WEBVTT 00:00:00.240 --> 00:00:09.560 They called me a madman. It'll leak. It'll break. It's heresy. 00:00:04.880 --> 00:00:09.560 Not heresy. Science. 00:00:10.960 --> 00:00:20.080 Through the power of 3D scanning and printing. At long last, we've done it. A 00:00:16.240 --> 00:00:23.119 hardline water cooled gaming PC with 00:00:20.080 --> 00:00:25.680 impossible tubing bends. And all it took 00:00:23.119 --> 00:00:30.720 was some really cool technology, a bit of modeling, and patience. Let's start 00:00:28.480 --> 00:00:37.840 with the first one. This thing is so freaking cool. The Einscan Pro2XV2 from 00:00:34.960 --> 00:00:43.280 Shining 3D is a multi-functional handheld 3D scanner that is exactly what 00:00:40.960 --> 00:00:48.879 it sounds like. This can be used to create a digital representation of small 00:00:46.000 --> 00:00:52.559 and mediumsiz objects, just like our computer here. After calibrating it 00:00:50.879 --> 00:00:58.000 against their handydandy calibration board, it works by shining a white LED 00:00:55.440 --> 00:01:02.160 pattern onto an object, then registering any deformations of that pattern through 00:01:00.160 --> 00:01:08.240 the attached camera. Of course, because our computer is black and even worse, 00:01:04.879 --> 00:01:11.119 kind of reflective, it uh works so good 00:01:08.240 --> 00:01:18.320 out of the box. Fortunately, fixing that is as simple as picking up a can of this 00:01:14.799 --> 00:01:21.280 Asub's matte vanishing spray, which no, 00:01:18.320 --> 00:01:25.920 before you ask, will not make your computer disappear. 00:01:23.200 --> 00:01:31.200 But what it will do is help out the iron scan by covering your object in an 8 to 00:01:28.320 --> 00:01:35.840 15 micrometer layer of matte material that can even be in different colors and 00:01:33.439 --> 00:01:40.720 will sublimate away or change from a solid to a gas after just a few hours. 00:01:38.880 --> 00:01:44.560 Now that our computer is no longer reflective or refractive, we're good to 00:01:43.040 --> 00:01:49.520 take a short break and tell you about our sponsor, Acronis. Their business 00:01:47.360 --> 00:01:53.360 solution, Acronis Cyberprotect, delivers protection against cyber threats along 00:01:51.360 --> 00:01:56.720 with a full suite of backup and recovery solutions. Check it out at the link in 00:01:55.200 --> 00:02:02.600 the description and unify your business's data protection and cyber 00:01:58.399 --> 00:02:02.600 security under one platform. 00:02:09.440 --> 00:02:15.360 We did a full ShortCircuit video on the Ein scan where you guys can see the 00:02:12.959 --> 00:02:20.480 whole process, but in a nutshell, it's basically Star Trek and then you export 00:02:18.800 --> 00:02:24.319 the rough model from their software for a bit of cleanup in Blender or the 00:02:22.080 --> 00:02:27.440 modeling software of your choice. That's it. Even before cleanup, actually, 00:02:26.319 --> 00:02:32.400 unless you're doing something more elaborate, it is more than usable for 00:02:29.599 --> 00:02:37.120 some quick, simple tubing runs. At this point, all we've got to do then is model 00:02:34.720 --> 00:02:41.760 our tubes from one fitting to the next using Fusion 360, which by the way has a 00:02:39.680 --> 00:02:46.319 hobby license that is free, and then hit the print button, right? Well, not 00:02:44.080 --> 00:02:50.480 quite. Part of the point of this project was that we didn't want to bring out the 00:02:48.319 --> 00:02:54.959 saws and pipe cutters, heat guns, torches, and deburring tools. Though, 00:02:53.440 --> 00:02:59.040 one final step that we're going to add to the modeling process is to add our 00:02:57.040 --> 00:03:04.080 own chamered edge to the end of every tube so that it can slide right into our 00:03:01.360 --> 00:03:07.360 compression fittings. Now, we print. And while you wait, you can head over to 00:03:05.360 --> 00:03:12.239 ltdstore.com to check out our new Labs phase shirts and sweaters. For our 00:03:10.000 --> 00:03:18.159 tubing, we started by testing the most common 3D printing material, PLA, on a 00:03:15.040 --> 00:03:20.239 Prusso Mark IV. I mean, in theory, it 00:03:18.159 --> 00:03:26.480 should work, and you can't beat the ease of use. But as it turns out, people 00:03:23.200 --> 00:03:28.879 don't do this for a reason. The glass 00:03:26.480 --> 00:03:35.760 transition temperature of PLA is around 60° C, and our water shouldn't ever get 00:03:32.720 --> 00:03:38.239 that high. But in the event of a pump or 00:03:35.760 --> 00:03:42.959 a fan failure, it could happen, which would cause our tubing to soften and 00:03:40.080 --> 00:03:47.760 deform, turning a bad situation into an absolute catastrophe. Now we could 00:03:45.680 --> 00:03:52.239 switch to PETG which is the same material used by many hardline tubing 00:03:49.840 --> 00:03:57.840 suppliers. But solving the material issue still doesn't solve the paracity 00:03:54.720 --> 00:04:00.239 issue. This style of layered deposition 00:03:57.840 --> 00:04:04.640 means that there are teeny tiny gaps between the layers that prevent our 00:04:02.159 --> 00:04:09.599 system from being completely watertight. Now we could fix it if we really wanted 00:04:06.879 --> 00:04:13.599 to by smearing epoxy along the entire length of the tube. I mean, come on. 00:04:11.840 --> 00:04:20.000 That's a pretty goofy solution when we can do this instead. 00:04:16.959 --> 00:04:21.519 Man, these are cool. Do I get any credit 00:04:20.000 --> 00:04:24.400 for this being my idea to try? No. Justin said yes. 00:04:23.199 --> 00:04:27.919 This was my idea. What? The resin? Well, just 3D printing in general. 00:04:26.720 --> 00:04:33.280 I suggested the resin printing. Fine. Fine. Whoever's idea it was or wasn't. This is 00:04:30.560 --> 00:04:37.919 so freaking cool. Some of these are really clean. Like, look at this bend. 00:04:36.240 --> 00:04:41.120 Yeah, it's like perfect, right? It's It's almost too perfect. 00:04:39.280 --> 00:04:46.320 It's perfect. I mean, there are parts that are less perfect. Clearly, we had 00:04:43.440 --> 00:04:50.560 to do some sanding here with a resin goo kind of drip. But like, dude, have you 00:04:48.639 --> 00:04:53.040 ever tried to hardline bend a tube before? Yeah. 00:04:51.919 --> 00:04:59.520 Could you do this shape? Oh, no. Oh, no. No, no, no. I mean, look at this 90° coming out of 00:04:56.560 --> 00:05:02.320 one of the fittings. You cannot do that. You can do it and then cut down. 00:05:01.520 --> 00:05:07.360 Sure. But, but it still won't be that tight. If you have a clearance issue, you cannot 00:05:05.440 --> 00:05:11.199 achieve that radius by hand bending. Now, obviously to get things this exact, 00:05:09.680 --> 00:05:14.160 you're going to need models of the fittings that you're going into so that 00:05:12.800 --> 00:05:17.680 you can ensure it's going to line up perfectly, but you don't even need to do 00:05:16.320 --> 00:05:21.440 that. You can do a little bit of trial and error, and you can actually cut or 00:05:19.680 --> 00:05:26.800 sand these down to length just like you would with regular hardline tubing if 00:05:23.280 --> 00:05:29.759 you overdo it. SLA printing is super 00:05:26.800 --> 00:05:34.400 cool. It works by UV hardening resin in a vat. And one of the biggest 00:05:31.680 --> 00:05:38.160 advantages, aside from a lack of paracity and a glass transition temp 00:05:36.320 --> 00:05:41.759 that's higher than PLA, is customization. 00:05:39.680 --> 00:05:46.479 I mean, you could buy resin in various colors and keep them all in stock. But 00:05:44.000 --> 00:05:52.639 why do that when you can buy clear resin and then dye batches whatever color you 00:05:49.280 --> 00:05:54.800 want, or better yet, use UV reactive dye 00:05:52.639 --> 00:05:59.520 to make your tubing glow under a black light. SLA printing is also the secret 00:05:57.759 --> 00:06:05.039 to being able to achieve shapes that would be impossible with traditional 00:06:01.440 --> 00:06:07.199 manual bending. Guys, I've done a fair 00:06:05.039 --> 00:06:13.280 number of hardline water cooled systems and even I am not about to sit here and 00:06:10.479 --> 00:06:17.840 try and bend my own shell spiral. Let's see what our system looks like with 00:06:14.800 --> 00:06:20.880 something more like this. 00:06:17.840 --> 00:06:23.039 Oh, I love it so much. But Lionus, you 00:06:20.880 --> 00:06:27.120 might warn, resin is brittle. it's going to crack or even shatter under the 00:06:24.960 --> 00:06:31.759 pressure of a custom water cooling loop. Well, I mean, you're not wrong, but 00:06:29.759 --> 00:06:36.720 you'd also be surprised by just how strong this stuff actually is. We bought 00:06:34.400 --> 00:06:42.000 some Monocure 3D tough hightensile strength resin from Spool 3D Canada, and 00:06:39.120 --> 00:06:45.280 it's worked out pretty great so far. And while we of course aren't going to 00:06:43.520 --> 00:06:49.840 recommend that you do this at home because I don't want to be liable for 00:06:46.880 --> 00:06:54.319 your RTX 480 or whatever, I'm confident enough to risk my own RTX 480. And we're 00:06:52.479 --> 00:07:01.000 going to leave this computer running for 6 months or until it leaks with regular 00:06:56.960 --> 00:07:01.000 updates on Twitter and Flowplane. 