Building a custom loop for Dry Ice... IMPROVED!

we're not ready to give up just yet this box is really dirty again round two NZXT is BLD is the simplest way to get a gaming customized pc NZXT SB LD is a simplest way to get a gaming pc customize the way you want it choose from the games you want to play and set a budget and let BLD take care of the rest the starter streaming and creator pcs take the guesswork out of finding a PC that meets your budget and gaming needs are now available with Intel's 10th gen core processors and with a 48-hour turnaround backed by NZ XTS 2-year all-in-one warranty and same-day shipping with blitz mode you can start enjoying your new PC right away to see what NZXT spld can do for you click the sponsored link in the description below now I'm gonna start off by just saying I am absolutely thoroughly impressed with the amount of Science licensed holding people in the comment section of our videos and I'm not saying that facetiously but the right word facetiously yeah I mean that in all seriousness in fact I even pinned a bonafide science doctor guy with a license he's like I actually have a science license but I have a PhD so hope in physics so hopefully this will help he did a very well written so it's obvious he's a doctor guy with how well his comment was written I even pinned it so that people could see that we're taking a lot of the advice that he had and we're gonna try to apply it to getting this dry-ice loop to work it tells me we were on the right track Phil and I knew that we were close we were really close to success paper town and whatever we were really close to success but we just weren't quite sure where we were stuck so two main takeaways that he kind of mentioned that we sort of thought about but didn't really know how important it was is one he said if I built a custom block for the waterblock that was just thicker it's not gonna do any better at actually cooling down the CPU he said is this gonna take longer for it to reach reach that equilibrium which is still going to happen no matter what and I should have known that because it's the same reason I've always talked about having a larger volume of water in a water loop and or your water cooling loop does not lower your temperatures it just takes longer to reach its max temperature because of the amount of energy the water can absorb but the total temperature which it's going to reach is always going to be dependent upon the radiator sizing that you have in your loop so essentially this is the water if you will like the block the amount of cooling that the block is able to provide adding more metal is not going to necessarily provide more cooling the other thing as he mentioned glycol is great at being freeze or heavy freeze protection it's also a terrible thermal transmitter which means it's not doing a very good job at actually picking up the heat from the block sure it's nice and cold because of the amount of time that it's spending in our ice bath that we created with the dry ice but it's not doing its job at transferring the heat from the metal component that's touching the thing that's hot so here's what we're gonna do today we've sort of changed a few things we also know for a fact the glycol makes that full mix was too thick to move through our water block and then we had to thin it down a little bit more because it wasn't flowing through this alpha cool block very good and then I remembered I had this guy this is actually the EK version of this socket type for the 3175 ex it is metal on both the top and the bottom but more importantly it is much higher flow this one is when I went back I watched my own video about four times now honestly not only that I experienced it but you know you go back and you watch things and you see mistakes and you go oh wait a minute if I just stood back and thought about this for a second the glaringly obvious is obvious when we started the ice bucket challenge with Steve about almost two years ago I initially started by having a pump pulling the ice water out of the the cooler and pumping the ice water directly through the loop but what I found by copying Steve is by sealing it in a radiator and having a thing in charge of transferring the temperature of the heat from the loop because remember although the loop may be really cold it's still picking up heat from a heat source and this the CPU and that heat has to be expelled and that's through the heat exchanger or the radiator so that gets exchanged to the cold water but having it sealed in a loop and then however having cold water cycled through the radiator which is the pond pump that we used was way better at cooling and we brought it all the way down to like to see under load where the coldest I could get it before was like 10 C when it was just pumping the water through so what we've decided is we're also going to recreate it a cooling coil and this was also another direct piece of advice that we took from our doctor PhD license holder guy create a coil so that's what I did went down to the local hardware store and we picked up 10-foot coil of copper this is what a lot people would use in like things that need copper and so he also suggested instead of trying to pump the fluid that's being cooled by the ln2 or the dry ice now Ellen - and worried about picking up those bubbles he's like isolate it make a cylinder or something to set inside of this bucket that