Swiftech H220 Video Review Redux

Linus Tech Tips ·Linus Tech Tips ·2013-05-07 · 1,693 words · ~8 min read
Floatplane YouTube

Transcript

JSON SRT VTT 121
0:00 This video is brought to you by CM Storm, preferred choice of mice,
0:03 headsets, and mechanical keyboards of Curse Gaming. Visit
0:09 www.coolermaster-usea.com/team_curse for more details. So, we made a mistake a
0:13 little while ago. We had some hiccups in our testing method. Well, no, our
0:17 methodology was fine. We just had some hiccups in our testing of Swiftex H220.
0:23 Now, it turned out there were a number of different issues, um, including us
0:28 using fan controller inline modules that
0:31 weren't compatible with these particular PWM fans, as well as the fact that our
0:36 unit was an engineering sample unit that had a bit of a problem with it. So, we
0:41 have been working with SwiftTech on what the heck happened where our test unit
0:45 performed terribly with Swift's own Helix fans and performed admirably with
0:51 third party fans like the Noctua NFF12s.
0:54 And we committed to SwiftTech as well as to our viewers to redo our testing much
1:00 more in depth and take a very close look at how the H220 performs on our test
1:05 bench with all the variables being eliminated as best as possible. So, I'm
1:10 going to start with a look at our cooling test bench. We use a closed
1:13 platform. We do not use an open test bench for cooler reviews anymore because
1:18 this is more of a realworld scenario. All right, so we use a Corsair C70 case,
1:23 which represents an above average case for cooling. Hey, there's my uh
1:27 screwdriver. I was looking for that. Dang it, Slick. Or I guess I'll leave it
1:31 here, but anyway. Um, okay. We use a real system inside
1:38 with a GTX 580 that runs firmark during
1:41 our load test so that we create heat inside the system so that it's more like
1:46 a real system would be if you're actually playing a game. We have two
1:49 fans here that run at 1100 or rather at
1:52 um crap I forget the RPM but they run at 7 volts. We have a fan back here that
1:57 runs at 1100 RPM and is held on with rubber washers. So that's why it just
2:02 popped off just now. 16 gigs of RAM. We mount our coolers in the top or the back
2:06 depending on what kind of coolers they are. You can see we've got Swifteek's
2:10 PWM splitter over here. Making sure we're running everything properly. 1000
2:13 watt power supply, 32 gig SSD cuz that's
2:16 all you really need. We collect our ambient temperatures here at the front
2:20 of the case and then we uh we log them with this and then we correct all of our
2:24 results to as if the room was 20° C. Uh,
2:28 the last thing that's of note is when we're not running fans at full speed,
2:31 we're using this variable resistor in here in order to run our fans at 1100
2:35 RPM. We can dial them into exactly 1100 RPM using software RPM monitoring in
2:40 this guy right here. However, it should be noted running PWM fans such as the
2:45 Helix ones, there could be adverse effects to the longevity of the fan
2:49 running it using voltage control. So, we're only doing that for the sake of
2:52 eliminating variables, not because we actually recommend running them this
2:56 way. All right, so back to the H220. We
2:59 did idle and load comparison. So idle is at idle, giving it time to equalize, and
3:03 load is at load, giving it, you know, 15, 20 minutes to equalize and normalize
3:08 the temperature. We ran Corsair SP
3:11 series fans, so those are on there right now. That was the last run we did, but
3:15 this is what they look like. Okay, they're quiet edition fans, not
3:18 performance edition. We ran Swifteek's own Helix fans in both push and pull per
3:23 Swift request because they ship them in pull configuration, but we test all of
3:28 our cool or they ship them in push rather. But we test all of our
3:32 configurations in pull because I believe it's better because it's easier to
3:36 clean. Whereas when you have them in push, the dust all accumulates under the
3:39 fan. So you have to take the fans off in order to clean them. Uh what we discovered is they performed exactly the
3:44 same in push and pull, which is what I said was going to happen. So, we didn't
3:48 even need a different data point on the graph to show those two results, but
3:51 they were exactly the same. So, consider these pull results or push results,
3:55 whatever you want. We ran Silverstone's new AP123 air penetrator fans with the
4:01 three different blade sizes. That should make them more optimized for quiet. Hey,
4:05 Slick. Uh, how are they in terms of acoustics? Super quiet. Awesome. Okay,
4:10 good. I mean, I guess I kind of expected that. And we ran our golden standard.
