PC Build in a Fridge - Does it Work??

Linus Tech Tips ·Linus Tech Tips ·2016-05-06 · 2,250 words · ~11 min read
Floatplane YouTube

Transcript

JSON SRT VTT 181
0:00 all right we're locked out of our own video
0:06 but this video is going to be great we are going to be building a computer
0:10 inside a mini fridge we are going to answer the question once and for all is
0:15 it a good idea to build a computer inside a mini
0:19 fridge like the video if you think that's a good question to answer I
0:24 didn't have any other teasers ready okay I need a key
0:37 Corsair delivers real Mech or oh here I have one H real Mech or nothing with
0:41 their new strafe mechanical keyboard featuring genuine german-made Cherry MX
0:45 key switches click on my chin to learn
0:48 more I'm going to move my chin now just to throw you for a
0:51 loop this may not look like a computer
0:55 case to you nor does it look like a
0:58 computer case to me because it's not a computer case this is a fully functional
1:06 that is to say working but not plugged in mini fridge and if you're like me and
1:13 you've trolled many a forum you've either witnessed or come up with the
1:18 idea on your own of building a computer inside a mini fridge where well hey it's
1:23 cold in there must be great for overclocking right let's put it to the
1:29 test do you even lift bro yeah I lift all
1:37 right okay now before you go full bore
1:40 on any kind of project you want to at least validate the concept so we'll be
1:44 taking the Zotac ZBOX Pico Pi
1:49 320 and we will be experimenting to find
1:52 out if this Compact and power efficient
1:55 passively cooled PC can actually
1:59 experience an improvement in thermals and or performance inside the mini
2:05 fridge right but of course first we have to establish a baseline reading in our
2:11 ambient temperature test here so we're booting up the ZBOX and then we're going
2:15 to run a stress test to find out just how she runs before putting her inside
2:21 the fridge all right so we've got our Baseline numbers we're looking at
2:25 anywhere from about 46 to 49° C on our
2:28 course which puts us at about 22 to 25°
2:33 above our ambient temperature which is about 27° now let's go ahead and fire up a
2:39 stress test what is
2:44 that that's
2:49 weird what is that noise is my backpack
2:53 here it's more on shut up it got louder
3:01 maybe that's the maybe it's the low battery alert that is a pretty
3:05 aggressive low battery alert and yet apparently that's what it is apparently
3:09 that's what it was okay so with that crummy experience
3:12 out of the way it looks like our CPU temperatures have settled into 59 to 63°
3:18 and our ambient temperatures are actually a touch higher at
3:22 27.5 okay so what we'll find out now is
3:25 whether our fridge is even ready for us to put our system inside it I thought
3:31 it's supposed to be like 4° maybe that's a touch hot oh well then it is going
3:35 down very fast okay so let's go ahead and put this puppy on the Shelf then
3:40 let's start it boom okay so no real surprises here
3:45 our idle temperature test gives us anywhere from 25 to
3:50 27° our ambient temperature around the
3:53 Pico is around 5° and that stands to reason because
3:59 we're looking at about a 20° Delta between ambient and CPU temperature so
4:05 that means that at idle the mini fridge
4:09 works it cools the Pico so this is the
4:12 part where the video gets really interesting our fridge is compressor is
4:16 running all the time but it is maintaining a frosty internal
4:22 temperature as low as 2 to 2 1/ 12° C
4:26 and our CPU has settled in at around 38
4:30 to 40° C under load the mini fridge
4:35 concept is a massive success achieving a
4:40 let's see wow greater than
4:43 20° Improvement in load CPU temperatures
4:48 now we're going to take things up a notch I'm going to steal Luke's test
4:52 bench and find out if the type of rig that you would actually benefit from
4:58 cooling down for better overclock Lo ability uh will actually work in our
5:02 mini fridge setup so this is a Titan X
5:07 this is a 5930k or 5960x not sure which 16 gigs of
5:13 ddr4 memory let's go he'll never miss it
5:17 actually he'll come looking for me so now it's Baseline time with Luke's rig
5:21 here in the garage so our ambient temperatures are a little cooler than yesterday but that's okay we can correct
5:26 for that the most important thing is the difference between amb temperature and
5:31 CPU temperature when you're measuring this sort of thing and our CPU
5:34 temperature we're going to simplify things a little bit instead of looking at all the cores we're just going to
5:38 take the CPU reading and we're looking at it's fluctuating between 29 and
5:43 30° so now let's start the stress test
5:46 and see where we end up outside of the
5:49 cooling box we have our numbers so the
5:52 room cooled down a little bit thanks to some breeziness at about 22 and 1/2 de
5:57 and our CPU settled in just under 45° on
6:01 the hottest of its six cores it is a 5930k I have now confirmed that so now
6:07 it's time let's do it go ahead and start by
6:12 removing some shelves I do not recommend cooling down
6:17 hard drives um especially while they are
6:20 operating so we're not going to be putting our hard drive in there and uh
6:25 another thing this Pico now that it's cold and it's below ambient temperature
6:29 should not be powered on until it reaches ambient temperature on its own
6:33 because otherwise condensation can be a factor and any moisture with powered on
6:37 electronics is a terrible idea okay I
6:41 think we've actually got enough clearance on our fans here that they
6:44 should be able to cool the CPU just
6:47 fine there's a bit of a problem here I
6:51 have a solution
6:58 ah there we go problem
7:01 solved so we're getting ambient temperature readings the compressor is
7:07 running and we can see our idle CPU temperatures here so let's find out
7:12 where she settles in at idle bear in mind guys that we are talking a very
7:17 different level of heat load here I think the Pico has a power supply that's
7:22 something like 10 watts total whereas
