Today we’re going to talk about thermal paste because I was kind of sent by a viewer something that they wanted me to check, and that is specifically this Slice Engineering Boron Nitride Paste, which is specifically used for 3D printers and like their temperature probes and hotheads and stuff for things that need to conduct heat there.
So we’re going to compare it to the Keen Pin Extreme because when we talk about thermal paste conductivity, it’s kind of nuts what this is capable of. We’re going to see whether or not it even makes a difference for computers. So when it comes to measuring thermal paste conductivity, it’s done using a formula called watts per meter Kelvin. So we’re not going to go into the formulaic equation as to what that actually means, but when we’re comparing paste conductivity, that’s the number referenced. So to put this into perspective, the KPX, or Kingpin Xtreme, which I use pretty much everywhere, Oh, by the way, it’s like you, her best friend. Don’t worry about that guy. He’s only there on the weekends. Anyway, moving on.
I’m glad she’s wearing headphones back there. So KPX, which is something I’ve been using for a long time, mostly because of how easy it is to actually apply this paste, It’s not too thick, and it’s not too liquidy. It’s kind of perfect right in the middle. In terms of Goldilocks, it’s like the just-right porridge; you can easily apply and spread it and don’t have to warm it up first or anything like that. It’s got a wattage per meter (Kelvin) of 13.8. So there you go. So when it comes to the thermal grizzly, the little bitty guy with the little glam, this is like what you would get with an AIO cooler or something like that. You can get bigger; it’s under the table now. You can get bigger tubes if you want, but that’s just one gram. And we’ll talk about that in a sec, because that is 12.5 watts per meter Kelvin. and so already lower than the KPX, right? This is 13.8. The Slice Engineering Boron Nitride Thermal Paste is 31.8. Yeah, 31.8. It is extremely high. There’s also something to point out, though: it’s got very low conductivity, and it becomes non-conductive once it’s heated up to 100 °C. So there might be a slight amount of conductivity in this. It’s also actually, like, water-soluble. It dilutes nicely in water, so it’s easy to clean up, especially with this rubbing alcohol and stuff.
I’m not too concerned about this having a slight amount of conductivity. I bet you I could put it on stuff and it would be fine. It’s just that they’re upfront in saying that it is not fully non-conductive until it’s been heated up to 100 °C. It’s easy to do that these days. Just let any motherboard take any 14-series CPU with its Intel limits removed, and there you go. There’s your 100 C. So yeah, I know that’s the dye temperature, but that will transfer as well. Feel it run long enough. So I’ve got KPX pre-installed right now, and I’ve set some parameters here to sort of lock down our testing. I’ve got a 13900K, I’ve got my voltage locked, I’ve got my frequencies locked, and I’ve got my fan speeds and my pump speeds locked. So everything is just completely level. It won’t dynamically adjust or change. We also have the temperature in this room stuck at 70 degrees Fahrenheit. So it’s less important what those numbers are as compared to what the before and after tests are. So one of the things I want to check right now is what the initial spike temperature is. Because one of the first lines of defense for your cooler being able to do its job is the fact that you have to have a thermal material in between the IHS and the cooler filling in all those microscopic voids and air pockets that will form. And that’s what thermal paste does: it fills those voids and conducts that heat because, as we all know, air is a terrible, terrible conductor of heat. So, that’s why we need to make sure that we have paste in there. So, theoretically, a paste that has a higher conductivity should more efficiently move that heat from the heat source, in this case, which is the dye, to the cooler. Now, there’s a couple layers. There’s like a cheeseburger layer. Yeah, the fat guy talking about cheeseburgers and… Is it lunchtime yet? Okay, I digress. There’s like a big Mac happening here when it comes to different layers. You have your dye, which is the actual silicon substrate, or not the substrate, but the silicon itself on the substrate. Then you have the thermal interface material in the 13900K’s instance. It’s like a soldered Tim. It’s like a soldered paste. It’s not quite soldered. It’s something they put in there and then heat from the outside, and then it solidifies. So that is moving material from the die to the IHS. Now the IHS is that metal thing that, when you look at the CPU that has all the printing, numbers, and stuff on it, is responsible for now touching the cooler. But now you have the material between the IHS and your cooler, which is the thermal interface material. So we might be moving some of those efficiencies around based on how good or how bad they are. So if you have terrible thermal interface material, all that heat is going to have a bottleneck at the thermal interface material, which means it’ll build up at the die, and then your clock slows down, your voltages go down, and your computer starts to slow down because of thermal throttling. So we want to make sure that that heat can get out of the entire CPU array of Big Mac, patty whack, give your dog a bone, whatever. I don’t know why I went there. Having a bit of a day, let’s just go with it, okay? We want to make sure we get that heat to the cooler as fast as possible. We don’t want a thermal pace to be what’s holding us back. I don’t care about the score so much unless we start slowing down on our CPU. So the package is currently sitting at 33 °C. So I’m just going to do a single run first of multi-core, and let’s see what we spike up to. So it’s 82 or 84. On the package, you can see our cores are sitting, and the P cores are in the in the low 80s, mid 80s, a couple in the high 70s, mid 70s, 83, and 84. It’s actually pretty good. We drooped down to 1.285 or so on the bid. That’s normal, because we do have a droop enabled. They give us a score of 40,451. What I’m going to do now is do a 10-minute test because I want to see what our maximum temperature is going to be as well. Theoretically, we should not see that number necessarily go up too much higher than a few degrees because, keep in mind, we do have a 360 AIO. As that water temperature gets warmer, we’re going to see the temperature climb too because the thermal capacity of the water is becoming more saturated with the heat that it’s exchanging.
