WEBVTT 00:00:00.080 --> 00:00:05.839 it's been over a year since the last episode of sketchy heatsinks 00:00:04.480 --> 00:00:11.840 this project was actually probably alex's worst at 00:00:08.720 --> 00:00:14.920 the company it was over time over budget 00:00:11.840 --> 00:00:16.800 and the end result was absolutely 00:00:14.920 --> 00:00:20.960 abominable so i told him 00:00:18.480 --> 00:00:25.600 we're not doing sketchy heat sinks again unless you can do a better job than this 00:00:23.600 --> 00:00:31.279 which you might think is a reasonably low bar but then it took an entire year 00:00:29.119 --> 00:00:35.680 to bring another concept to the table so you're already over time and over budget 00:00:33.840 --> 00:00:39.560 but to your credit this does look 00:00:37.280 --> 00:00:46.079 a lot more promising 00:00:49.200 --> 00:00:54.239 private internet access encrypts your internet traffic and uses a safe 00:00:52.320 --> 00:00:57.440 protected ip check it out today through the link below that was just not fair at 00:00:56.719 --> 00:01:04.159 all but not fair at all what part of it was not fair first of all this is amazing it 00:01:02.079 --> 00:01:07.200 cooled the processor great but anyway what was not fair about that is that i 00:01:05.760 --> 00:01:12.640 just did not have the tools that i needed to properly do it 00:01:10.320 --> 00:01:15.840 you had a blank check to buy the tools you needed that's not an excuse okay 00:01:14.080 --> 00:01:21.759 okay let's just yeah let's yeah let's just get to let's put the past in the past 00:01:19.360 --> 00:01:24.560 all right so the plan for this one is we're actually going to put heat 00:01:23.439 --> 00:01:29.600 pipes into aluminum and then have 00:01:27.920 --> 00:01:33.759 actual aluminum fins so we're going to try to make a hyper 212 basically yeah 00:01:32.240 --> 00:01:37.920 pretty much exactly it's like largely ripped off of them if you look at the 00:01:35.680 --> 00:01:44.000 design here basically we're going to do is take one of these guys 00:01:40.960 --> 00:01:46.479 drill six holes directly down in then 00:01:44.000 --> 00:01:50.079 we're going to stick the heat pipes in bend them at 90 degrees and they'll go 00:01:48.320 --> 00:01:55.520 up through the fins which will have also six holes in them and are you gonna 00:01:51.759 --> 00:01:57.280 solder the heat pipes to the aluminum or 00:01:55.520 --> 00:02:01.439 um they're going to be thermal epoxied on you can't solder them 00:01:59.520 --> 00:02:05.759 because like pipes will explode if you eat them the only 00:02:03.600 --> 00:02:10.000 part that really sketches me out is bending the heat pipe 00:02:07.280 --> 00:02:14.239 which hopefully will be done with this but i don't know if we'll get a full 90 00:02:11.760 --> 00:02:20.400 degrees on it so if that doesn't work it's gonna have to sort of come out 00:02:16.800 --> 00:02:22.239 at two angles like a v8 or a v6 i guess 00:02:20.400 --> 00:02:26.720 that's going to make the holes through the fins a complete nightmare oh we'll 00:02:24.720 --> 00:02:30.879 have like fins on either side we'll like cut them in half 00:02:28.080 --> 00:02:34.959 and they'll like come out like this the first order of business was to see 00:02:32.959 --> 00:02:39.840 how far we could bend the tubes we ordered 10 heat pipes so we only have 00:02:37.440 --> 00:02:42.239 four chances to figure this out oh man this is 00:02:42.640 --> 00:02:51.519 this is such a terrible bender oh my god authority just stuck on itself 00:02:48.480 --> 00:02:53.360 oh this is about to get gross 00:02:51.519 --> 00:02:57.840 out of the box the guide on the bender wasn't correctly aligned how is this 00:02:55.680 --> 00:03:03.440 thing so balls so we needed