4:14 Well, our brown and tan standard, I suppose, uh, Noctua NFF12s, which are
4:20 the ones that we use as our standardized test bench for testing fans because we
4:25 believe that they're awesome. And as you're about to see, they're just freaking awesome. So, what we saw the
4:31 first time around was that at load, our NFF12s were outperforming the Helix fans
4:35 by like 10 or 15 degrees or something like that. We looked at it, we were like, "This is impossible." We retested
4:40 it and we got the same results. So, we published it, but we were wrong. We made
4:43 some mistakes. So these are the proper
4:47 results. We've tested our fans at full
4:50 speed as well as at 1100 RPM and we've
4:53 included some other coolers for reference as well. So the best
4:56 performing solution was the H220 with NFF12s running at full speed. Bear in
5:02 mind these only run at around 1,400 1450 RPM at full speed whereas the Helix's
5:07 run at around 1,800 RPM at full speed. So, the fact that they outperform the
5:11 Helix's is very admirable. The Helix's, however, did pull ahead of the air
5:16 penetrators as well as Corsair's static pressure fans, but those ones also run
5:21 at lower RPM than the Helix fans. So, the Helix fans have a very wide range of
5:25 RPM in part due to their PWM design as
5:29 well as due to the design of the motor overall, which is able to start up at
5:33 very low uh speeds as well as run at very high speeds, which is a great
5:37 flexible option. It's good to have. All right. So, at full speed, it goes 54.5
5:43 NF12, 56 H220, 58.5 for both of the
5:47 other fan solutions. Then, we took all the fans and we put them on a level
5:51 playing field, 1100 RPM only. The NFF12s
5:55 won by almost 3°, about 2 1/2°, which is
5:58 outstanding considering that they're running at the same RPM as everything
6:01 else. They're very pressure optimized. They perform incredibly well up against
6:05 a radiator in addition to being very quiet. So, we're going to be continuing
6:08 to use the NFF12s as our brown and tan standard, well, golden standard fan.
6:13 Next up was Corsair's static pressure fans, which were tied with the air
6:17 penetrators again at a lower RPM. So, those we can consider very equivalent
6:22 designs. And then bringing up the rear was the Helix fans, but they were only
6:26 one degree away from the static pressure and air penetrator fans. And we can
6:31 consider about one degree to be not that
6:34 significant of a difference in terms of performance. However, the three 2 and
6:38 1/2°ree victory is a clear victory for the NFF12s. After that, we finally find
6:44 other coolers. So, get this, you guys. The H220, regardless of which fans you
6:50 put on it, performs better than anything else because our the water 2.0 Extreme
6:54 from Thermaltake is our next best performing cooler, 3° behind, the worst
7:00 configuration for the H220. Then there's the H110, which is another almost a
7:05 degree behind that. Then there's the H100i, which is another almost degree
7:09 behind that. And then we get down to the single 120 millimeter coolers like the
7:14 H80i, or sorry, that was another four degrees or 5° behind the uh H110. Um,
7:19 and then the H80i brings up the rear at like 95°. So, not all liquid coolers are
7:24 built equal. Not all fans are built equal, but it's very clear to me that no
7:28 matter how inequally built your fan is, the H220 is the best pre-filled liquid
7:33 cooler on the market. almost no matter how hard you try to it. I mean,
7:38 we didn't it by using the included fans, but we didn't give it a
7:42 huge advantage like we do with every other cooler by putting some of the best
7:45 fans on the market on them. Thank you for checking out this corrected edition
7:50 of Lionus Tech Tips on the, you know, I haven't even held up the product at all.
7:54 So, here's the box. This is what it's actually going to look like if you buy
7:57 it. See, we have a retail sample now of the H220 from Swift. Don't forget to
8:02 subscribe to Linus Tech for more unboxings, reviews, and other computer videos. And if you haven't seen it
8:06 already, we did a build log a little while ago in a Silverstone TJ10 where we
8:10 used the H220, we expanded it. We added a GPU block as well as another radiator.
8:15 And that is the other advantage of the H220. So three-year warranty,
8:19 expandability, and the best performance on the market in spite of the fact that
8:22 it's using a dual 120 mm fan radiator instead of a dual 140mm fan radiator
8:27 like the H110, which still gets outperformed by at least four degrees.
8:32 And with the same fans, it looks like 110 is
8:37 about 8 degrees. Wow.