That’s why I have my fan set to 75% static speed instead of 100, just because of noise for the video, honestly. But 75% on this AIO should be plenty with these Vardar fans. And then we’ll see what our maximum temperature ceiling’s at. Change the thermal paste. We’ll take you along for the ride of the application. I’ve not opened one of these yet. I bought three of them, by the way. I got them on Amazon. and I want to see how easy it is to apply and spread because it is water-based. It should be very similar to Kingpin, where it should be easy to apply. And for the Kingpin, I did not pre-spread. I just did a line down the very center of the CPU. So we want to see how it spreads as well. Okay, so our package is at 93 currently in this part of the test. It looks like it spiked 94 at least once. That was it, and our cores are sitting here in the upper 80s. So we’ve only got eight seconds left on this test, and then we’ll see our package wattage actually went up to 310, because what you may not realize is that as temperature goes up, so can power draw. Our video is still around the same at 1.27 and 1.28, so that’s why our power ended up going up, or at least our wattage, because as the temperature goes up, it becomes more inefficient, which means more power to end up maintaining those clock speeds and stuff, and because we removed all those 7 gigahertz all cores or 5.6 gigahertz all cores. And we’re at 6.1 single core right now. So my overclock is apparently doing really well for a down and dirty didn’t really try. Okay, so these are what our numbers look like here. I guess what we’re looking for now is: is it going to be any different at all? 92 to 93 °C is what we were seeing on average right there after 10 minutes. Our cooler is fully saturated. I need to let it cool down right now. You can see our score dropped by about a thousand points as well. That’s typical, too, with high heat. Even though our clock speeds didn’t change, they might be changing very quickly, faster than the polling rate of the software, where it might be downclocking and clocking back up so quickly that it’s not picked up in the polling rate, which is what will end up affecting our score. All right, so after letting it cool down, we’ll apply the boron nitride, and we’ll see if this is going to be a suitable solution for an inexpensive thermal paste. That’s the teaser here. That’s also why I had that little tube. Okay, so let’s see how easy it is to actually apply. Okay, that’s a little flaky. So let’s try this on a paper towel real quick. Oh, okay. It really squishes out. Jesus. Hold on. Okay, so I’m trying to apply this about the same way I did the KPX. This is actually quite a bit thinner than I was expecting. There, I put a lot. Because I don’t want to, I need to be careful with the squish out because it needs to be hot to not be thermally conductive or electrically conductive. Yes, we want thermal conductivity. We don’t want electrical conductivity. Okay, so theoretically, if it’s squished out, it still should not have killed my computer if it did. Oops, there we go. I forgot to plug in the pump. Okay, we got it to 100 C. Look how fast I came down! Okay, so I was like, Oh, it’s at 100 C already. I forgot to plug in the pump. Guess what? I think we made it; we did the dehydration process. Oops. I was wondering, How am I going to make you hit 100 C? Like, should I go with the crazy overclock? No, forget the plug in your pump. That’ll do it. Okay, let’s clear those maxes. Pro tip: if you want to throttle, unplug your pump. Just one run. What do we have right now? So we’re at 61, 65, 68, 92, 93, 94, and 96. We’re in the 90s on all the courses. We did have our full score, 40,456, but that’s only because we have nothing in place to make it do any sort of dynamic throttling until it hits 100T. This tells me right now that we have pretty poor conductivity. And I don’t think that’s going to have anything to do with the thermal paste itself. I honestly think that might have to do with how watery it is. I think a lot of it just smooshed out the side. I really think that that’s what the problem is there. So let’s do that again. I won’t do the 10-minute test at those temperatures because it’s just 94, 96, and 97. Yeah, 39, 827 came down. Okay. So with that said, let’s go ahead and shut this down, let’s take off the pump, and let’s see what the thermal paste looks like. I have a feeling it’s so watery and so thin that as soon as I tighten it down, it just smooshes out. I feel like we need some of that thickness to be able to maintain its position between the two