to make up some of the slop to get consistent bends 00:03:00.319 --> 00:03:05.440 so the working idea 00:03:03.440 --> 00:03:10.239 to make this which goes the wrong way compared 00:03:08.080 --> 00:03:15.120 to the tubing sizes line up correctly 00:03:12.080 --> 00:03:18.920 is to just give her with the gaff tape 00:03:15.120 --> 00:03:18.920 hopefully it'll work 00:03:21.120 --> 00:03:28.930 i guess what you get for buying like a 20 thing 00:03:30.239 --> 00:03:33.840 i think that'll work 00:03:34.239 --> 00:03:36.959 will it fit 00:03:39.360 --> 00:03:46.120 so we've got the clearance there uh 00:03:42.799 --> 00:03:46.120 looks good 00:03:46.560 --> 00:03:49.840 yeah that one didn't quite work out 00:03:51.519 --> 00:03:58.959 oh that one did not work out either those are both totally mangled 00:03:55.599 --> 00:04:00.799 this one right here this one worked out 00:03:58.959 --> 00:04:04.560 so we're going 45 degrees with the heat pipe angles decided it was 00:04:02.959 --> 00:04:08.959 time to move on to creating the base that would sit on top of the CPU out of 00:04:06.720 --> 00:04:13.200 aluminum stock we have a guide for the CPU block what we 00:04:11.360 --> 00:04:17.919 really needed was a three and a half by three and a half inch square but it was 00:04:15.360 --> 00:04:21.840 cheaper to buy four inch stock so the first order of business was to cut it 00:04:19.680 --> 00:04:28.080 down then with our square cut down to the correct size we needed to drill the 00:04:23.680 --> 00:04:31.440 holes for the heat pipes to go in huh 00:04:28.080 --> 00:04:31.440 it's basically a milling machine 00:04:31.840 --> 00:04:35.840 ah that's not gonna work 00:04:39.120 --> 00:04:42.440 problem solved 00:04:45.360 --> 00:04:50.479 perfect could not have been better 00:04:49.360 --> 00:04:55.120 well it definitely could have been better but is this even going to hold this in place 00:04:53.520 --> 00:04:58.160 oh yeah it should be fine i didn't want to like take the time to 00:04:56.720 --> 00:05:03.360 measure all the holes so i just laser cut them captain should be fine everyone 00:05:00.880 --> 00:05:06.960 this this one is fine that one this 00:05:04.960 --> 00:05:11.759 is gonna go through there yep i don't think you have a plan 00:05:09.120 --> 00:05:15.440 i i was just gonna eyeball it sweet about 00:05:12.880 --> 00:05:19.680 thera look good no this way but that way that too hey look it's got a 00:05:17.280 --> 00:05:19.680 laser 00:05:20.800 --> 00:05:26.639 yeah it doesn't work well 00:05:24.479 --> 00:05:30.240 if only we had a rubber mallet prepare for some lube to go everywhere 00:05:35.680 --> 00:05:41.120 well i'm pretty sure you're not supposed to 00:05:39.039 --> 00:05:44.080 apply that while it's spinning if only we had brushes here 00:05:43.120 --> 00:05:49.199 you know i approve everything you want to buy 00:05:47.360 --> 00:05:55.680 all you have to do is order it well we also literally have brushes so 00:05:53.039 --> 00:05:55.680 i want to try it 00:05:56.880 --> 00:06:02.450 i mean you wanted something to go wrong so now we can both get our wish yeah 00:06:00.800 --> 00:06:05.660 that's lots 00:06:07.919 --> 00:06:13.280 oh yeah you want that 00:06:13.280 --> 00:06:18.240 this actually isn't as far off as i was expecting why are you expecting it to be 00:06:16.880 --> 00:06:23.039 off gee i i don't know 00:06:20.800 --> 00:06:26.240 wow these don't fit that tightly they fit tightly enough 00:06:24.560 --> 00:06:29.280 some of them are tighter than others this one's pretty loose 00:06:27.919 --> 00:06:35.840 how are you gonna get all that oil out of there i was planning on just kind of 00:06:31.199 --> 00:06:35.840 cramming it full of thermal paste and uh 00:06:38.960 --> 00:06:44.400 you can't just slap a block of aluminum on the CPU though since it will 00:06:42.639 --> 00:06:48.400 interfere with the capacitors around the socket so this side with the holes i 00:06:46.080 --> 00:06:52.639 don't want to destroy them so that's why i'm using the hacksaw for this 00:06:51.039 --> 00:06:56.440 i can see this really quickly just turning into my angle grinding 00:07:42.560 --> 00:07:49.520 with capacitor clearance cut and sharp edges removed with the sander and file 00:07:46.720 --> 00:07:52.880 the block was looking like it might work it's drill pressing time again well i'll 00:07:51.520 --> 00:07:56.960 have to let you do that yourself i have other pressing matters to attend to 00:07:56.000 --> 00:08:01.120 all right bye no satisfaction it's excellent time 00:08:00.000 --> 00:08:05.120 no no i'm still satisfied i don't need you to 00:08:03.199 --> 00:08:09.680 get off to be satisfied oh it's so much more on the edge i love 00:08:07.599 --> 00:08:13.639 it with more lube 00:08:16.800 --> 00:08:23.520 perfect i love how proud you are such a great fit cool with our plan fully 00:08:21.680 --> 00:08:26.879 worked out it was time to bend the rest of the heat pipes all right now we need 00:08:25.280 --> 00:08:31.759 to do six of these 00:08:28.560 --> 00:08:33.360 and i can make two errors during that 00:08:31.759 --> 00:08:37.440 and i don't know if i'm capable of that really these are kind of finicky and 00:08:35.599 --> 00:08:42.320 this tool sucks and i say 00:08:38.839 --> 00:08:44.159 that but i just made it look easy so 00:08:42.320 --> 00:08:47.920 we have one left over we can tell Linus to do the last one and tell them that it 00:08:46.399 --> 00:08:50.800 actually matters and see what happens 00:08:53.040 --> 00:09:00.080 nice well that was easy enough 00:08:56.800 --> 00:09:02.160 the problem that we have to go over now 00:09:00.080 --> 00:09:06.080 we have all of these sheets of aluminum yeah and they need to go on these 00:09:04.800 --> 00:09:09.519 yes so we need to drill 00:09:07.839 --> 00:09:12.880 these holes in these and we need to figure out a 00:09:11.200 --> 00:09:15.760 good way to hold these got the clamps 00:09:16.640 --> 00:09:22.240 yeah we seem to be on the same page here so do we want to do half of them like 00:09:20.399 --> 00:09:26.240 that i don't know you're you claim to be the 00:09:24.000 --> 00:09:30.080 expert that moved a fair bit while you were 00:09:27.600 --> 00:09:34.800 putting it on i don't even know how much that's going 00:09:31.519 --> 00:09:36.880 to affect the bit wandering 00:09:34.800 --> 00:09:39.839 oh that moved so much 00:09:38.800 --> 00:09:44.800 okay uh oh wow how are we gonna do this 00:09:43.440 --> 00:09:51.240 professional this is where we find out if 00:09:46.800 --> 00:09:51.240 this project's gonna be on time or not 00:09:53.760 --> 00:09:58.320 yeah it cuts through these like butter 00:09:56.959 --> 00:10:02.800 not bad they are mostly in a straight line-ish 00:10:00.800 --> 00:10:06.640 yeah i think so just some cleaning and prep was left until it was time to do a 00:10:04.880 --> 00:10:10.640 test fit and make sure it would all go together all those clear that awesome 00:10:11.360 --> 00:10:18.000 come on okay so about there 00:10:15.440 --> 00:10:21.120 the fit surprisingly was pretty good although a couple needed to be trimmed 00:10:19.600 --> 00:10:24.800 down so that we could get the maximum number of fins onto the heatsink 00:10:27.920 --> 00:10:35.680 oh nice so we got this thermally conductive adhesive from mg chemicals 00:10:33.600 --> 00:10:40.720 which hopefully when put between the fins and the heat pipes will 00:10:38.079 --> 00:10:45.839 conduct most of the heat i