7:26 even at idle this computer is going to consume somewhere in the neighborhood of
7:31 80 to 90 Watts now I was really hoping I
7:35 was rooting for it I was hoping that the
7:38 idle scenario would give us subambient
7:42 temperatures but in fact it has settled
7:45 in around 23.3 de almost exactly the
7:50 ambient temperature of the room which means we are
7:54 basically achieving nothing other than consuming electricity by putting the
7:59 computer inside the fridge but I promised you guys a load scenario and
8:05 that is what I am going to deliver it's
8:08 should be noted that I observed that to the touch the side of the fridge which
8:12 is where this particular unit seems to dissipate heat from the hot side of the
8:17 HVAC unit is hotter than I have
8:20 experienced before you can see we actually just crept up another fraction
8:23 of a degree so at 23.4 right now let's
8:27 go ahead and kick it into high gear kids okay so you can see they've already
8:33 jumped to about what we observed in our
8:37 outside control test but what this
8:41 doesn't account for is that the ambient temperature in here is going to increase
8:47 now that the compressor and refrigeration system inside here is
8:53 overwhelmed so let's watch it happen shall we all right so this is taking
8:58 longer than I thought to completely reach unreasonable temperatures and here
9:02 we are 12 minutes later the system is still running but we know enough to know
9:08 that it doesn't work this is not even an overclocked rig and our ambient
9:13 temperatures inside the fridge are steadily climbing I took a reading about
9:19 30 seconds ago it was 26 and2 now it's
9:22 peaking at 27 depending on how I jiggle the probe um so all that's really left
9:27 is the explanation for or why this
9:30 doesn't work well there's no airflow this is a closed system so this system
9:36 relies on itself to move any heat that's
9:39 generated inside out so that's going to be done through the sides of the unit or
9:44 in many cases on fridges a coil at the
9:48 back a heat sink effectively so unless you're using a unit with enough capacity
9:53 to handle not only keeping the insides cold but also removing a constant heat
9:58 load it is n going to work so we can't
10:02 just introduce airf Flow by like you know punching fans in the side so a
10:06 closed system depends on the capacity of
10:09 the system for cooling now with that
10:12 said you guys might go okay well hold on a second lus you showed off that the low
10:17 powered PC could work and even the idle PC was not failing at the rate that the
10:22 load PC is well that's only sort of true
10:26 because fridges are designed for occasional use some heat from the
10:31 outside will transfer inside through the insulated barrier and it'll have to you
10:36 know maintain that also the user will open it up from time to time to remove a
10:40 drink so the compressor fires up and it gets things back down to a nice cool
10:44 temperature they're not designed with components that are rated for 24-Hour
10:50 7day a week operation that type of
10:53 workload rating is only going to be found in a more industrial unit
10:57 something like an air conditioner and I guess that's where we get into the
11:02 Practical sort of side of phase change cooling for computers if you have a
11:07 large high capacity air conditioner say for example designed for a room that can
11:12 handle that kind of a heat load or something where an entire building is
11:15 air conditioned then you will get the benefit of the lower ambient temperature
11:20 not only for yourself but also for your computer also if you checked out our
11:24 phase change BC build guide which you
11:27 can check out in the little I in the corner well if you apply the phase
11:31 change cooling directly to the heat load and you're controlling the scope of what
11:35 you're cooling we haven't even fired up the graphics card in here and we're
11:38 overwhelming it then it can be useful as well but again we're talking high-end
11:44 components not a you know couple hundred mini fridge and we're talking custom
11:49 designed stuff in that case from LD cooling so yeah I mean most of you
11:55 probably knew how this was going to turn out some of you might not have cuz it's
11:58 all about how how you think about it if you think about your computer in terms of that it runs at 30 or 40° and so you
12:04 know it should be able to be kept cool then I get where you're coming from but
12:08 it's more to do with the total thermal energy not the actual temperature which
12:13 are two completely different things and if you think about it this way a
12:17 computer that consumes you know 300 400 watts at load is going to be kicking off
12:22 basically equivalent amounts of heat that's about the same as a as a modest
12:26 space heater and you wouldn't think putting a space heater inside a fridge
12:29 would keep that cool either would you so
12:32 there you go guys the answer has finally
12:36 been provided online as Tech tips not that anyone else hasn't ever proved this
12:40 before you can see in the time I've even been talking here we've gone up almost
12:44 an entire additional degree so I'm going to go ahead and put this system out of
12:48 its misery now thank you for watching if you disliked this video bam you know
12:53 what to do but if you liked it then go ahead click that like button get
12:57 subscribed to our videos even consider supporting us by buying a cool t-shirt
13:01 like this one shopping on Amazon using our affiliate code we've actually got a
13:04 video guiding you guys through how to do that or even by supporting us directly
13:08 through the Linus Tech tips Community if you guys are looking for something else
13:12 to watch now that you've done all that stuff I'm going to pull up my calendar and find out what I've done recently ah
13:18 yes Luke should have just uploaded the
13:21 Doom Hama machine where we built the ultimate possible machine out of all the
13:26 components in our office and ran it through the gauntlet for you so check
13:29 that out I'll see you guys again next time