00:07:02.000 --> 00:07:09.840 Oh, this looks incredible. 00:07:06.800 --> 00:07:12.560 And best part, guys, is that it looks 00:07:09.840 --> 00:07:16.560 like any other hardline build except for people are going to look at it and go, 00:07:14.400 --> 00:07:20.720 "How on earth did you do that?" Because these builds are wild. I mean, even a 00:07:18.960 --> 00:07:25.599 skilled artisan like Charles Harwood, creator of the MBX MK2 Murder Box, 00:07:24.080 --> 00:07:29.759 couldn't do something like this. Charles, that's not a challenge. Just 00:07:27.280 --> 00:07:33.919 sit down. And look, I'm not going to sit here and tell you guys that this loop 00:07:31.520 --> 00:07:38.080 has the best possible flow rate. And sure, um, some of it looks more like a 00:07:36.000 --> 00:07:42.319 silly straw or a water slide than an actual, you know, tubing job, but the 00:07:40.880 --> 00:07:47.039 point of the experiment was to find out where the limits are. And I feel like 00:07:44.639 --> 00:07:51.120 we're just scratching the surface. I mean, okay, we did kind of find some 00:07:49.440 --> 00:07:54.960 limits. Um, can you tell me what happened with this? 00:07:52.479 --> 00:07:58.879 Yeah, I basically showed up, installed a tube, it leaked and delayed the shoot by 00:07:56.960 --> 00:08:04.000 like a whole day because all of this was covered in water. So, we had a small pinhole leak. But I 00:08:02.400 --> 00:08:09.039 don't think that that invalidates the concept. Especially because looking at 00:08:06.240 --> 00:08:12.240 this tube, it's not even immediately obvious where you guys fixed it. What 00:08:10.960 --> 00:08:15.520 did you even do? Just put a little bit of resin on it or CA glue. That's a really simple fix then. 00:08:15.199 --> 00:08:20.960 Yeah. Okay. Tell me this. Knowing everything you know about this. Okay, we haven't 00:08:19.280 --> 00:08:23.599 done the durability testing yet. Would you do a system like this? 00:08:22.639 --> 00:08:27.680 Probably not. Oh, I just hate resin printing. Everything 00:08:25.680 --> 00:08:30.319 else is awesome. I just hate the resin. It's so terrible to work with. The 00:08:29.360 --> 00:08:35.120 fumes. Okay, that's why we have Justin. If someone else could print it for you 00:08:33.440 --> 00:08:38.880 if you just used like a 3D printing service. Sure. Yeah, this is actually pretty 00:08:36.959 --> 00:08:43.240 awesome. Okay, moment of truth. 00:08:43.360 --> 00:08:50.800 That's pretty fun watching their photos. That is so cool. It's the impossibility 00:08:48.320 --> 00:08:55.120 of it. Like, yeah, you could maybe create a custom mold for something like 00:08:52.720 --> 00:08:59.519 this, but it would be specific to that model of PC, it would not be fun, 00:08:57.279 --> 00:09:04.160 that particular case and damn near impossible to injection mold. But with 00:09:01.760 --> 00:09:08.800 the beauty of additive manufacturer, look at it go. 00:09:06.480 --> 00:09:11.760 To be clear, even 3D printing this, there could be some trickiness. You'll 00:09:10.240 --> 00:09:16.560 have to play around with supports. We might have to do some finishing to get 00:09:13.680 --> 00:09:23.360 it to look a little bit nicer, but the sky is the limit. 00:09:19.600 --> 00:09:26.000 We went with 10/12 mm tubing, but you 00:09:23.360 --> 00:09:31.120 could do bigger tubes, smaller tubes, and this is kind of cool. Okay, you've 00:09:28.800 --> 00:09:36.000 played with Solid Works flow simulation. Yeah. Would anything prevent you from like 00:09:34.399 --> 00:09:38.959 Okay, this is so close to being assigned to 00:09:37.760 --> 00:09:43.600 you already, and I haven't even talked to you about it yet. Would anything 00:09:41.120 --> 00:09:47.920 would anything prevent you from creating a an open loop? So you actually have 00:09:45.760 --> 00:09:52.959 like a a water slide that you like create laminer flow and like blast the 00:09:50.480 --> 00:09:55.839 water over to a receptacle and like a waterfall. 