then is being chilled or cooled by the dry ice and then this is sealed so it can't pick up any of that air so only guys a 1 quart we went and we got these fittings that we put on here that way we can reduce it down to G quarter thread this will be sitting in the middle of it inside of our bucket I have to also have water flow inside and over the coil just having the coolant moving through the coil while the coil is sitting stagnant in the fluid is not going to necessarily work and we found that out with the ice bucket because when I didn't have a pond pump in there and I had just a radiator sitting and I had put some screws long screws in the bottom of the rad so it didn't sit flat on the floor if I took the radiator and I started like doing this with it like a like a handle forcing water through the fins by just sloshing it through like that we saw the temps just start to come down because you need active movement over the cooling coil in this case we I couldn't put a radiator there and dumped it but I think it'll spend more time sitting in this coil than it would in a radiator which will help and this doesn't work well I'm well on my way to make a moonshine now to get inside the ek waterblock to show you what it looks like in there I've got to do that using tools and to do that I would use our trusty odd Richter you guys didn't know here in the United States the month of June is actually Father's Day month so if you want to show some love for the baby daddy in your life why don't you consider by showing him how much you love him my saving $10 off a the order of $50 or more with iFixit you can find a link and a coupon in the description below Phil and I were kind of laughing because this reminds us of like a two-and-a-half-inch like raptor drive from back in the day this is the look how tiny those fins are this is the Alpha cool it's got this flow plate like we've already discussed it was a good design but if we take a look at the ek block you can see right here it has the micro fins and already you can see there's more flow through their jet plate with the opening right here and then it forces it down through these fins right then here and then we think we're gonna start with a 50-50 mix of the glycol yeah because we know it 50/50 it says that - 34 C is where it can go for freeze protection I don't know what's gonna necessarily happen to the transfer I mean that still might be a problem like we still might need to go with something crazy like a lot people suggested that we go with like just a vodka but a lot of alcohol enthusiasts in my comments were like dude vodkas like mostly water or lots of water all I know is I put vodka in the freezer and it doesn't freeze but the freezer also doesn't go to minus 35 C one thing I want to address is a lot of people were like you're thermal paste is clearly your bottleneck I just I just need to point out this is one of those times where you you might have thought you were helping but you really don't know what you're talking about if you understood the the compound of the kpx this stuff is designed to go down below minus 100 C it's the way kingpin designed it we've used it for ln2 we've used it on completely frozen blocks and it stays in this soft State our thermal paste is not our problem so here is the CPU all wrapped up and it's nice little blankets this blanket ie blue shop towels what it's more or less gonna do is it's gonna allow us to visibly see if it's getting wet through all the layers then we'll know it's it's pretty it's time to start maybe we consider drying it up the other thing too is we aren't trying to necessarily create a situation where we're soaking up all the water we're trying to keep the metal from touching the air as much as possible it's the direct contact with air that's gonna cause it to condensate obviously air is in there it's not airtight but we've had pretty good luck with this when we did lnto and stuff the other thing that really helps is the kingpin block cooler I'm sorry I normally think of kingpin and cooling but anyway because it's just got this really thick thermal pad on there basically I'm just gonna and it won't come off at that point so what that's gonna do is it's gonna keep the underside of the socket nice and warm so that it can't it's not going to create condensation the heat will actually dry up any droplets that would start to form before it can actually turn into bigger droplets and then to run so I also need to create water flow through the bucket I think we'll just use this guy we're gonna Bank on the fact that the acrylic isn't gonna have a problem with the temperatures as long as we don't get a cold in the go we should be fine so here's our CPU pump we are running the D DCs which are our single speed which means they're gonna move 200 liters per hour is the speed that they're gonna run at so is no control which is perfect reservoir plumbed into the in simply because we have to have a way to fill the loop because it's gonna be sealed now obviously with this so what we're moving on to now have this fitting on this end but then on this end one of these pointing down channel I'm gonna do is have this fitting inside that now I can control and push down to put a piece of tube coming off of that going into the loop now that I can control like where it returns and how we pick up this is the equivalent in Kerbal space program when you're Rockets all wobbly you