCs for sending this over to us on super short notice. Now, as
2:00
much as I love to spend my day building, testing, and tearing down various water
2:05
cooling loops with different tubing sizes, there is a reason that I hired
2:09
writers. So, I'm going to let Jake take
2:12
care of that while I go do very important CEO stuffs.
2:16
Really? You put Stoofs in there? Yeah. No, you put Stoofs in there.
2:20
Oh, sure. It said stuff before. Okay, sure.
2:24
Whatever. Haha. The joke's on me, I guess, because
2:28
I still have to record this narration.
2:31
So, then to keep our testing relevant to the real world, all of the loops were
2:36
built into an actual case, the Fractal Design Define R5. So to establish a
2:41
baseline, we're starting with a typical configuration. So we've got our pump
2:46
going out to our CPU over to our radiator and then back to our pump which
2:51
has a reservoir built into it. The whole thing is running 3/8 in ID 5/8 in OD
2:57
clear tubing from Primochill. As for coolant, we're just using plain Jane
3:02
water, which is going to be replaced for each one of our tests. I I mean I mean
3:07
not because we would expect that new water would affect the results. It's
3:11
just because recapturing and reusing it
3:15
would actually be very inconvenient. That is how you fill a water cooling
3:19
loop. Good thing bro.
3:22
So to put a little bit of extra heat into our loop, this will help to
3:26
separate the performance between the different tubing sizes. We're going to
3:29
set it up at 4 GHz with 1.15
3:34
volts. Uh, it's not like an impressive
3:37
overclock for 5960X, but it should be stable. And really, that's all that we
3:42
need for this kind of an experiment. All right, so let's get IDA going. Uh,
3:47
stress FPU and bippity boppity, we will
3:50
be back in a while. We ran each of our
3:54
tests for a good chunk of time. Start a 15-minute timer. And then we
3:58
reset the statistics, collecting them over a 3minut window to get meaningful
4:02
averages. In between our 50% and 100%
4:06
pump speed tests, we allowed a grace period of 10 minutes with the fans on
4:10
full blast to cool our water back to ambient. After the first set was done,
4:14
the process was repeated with 12 mm inner diameter tubing and then 13 mm
4:19
inner diameter. Okay, so uh Wow, you
4:23
made quite a mess here. Yeah. Okay. How are our results then? Well, they're
4:26
a little depressing. They're all within margin of error. And I didn't even get
4:30
to use the flow meter because there's no way to get any results from the RPM that
4:34
it reads. No liters per gallon or liters per gallon. Liters per
4:38
So, it's just an RPM. Yeah, it's just an RPM. Oh,
4:41
yeah. So, what? It's just no difference. Yeah, it's it's no difference.
4:49
What? Why are you going over there? Hey, you stay away from that. What are you?
4:54
What are you? Wait a minute. Wait. Hey,
5:00
I knew it. You weren't supposed to see that.
5:04
It's the workshop. All right. So, you had your fun, Jake.
5:10
So, back in the early days of water cooling, enthusiasts would often
5:14
repurpose aquarium pumps, which are optimized for flow rate and not head
5:19
pressure. So using wider diameter tubing
5:23
could make a big difference. But thanks to more pressure optimized pumps like
5:28
the D5 and the DDC, even if this might not seem intuitive,
5:33
even dropping down to like a 10 mm diameter is a non-issue. I mean, even
5:39
1/4 in, which the most basic all-in-one
5:43
water coolers use, should not affect the performance of a modern loop in any way
5:49
other than the bend radius of the tubing. So, in most cases, it's far more
5:54
important to consider other things like the restriction of your water blocks,
5:58
any tight bends, the inclusion of restrictive fittings like quick
6:02
connects. But even in this worst case
6:06
scenario, which we actually built fairly recently, the D5 didn't break a sweat.
6:10
So the final answer then, what tubing do you buy? Well, it's the good old same
6:15
answer that we started out with. Any good quality tubing like Primo Chill or
6:19
EK in 10 mm inner. Uh that would be 3/8
6:23
in or 16 millimeter outer diameter. That's 5/8 inch for our Yankee friends.
6:27
And the thicker the wall, the kind of better because it'll help with any
6:31
kinking in your bends if you don't want to use angled fittings like we've done
6:34
on some of our builds. So, thanks for watching, guys. If this
6:39
video sucked, you know what to do. But if it was awesome, hit that like button,
6:44
uh, get subscribed or check out the link to where to buy the stuff we featured in
6:48
the video description. Also, linked down there is our merch store has cool shirts
6:51
like this one and our community forum, which you should totally join.
6:56
workshop. Get Get that out of here. I like it. It's It's pretty nice,
7:00
actually. You know what? We should probably put it up. We should put this in Luke's office. How about like right here?
7:04
Oh, man. You're just going to hold it up there for the next That seems dangerous.
7:09
How about like right there? That's pretty cool, right? In orange. Never I don't know.