don't know the thermal conductivity on it isn't 00:10:42.399 --> 00:10:48.320 great but it's the best that we've got 00:10:45.839 --> 00:10:51.440 or that exists really this kind of looks like candy tuna i 00:10:50.000 --> 00:10:56.640 don't think i'd recommend putting this on your sushi sushi 00:10:53.680 --> 00:11:01.360 sushi oh man that smells like tuna too our choice of medium cure time epoxy 00:10:58.959 --> 00:11:05.440 means we'd have about 45 minutes to work with it before it cures first the heat 00:11:03.600 --> 00:11:10.000 pipe holes in the block were filled with nth-1 thermal compound and then the 00:11:07.839 --> 00:11:15.600 thermal epoxy was put on the end to hold them in place 00:11:11.120 --> 00:11:17.600 suppose it's done i guess 00:11:15.600 --> 00:11:23.040 oh my god these are like almost gone from here the heat pipes were covered in 00:11:19.360 --> 00:11:25.360 thermal epoxy and the plan anyway was to 00:11:23.040 --> 00:11:28.399 just jam the fins on i guess before i put the epoxy on i should have made sure 00:11:26.959 --> 00:11:32.160 that all six heat pipes will go into this all right 00:11:30.399 --> 00:11:36.160 to keep the fins from sticking together we resorted to some good old-fashioned 00:11:33.920 --> 00:11:40.320 canadian engineering using pieces of a cut up tim hortons box to space them out 00:11:38.320 --> 00:11:44.320 we repeated this process until we were out of fins so while you're off spending 00:11:42.320 --> 00:11:49.360 company money on lunch actually that was personal lunch with 00:11:46.160 --> 00:11:50.560 personal money same thing so i basically 00:11:49.360 --> 00:11:54.480 finished it oh finished is a word 00:11:52.399 --> 00:11:59.519 so don't move it too much it's still setting but i noticed that something i 00:11:57.440 --> 00:12:03.440 hadn't really thought of until now uh all the heat pipes are together 00:12:01.600 --> 00:12:07.040 which oh yeah probably isn't the most efficient way to do that no yeah it's 00:12:05.360 --> 00:12:10.639 going to be like right here is where all the heat gets 00:12:09.200 --> 00:12:15.920 dissipated and then this is all just worth it yeah like probably the 00:12:12.480 --> 00:12:18.800 like outer inch or so is doing nothing 00:12:15.920 --> 00:12:21.519 so aluminum here doesn't help us a whole lot it probably should have been copper 00:12:20.160 --> 00:12:26.720 but it would have been way more difficult to machine and stuff you're 00:12:23.440 --> 00:12:26.720 calling what you did machining 00:12:26.880 --> 00:12:33.200 yes at this point all that was left to do was leave it to cure overnight and 00:12:31.120 --> 00:12:38.160 gather some comparative benchmark data so we could determine just how good or 00:12:35.440 --> 00:12:41.680 bad our creation really is it's really dumb that Windows doesn't automatically 00:12:39.680 --> 00:12:44.720 restore connection to network drives all the time i mean far be it from us to 00:12:43.360 --> 00:12:47.519 criticize other people's products at this moment 00:12:46.639 --> 00:12:52.880 though we pay for their products though 00:12:50.079 --> 00:12:56.480 no one's buying this i paid for it uh you're getting paid for 00:12:55.760 --> 00:13:00.040 it the total cost of this monstrosity 00:12:58.959 --> 00:13:07.360 was 107.38 which is canadian dollars so that's two 00:13:04.560 --> 00:13:11.120 dollars less than an hd15 and let's find out how long it takes to 00:13:08.959 --> 00:13:15.120 mount i guess now you've stolen some mounting hardware i see here this looks 00:13:13.040 --> 00:13:18.399 like uh Noctua mounting hardware that's sort of cheating that's revised to 00:13:16.959 --> 00:13:23.760 everyone though so i think it's okay and it's the first one that wasn't