00:09:54.160 --> 00:10:01.839 Can we do the loop such that we're cooling the GPU and the CPU has an NHD15 00:09:59.760 --> 00:10:05.519 and it waterfalls over it? I mean, can we? 00:10:05.040 --> 00:10:12.800 Yeah. Okay. No promises though. I think the best part of this though is that while 00:10:09.440 --> 00:10:14.800 we did use a 3D scanner, you don't have 00:10:12.800 --> 00:10:19.600 to. You can totally do the work the traditional way by just measuring out 00:10:16.880 --> 00:10:23.839 your runs and then modeling accordingly. And if you think about it, even the 3D 00:10:21.440 --> 00:10:28.480 printer is kind of optional, right? Like we've got an Elgu Saturn 4 Ultra, which 00:10:26.079 --> 00:10:32.320 retails for around $400, but there are online services that will print things 00:10:30.160 --> 00:10:36.480 for you as long as you find something with a good amount of vertical print 00:10:34.079 --> 00:10:40.320 height for your lengthier tubes or a wide printing bed. Just make sure that 00:10:38.800 --> 00:10:44.560 you're using something that's precise enough that your tolerances aren't going 00:10:42.320 --> 00:10:49.519 to cause leaks at your fittings. So, an Elgu Mars 4 or better is what we would 00:10:47.200 --> 00:10:54.720 recommend without breaking the bank. This limits us to either smaller form 00:10:52.000 --> 00:10:59.600 factor machines though or a lot of money spent on interconnecting fittings for 00:10:56.880 --> 00:11:04.480 extensions. As for the resin itself, the price does vary quite a bit. But we can 00:11:02.320 --> 00:11:09.920 get a liter of the same stuff that we used for about $80, which will create 00:11:07.360 --> 00:11:16.480 quite a lot of tubing. For example, our proof of concept here took only 87.4 ms 00:11:13.360 --> 00:11:19.200 out of our 1 L jug, which is about $8. 00:11:16.480 --> 00:11:24.560 Now, realistically, I don't see that many folks putting this into action um 00:11:22.000 --> 00:11:30.959 outside of maybe niche small form factor builders, but just like everything in 3D 00:11:28.320 --> 00:11:36.399 printing, as the quality goes up and as the costs come down, I could conceivably 00:11:33.440 --> 00:11:41.839 see a small to midsize system integrator shipping hardline systems at a super low 00:11:39.440 --> 00:11:47.040 cost. Because most of the cost of a hardline system is not in the tubes. 00:11:44.320 --> 00:11:51.519 It's in the labor, all the measuring and all the cutting and bending and 00:11:49.040 --> 00:11:56.079 installing. Or I could even see home users sharing models with each other. So 00:11:54.240 --> 00:12:00.320 that hardline tubing could become as simple as copying your buddies case 00:11:58.480 --> 00:12:04.320 selection and build list and measurements and then just printing out 00:12:02.480 --> 00:12:08.320 the same set of tubes, throwing them in, and calling it a day. If you do want to 00:12:06.560 --> 00:12:13.360 get your own SLA printer, by the way, just make sure you take the proper 00:12:09.920 --> 00:12:16.399 safety precautions. Resin is hella 00:12:13.360 --> 00:12:19.040 toxic. Like gaming subreddit toxic. So, 00:12:16.399 --> 00:12:24.720 you want to avoid that. But don't avoid our sponsor, Charge. Their Ice Mag isn't 00:12:22.160 --> 00:12:29.120 just any regular old power bank. It's the world's first magnetic power bank 00:12:26.800 --> 00:12:32.880 with active cooling built right in. That's right. It not only runs cool, but 00:12:31.279 --> 00:12:37.839 it also looks cool thanks to a transparent window and full spectrum 00:12:34.800 --> 00:12:40.079 dynamic RGB lighting. Supported is 7.5 00:12:37.839 --> 00:12:45.200 watt max wireless charging with MagSafe and Gi compatibility with wirefree use. 00:12:42.959 --> 00:12:49.440 Or you can plug in a USBC cable and reach up to 20 watts of charging power 00:12:47.120 --> 00:12:53.360 for a quick boost. Plus, it features a unique low current mode for efficient 00:12:51.279 --> 00:12:57.360 charging without draining the bank. It's even airline safe, making it a fantastic 00:12:55.440 --> 00:13:01.440 travel companion for your adventures. Check out the Charge Ice Magnetic Power 00:12:59.680 --> 00:13:05.120 Bank at the link in the description. Man, what on earth are we going to think 00:13:02.959 --> 00:13:08.800 of next? If you guys liked this video, make sure you check out whole room water 00:13:06.959 --> 00:13:14.160 cooling. Maybe next time we do it, we will 3D print everything. 00:13:11.760 --> 00:13:18.880 God, what a time to be alive. Look at this tube. It works.