just add more struts this is starting to look too good that's my problem now is this looks too good this might be the most like professional looking thing I've ever the idea with the tube going around the outside like that is it's gonna force the water to spiral that's the first thing we need to test right now so now we're going to insulate the bucket this is way cooler looking than that thing we had before now I'm just proving but I'm not a complete just a partial one speaking of complete we know the title is misspelled or the thumbnails misspelled in our last video we're aware of it we did it on purpose the enemy mad people that were like you miss build failure like we also miss built complete and then in catch that one I don't think this foam has even got a very good our rating Joyner the fact but it's all we got I am going to now fill our loop even though a lot of people suggested we try different types of alcohols vodka and stuff like that because I don't want to have to throw away any parts that the alcohols won't you touch and therefore possibly ruin fingers crossed so if I don't put the lid on have it sealed that will just pump into the tank and then that will empty and it won't pull a vacuum it has to be sealed to pull water back through it it's pulling oh my goodness it worked oh it's gonna start shooting out of the hey it's spiraling like I said it would [Music] built a thing so obviously we did the rest of it after lunch proof of concept worked on the bucket you guys saw it spins so or the fluid goes in for Texas so it's gonna definitely have movement over the coil which is important this does not empty although the fluid level does cause I put this tape on there so I can monitor it the fluid levels doing because it's not sealed and this is like totally under vacuum I had to actually put these zip ties on this tube and down here because this is the return line that was actually compressing and it's a problem with the thin walled tubing like this is it it's so soft it'll compress yeah so that separate loop is set the cooling loop is set for the bucket and then this is obviously our cooling loop for the computer I just want to say regardless of how this turns out today this has been fun it's a it's a bit of a mess to clean up when we're done but it's been fun now we need to start the long process of reducing the temperatures of the loop so the nice thing about having this little container rather than just pouring it in the bucket is now because the bubbles are gonna be isolated from the coop the circulation loop I can pour in all these little crystals that break off the big pieces and this just adds a crap-ton of surface area to our cooling okay our CPU loop has now gone - where's now getting to the point to where everything around it the bucket the can the tubes the copper tubing is all like starting to freeze the pumps are starting to condensate which means they're gonna start freezing soon they're extremely cold obviously - 9 now we're still got plenty of flow rate right there with the D DCs problem number one we've reached the temperature limit of our pump watch it go down and it goes back up it's gonna go down and it's gonna go back up so the pump is turning on turning off turning on turning off we have effectively reached the point where the electronic oh my god those are they feel like plastic they're so hard okay the CPU says it's at sea this won't show native negative temperature so I have no idea what we are because I don't it's not like a pot where I had the temperature probe set to the bottom of the pot I have no idea where the temperature is that I bet you the CPU on here is just gonna show zero see it works nice oh my god I don't know how would they found me whoo yeah sorry we've been watching a lot of Back to the Future the kids felt a little bit anyway yeah I'm gonna let the loop come back down it's gonna take time for it to circulate and then I think we go for the overclocks this is 4.7 all court automatic voltage I don't know what voltage is gonna apply and we just did a 16,000 101 that's for seven all four all right I just want to see if I can get one to pass at five oh I don't think it's gonna I think it's just gonna crash I want to see the temps do holy Sh they're staying in the 50s at five gigs go baby go baby 17,000 Oh 32 we just passed an image of five gigs and I have a finite amount of time to get this done because everything is freezing all right so our flow system needs work we're getting ready to our 3d mark scores I am now to just brute force down the temperature in this bucket because as you can see it's not flowing anymore so now we're just brute forcing we're just throwing ice in the bucket now because our loop came up nearly 10 C once that stopped flowing the glycol is freezing look you gotta just go for it see if it'll pass because we now are on borrowed time but look I got the CPU loop back down maybe we didn't need the pump all at all it pulled it down maybe the water flowing through the coil is the same as water flowing on the outside of the coil so now if the coil is just sitting here in this frozen gel and then the fluid is moving through that coil is that the that's the same as it moving on the outside too right maybe we never needed this other loop I don't think we did but now we know that's simpler this way Wow 15,000 219 we are way off so smart me and Pass saved a profile of the memory timings