mounted with zip ties so 00:13:21.519 --> 00:13:29.440 it's a pretty big improvement what we're looking for so Intel stock 00:13:25.760 --> 00:13:31.120 heatsink okay took 40 seconds to mount 00:13:29.440 --> 00:13:36.560 okay the cooler master hyper 212 took 9 minutes 00:13:33.680 --> 00:13:41.120 and 30 seconds okay so that's what we're looking to beat so we start the time 00:13:38.480 --> 00:13:44.320 right now um sure this is kind of cheating because we're using two people 00:13:42.399 --> 00:13:47.360 but i think this thing needs every advantage it can get hold on no i don't 00:13:45.839 --> 00:13:51.440 have to hold it from the bottom actually and of course this is going to be a hard 00:13:48.959 --> 00:13:55.760 mount now not a spring mount because our block is so thick these 00:13:53.360 --> 00:14:00.160 aren't going to reach oh did you check you didn't check 00:13:57.760 --> 00:14:05.760 i might still have an old school water block mount in my office this is 00:14:02.560 --> 00:14:07.040 like training wheels off CPU mounting 00:14:05.760 --> 00:14:10.880 we could probably go a little tighter but it's fine 00:14:08.880 --> 00:14:15.519 no we're not going to use a blowy on this that's cheating well i figured that 00:14:13.199 --> 00:14:20.959 if that doesn't work i think rubber bands is actually probably a better bet 00:14:17.680 --> 00:14:24.079 yeah sure good enough all right it's 54 00:14:20.959 --> 00:14:26.959 seconds still being hyper 212 00:14:24.079 --> 00:14:31.760 so the hyper 212 idles were like 30 40 degrees the Intel stock was 00:14:29.360 --> 00:14:36.959 regularly getting up to 60. it's still 30 degrees it's not going up 00:14:34.079 --> 00:14:40.639 yeah so do we hit it with a load it's kind of cheating to put like a 30 00:14:39.199 --> 00:14:46.800 fan on it too i realized that but i just wasn't gonna mention it 00:14:45.519 --> 00:14:51.199 like we definitely could have just like c and seed out the whole block 00:14:49.600 --> 00:14:55.240 got the pipes professionally done it wouldn't even cost that much probably 00:14:53.120 --> 00:14:56.480 would have cost less than a drill press 00:14:56.480 --> 00:15:03.920 there's no fun in this exactly that's why it's our job to do it wrong it's not 00:15:00.639 --> 00:15:06.639 even warming up like our heat pipes with 00:15:03.920 --> 00:15:10.959 their lesser bend may actually have some kind of 00:15:08.959 --> 00:15:15.519 efficiency improvement in terms of their evaporation and condensation another 00:15:13.120 --> 00:15:19.199 thing that we haven't really considered is that even though our heat pipes look 00:15:17.120 --> 00:15:23.279 like they're the same diameter they might be like a high quality heat pipe 00:15:21.680 --> 00:15:27.839 so we don't know if these are good heat pipes or not i would say that they are 00:15:24.720 --> 00:15:27.839 they cost like 80 bucks 00:15:28.720 --> 00:15:34.720 and you were just like ruining them all right hit me with the flear 00:15:32.959 --> 00:15:38.320 i want to see this first whoa it's really cool yeah the hottest 00:15:37.440 --> 00:15:44.160 part is the fan hub yeah like if you look at the motherboard 00:15:42.399 --> 00:15:50.639 i mean there's the coolest spot is our heatsink the base of our heatsink 00:15:47.440 --> 00:15:52.959 is tops like 31 degrees as measured by 00:15:50.639 --> 00:15:59.759 this you compare that to like this capacitor 00:15:54.639 --> 00:16:03.360 at 35 so CPU score is seven 00:15:59.759 --> 00:16:05.040 like 73 26 73 26 how does that compare 00:16:03.360 --> 00:16:08.079 we're right up there with the hype same as the hyper 212 00:16:06.560 --> 00:16:12.000 i mean there's plenty of time for it to buckle under a sustained load i guess 