surfaces. Because typically what you’re doing with this stuff here is taking it and putting it on, like thermal couplers or thermal probes, and you’re just gooping it all over it, and then you’re sticking it in a hole that’s reading temperatures of things like the print head and stuff. Because, remember, 3D printers have to be able to monitor the temperature. It has to know what the temperature of the print head and stuff is. That way, it can monitor when you set PLA, or whatever it’s at 200, and whatever C, or whatever it has to know what those temperatures are. So all this, you can just fill that whole void with the white stuff, and then it’s conducting and sending it back, but it’s not having to do it between two pieces of compressed metal at this point. So I have a feeling that all squished out, which means I now have cleanup to do. Actually, no, it didn’t. Not as much as I thought it would, but you can see right there that, with how liquidy it is, it created these like veins, which are really interesting-looking. I have a feeling the veins that are left on the IHS will match up with the low spots and the high spots here will match up with those low spots but clearly it was not transferring heat like if we had thermal paste that looked like that we definitely wouldn’t have had the we wouldn’t have had the best performance but we would not have seen the instant temperatures that we saw in this case so now let’s see how well it cleans up since it’s water-based it should work pretty well with isopropyl alcohol just like wipes it It looks like Elmer’s glue like I still have to take it out I think to get that cleaned up perfectly but yeah well that’s too bad I was expecting I was hoping that we would see some good results out of this because five ccs of this was $12.99 from Amazon and that brings it down to like two dollars and sixty cents a gram whereas like the thermal grizzly was like eight dollars to 99 cents a gram. That’s what that little tube was. But clearly, not all thermal paste types are created equal. All right, so it’s all cleaned up. I had to take the CPU out because it definitely got in there. Little big on the pins, but I was able to just kind of take care of that with some alcohol, fortunately because it’s water-based. It’s nice that it just dries it up and evaporates it. So I wanted to, and now we’re back on KPX. I wanted to just do a sanity check to make sure everything’s running as it should. So if I hit multi-core, it should spike in the mid-80s: 84, 83, and 85, exactly where it was before. I think I said it was like between 84 and 86, and there are 84 and 86. So there we go. Imagine chemists knowing more than some average fat YouTuber. Look, the difference is that every thermal paste is made up of a couple of different materials. Whatever the actual conductive material is, whether it be metal, ceramic, or crystals—in this case, boron nitride—it’s crystals.
So there’s also the suspension material. How is that crystal being applied, right? So this one was actually very watery. It’s very water-based. Obviously, the chemists that have created thermal paste for computers have figured out the right consistency, where it’s not too thin when hot but not too thick when cold. There’s a whole process to that. Clearly, because of the application that the boron nitrite’s being used in for 3D printing, it’s a completely different type of application. Like I’ve already said, it’s not going to deal with the pressure of components pushing against it, which, as it’s very thin and watery, will squish out and create a moat around the CPU, which is pretty much what happened. The difference, too, is that it has to be able to keep that state under those different phase changes. Or in this instance, it did phase change because it got really hot. So it just dried up all the water. those bits and then turn into a solid. But at the same time, you know, those crystals themselves, like the boron nitride itself, have a wattage per meter Kelvin of, like, 700, or something, which is insane. But obviously, you can’t just have a solid connection between our CPU and our video idea. So let’s put sand in it. No, I’m just kidding. Grow a boron crystal for this. Anyway, there you go. The moral of the story is to use the proper stuff. I mean, we’ve done this video before with things like peanut butter and, uh, Nutella. Once, I think we used Oreo cookie filling. We’ve done crazy stuff just to see what would happen. There you go. We learned something today. We learned we’re still stupid when it comes to things that are above our pay grade. We’re looking at you