and stuff that all worked we hope the fingers crossed that they work now because you take the memory out and you retrains when you change its positions and stuff so this is five-o with good memory timings I've no idea what the voltage is set to you either I want to see if that is two point one one four on that I tried to overclock the RAM like farther and didn't actually voltage alright so we were so close we got to get cinnamon run in although it wasn't a great score our CPU loop has frozen we think the CPU block might be frozen and we might be dealing with like how the glycol in here is all frozen so I probably over cooled it which is fine we use it to wait now to where we can get it to move again like let it warm back up what I think it's funny is the CPU is chilling at 71 72 and a bump it's like 73 because the block itself is so cold that it can actually keep like the CPU under check right now with no water flow I'm like that crazy used-car salesman that just refuses to be bleep what we're gonna do now is not use anymore dry ice because that was too much at one point I looked out the probe and it said minus 50 and it was sitting in there but what I've done is I've unhooked the coil we're gonna get frostbite ha ha I'm gonna try and take the coil out get the coil out run some warm water over it get it flowing again okay let's try this again that is now just sitting submerged in a cooling gel we have to hope the motion otherwise if I need to cool it more I can just do this you got this there's a lot of area to overcome you got this come on it moves for a second and it stops okay so we figured out why it froze so fast maybe you guys did too I flushed it with water and then immediately dipped it back in there that water froze fast now we are gonna get the loop nice and full of glycol first then we can dip it back in there and start monitoring the temps to come down so like v2 of this would be make it so that I could like to position right out of it I guess v3 yeah v2 of this design but v3 now you know whether it's decimals in versioning look at that look how fast the temperatures dropping now that I stuck it back in there so the flow rate is the same as something moving over the coil that's interesting I've never really played around with coil cooling like this like I know why now like air conditioners and stuff use this but we're just seeing now if we can't get through a score at this point if I can just get through a score that's halfway decent I'll call it a victory even if it doesn't beat my old score so obviously in terms of practicality if I got a higher score with ice bucket then this thing clearly this is done right with ice I was telling Phil you just throw the ice in the in the chest and then you know how low it could possibly go zero and you want to worry about freezing it up and then you're like oh it's getting too warm just throw a bunch more ice in maybe take some water out replace it with ice and then keep going have all the time in the world this is like okay now I'm managing temps can you let it get too cold we know it completely froze up at around minus nineteen seventeen three to eight I pushed the voltage a lot on this one we're at one point four five point one gigahertz ah I knew it was gonna do that 5.0 is my I like hard limit here like we're idiot cold we're not getting it cold enough to go farther so I'm hitting the same wall as dry eye or ice bucket cuz we're only tendency and not even ten feet lower than ice oh no okay well yeah that's our last yeah it tipped over and then it was going boom like it was pushing out the side yeah but it was bubbling out and spitting everywhere okay it's at 44 C uh-huh the one more thing I got time to try one more time I am just when it won't go just give it more just like when your cars are going fast enough push the pedal harder seventeen - nine eight okay eject Oh pump flow Oh cuz seventeen 5:36 17:55 too hard we are all out of dry ice little buddy this is all we get our it's at 47 48 49 50 - 50 okay it's going the other way ref 5.1 gigahertz I got another 100 out of it with the faster memory it didn't even hit 20 milliseconds like one time where but the other runs it did you'd be surprised what point two milliseconds will do to your score this might be it 17 931 we win what did we win and I just want to see now that five - we'll even have a pipedream I didn't really up the voltage much that's a little bit that's exactly that bonafide Bill Nye the Science licensed guy in our other comments between that and a lot of discussions that took place you guys helped make this possible where do I win from this nothing I actually gained a big big cleanup project so I'll go edit this video guys thanks for watching and don't forget Father's Day is around the corner get your iFixit kit for the dad in your life or the husband if he's not a dad woods well if you have kids you got dogs or pets or computers your parents look $10 off orders $50 or more go click the link in description below and you guys will see you the offer okay thanks bye oh you sir [Laughter]

Channel: JayzTwoCents

Views: 563,096

Keywords: jayztwocents, dry ice, dry ice cooling, jayztwocents stupid experiment, experiment, science experiment, cpu cooling, custom cpu cooling, how to cool a cpu with dry ice, jayztwocents is angry, angry jay, custom cpu loop

Id: eBAiY-q6tqA

Length: 22min 42sec (1362 seconds)

Published: Sat Jun 13 2020