00:16:10.800 --> 00:16:16.399 oh okay so that jumped right up to 83 00:16:14.160 --> 00:16:20.800 degrees on the hottest core so far still turboing huh 00:16:18.240 --> 00:16:25.680 full turbo no throttle oh it's still climbing though 86 degrees yeah so our 00:16:23.680 --> 00:16:29.680 victory there might have just been purely on mass alone so i have a 00:16:28.079 --> 00:16:36.639 solution to this what's that we call it a gaming heatsink 00:16:34.639 --> 00:16:40.800 then when people say well you know under a sustained all-core load it actually 00:16:38.959 --> 00:16:45.360 doesn't perform as well we say well it's designed for gaming so if we had spread 00:16:42.880 --> 00:16:50.880 out our heat pipes more which we knew yeah was better that's why they design 00:16:48.320 --> 00:16:54.320 it like this so that way that little area around the heat pipe 00:16:52.800 --> 00:16:59.759 right in the path of the most optimal parts of the fan 00:16:57.519 --> 00:17:03.759 is exactly right yeah and that's why it ends like a fingers width away from the 00:17:02.000 --> 00:17:06.799 heat pipe because you're just nothing's happening if you have more 00:17:05.520 --> 00:17:12.480 aluminum here all right so it hasn't thermal throttle though or it's thermal throttle but it 00:17:10.319 --> 00:17:16.799 hasn't it has it doesn't it's not turboing anymore it's actually no you 00:17:14.480 --> 00:17:21.280 know it's still above base clock 3.85 00:17:18.319 --> 00:17:24.240 is above the base 3.7 clock speed of an 8700k 00:17:23.199 --> 00:17:28.240 so do we declare victory 00:17:26.319 --> 00:17:33.440 it's better than Intel stock it's worse than a hyper 212 so assuming 00:17:31.200 --> 00:17:39.360 that you have a half a dozen heat pipes some blocks of aluminum and a bunch of 00:17:36.320 --> 00:17:41.520 aluminum sheets a drill press some 00:17:39.360 --> 00:17:45.840 thermal epoxy and a lot of time on your hands 00:17:42.880 --> 00:17:49.760 you can conceivably have a better cooler than the Intel stock cooler that your 00:17:47.520 --> 00:17:54.000 CPU shipped with this one didn't though by following this method mission 00:17:51.440 --> 00:17:57.679 accomplished boys speaking of missions being accomplished 00:17:56.080 --> 00:18:04.000 thermal takes mission when they were created the vue 71 snow was to make a 00:18:01.039 --> 00:18:10.080 four-sided tempered glass case and they pulled it off it supports mini itx micro 00:18:06.480 --> 00:18:12.640 atx atx and the atx motherboards it has 00:18:10.080 --> 00:18:17.280 four tempered glass panels and a dual swing door it allows vertical mounting 00:18:14.799 --> 00:18:21.600 options for two and a half slot GPU solutions so even massive graphics cards 00:18:19.919 --> 00:18:26.559 with the riser cable sold separately it's got three-way radiator support so 00:18:23.520 --> 00:18:29.440 top front or vertical and it includes 00:18:26.559 --> 00:18:33.120 two of their ring 140 millimeter white led fans and you can check out the full 00:18:31.120 --> 00:18:38.320 details at the link in the video description 00:18:35.520 --> 00:18:41.440 so thanks for watching you guys i know it was painful if you dislike this video 00:18:40.000 --> 00:18:45.760 you can hit that button but if you liked it hit like get subscribed or maybe 00:18:43.679 --> 00:18:50.240 consider checking out where to buy the things that are actually any 00:18:48.480 --> 00:18:53.919 good that we feature at the link yeah video description 00:18:52.000 --> 00:18:57.919 these fans are good too also down there is our merch store which has cool shirts 00:18:55.679 --> 00:19:02.960 like this one and our community forum which you should totally join 00:19:00.880 --> 00:19:06.880 am i hot yeah yes i do it