[Music] oh it's like smoking really tempered glass cold plates so before any manufacturers get any ideas from our contraption we made this not because we want to see tempered glass rgb cold plate liquid coolers but because we want everyone to stop doing it wrong there's a lot of pictures from manufacturers of radiators and aios front mounted in cases with the tubes at the top or we've seen builds by people where the radiators mounted at the bottom with the pump at the top of the loop in a closed loop system this is a very bad idea it's going to be bad either acoustically in the very least with pump wine and gurgling noises and things like that or in the worst case it's bad because it potentially shortens the life of the cooling product that you've used and you end up rma in it when it otherwise would not have needed it and would have been totally fine if it had been installed correctly so it's a bit irresponsible of the component manufacturers to show things mismounted but we're going to correct that today before that this video is brought to you by thermal grizzlies conductor not liquid metal conductor knot is what we've used in all of our liquid metal and delid thermal tests capable of dropping cb thermals significantly when replacing the stock thermal interface lower cpu thermals don't just allow better overclocks but also lower noise levels because the transfer efficiency is increased the mix of gallium and indium makes for a thermal conductivity of 73 watts per meter kelvin outclassing traditional pastes significantly learn more at the link in the description below all right so recapping the issue then this has been something we've kind of noted on and off and reviews for a long time now but most recently in case reviews over the last six months or so we've noted that closed-loop liquid cooling products should generally be mounted with tubes at the bottom if they're front mounted although this isn't quite as bad as mounting the radiator at the bottom with the pump at the top and these are both sub-optimal at best to damaging at worst configurations for an all-in-one cooler and it's what causes probably one of the highest reasons for rma with liquid cooling manufacturers now case manufacturers are particularly to blame for showing the products installed incorrectly and since they are actually selling a product to people and providing recommendations on how to assemble it it's particularly irresponsible to show coolers installed incorrectly on manufacturer websites we've also seen just various system builders online on forums youtubers people like that installing radiators in a way which is again maximally damaging to the product minimally acoustically annoying causing rmas because people think that their pumps whining their pumps loud they think liquid coolers are always loud they think pumps are always loud because it's actually just been installed incorrectly uh and we don't really blame the youtubers as much for that and this obviously everyone will well we'll know now it's more about blaming the manufacturers for sort of setting that undertone to begin with that it was okay to install these raiders in a certain way so this isn't going after anyone in particular although maybe the manufacturers were posting those images on their websites showing coolers installed incorrectly okay so what we've done is we've cut up a radiator we're going to be using this as a test vehicle to show you how the water actually flows inside of the loop because since saying it and case reviews for the last six months hasn't been good enough we thought we'd actually just cut one of these open and show you how it works to recap the problem briefly this we originally wanted to make this content piece because about maybe eight months ago on the patreon discord i received a message from someone who one of our backers on there who said hey steve i bought this cooler that you thought was pretty good but i have this pump wine issue i was wondering do you have that problem too and so i messaged back and forth with him i got a photo of the system and i said try flipping the radiator over and get the tubes at the bottom of the case if you can manage if you have enough tube length to get to the cpu still and so the user did that and then came back to me a day later and said thanks so much the problem solved that doesn't make any noise anymore and so we've been through this process with a couple of other people and one-on-one scenarios on the patreon discord over the past year or so and almost always flipping the cooler the correct way around fixes the problem why manufacturers aren't more forward with this being the recommended installation is beyond me but arctic as an example is starting to revamp its manuals to include specific documentation on not just best use in terms of product survivability which they've already got the best use in terms of human factors noise things like that so we took an older closed-loop liquid cooler that we no longer needed we drained it and then we took a dremel tool to each side of the tank we should probably pause here to define the parts of a liquid cooler because this is something we did previously in an animated video and it might prove helpful water enters the radiator from one of the barbs typically this is after going through the cold plates microfins atop the cpu some pumps are located in the radiator but we're not talking about those today we're only talking about the pump and block installations the water goes through a barb that's tightened with a piano wire internally and then it enters the tank the tank is split about 40 60 or so for inbound and outbound water the inbound water flows down the tubes which are narrow but very tall and it maintains a higher flow once it hits the other tank and turns around to return flow slows down something we'll talk about later the fins contacting the tubes are responsible for spreading heat across as much of the surface area as can be crammed in without causing excessive flow impedance to the attached fan which dissipates said heat when the water routes back to the other tank where it enters the barb positioned about halfway down the tank it continues on to the cpu coolers pump housing pump and micro fence we'll talk more about how pumps work in a moment because that's another important part to the damage people are causing by improperly mounting radiators back to our experiment we took the dremel to the tank wall which is made out of aluminum and cut in a rectangular viewing port on each side they're not clean cuts but that's because we weren't exactly sure how deep the barbs would protrude into the tank so we repositioned a few times turns out that the barb terminates almost exactly where the tank starts making this easy each window is isolated in the middle by a dividing wall which splits the inbound from the outbound water and separates it we next took a drive to buy some acrylic super glue and a super glue curing accelerant the accelerant chemically bonds with the super glue to rapidly dry it generally in under 10 seconds it's also a watertight seal for at least this experiment after one failed attempt where we smudged up the plexiglass we tried again and got a good mount to the radiator tank with the window we then applied glue to the entire exterior of the glass to securely seal it from leaks there's still some aluminum debris in the radiator from our cuts we made but just know that that wasn't in there when we started that's from our hack job of cutting it open we next removed the cold plate and hastily cut a piece of acrylic for a viewing port the goal of this window will be to view how the pump can run dry when mounted with the tubes up and the radiator mounted down at the bottom like in the nr200p for example or with the pump at the top of the loop and with the radiator in the front or the radiator in the front with the tubes at the top and the pump in the middle of the loop it'll also show how the pump works which makes for really cool footage we kept the gasket to keep flow guidance but obviously removed the jet plate and the cold plate for visibility into the system this was again sealed with glue although we discovered some flaws in this old design where it could leak from under the pump plate we glued that as well there are two really important points behind the science of liquid coolers that we need to get to before we get into the experiment obviously we get two or three times the amount of views if we just started with the experiment and made a four minute long video but our goal is to hopefully answer this once and for all so no one has questions anymore unless nonsense stops and people stop killing their coolers prematurely or mounting them in the bottom of the nr 200p so the first point is that liquid coolers have air they will also develop more of an air gap over time as permeation takes place and that air pocket increases but even shipping from the factory they're going to have air in them we asked manufacturers three actually three specific liquid coil manufacturers and their suppliers there's two suppliers in here those are the people who actually engineer and manufacture these things and the manufacturers are the ones who set the specs and the average distribution of liquid to air volume is between 10 uh to 2 percent air with the rest being liquid a propylene glycol mixture so we can't name the two suppliers explicitly we got this particular information from but it's the two largest suppliers in the space one of them noted that they have a two percent air target for their closed loop liquid coolers the other one said 10 percent the reasoning for that is one manufacturing challenges significant ones and two is allowing some room for expansion and for preventing a burst of a weak seal in the event of a higher pressure scenario from running at higher temperatures especially high liquid temperatures as for the manufacturing side of things we've actually been to cooler factories and can show how that works after all the assembly processes are complete and the radiator is sourced from a supplier and connected with the loop the cooler is taken to an infusion machine where it is vacuum sealed for a couple of seconds making the loop airtight the machine then relies on the natural physics of air pressure to suck liquid into the cooler and its radiator from an external tank that's filled with said coolant instead of filling it forcefully pressure is used to pull the liquid into the tanks cooler masters factories is one of the highest end that we've been to for this and it uses a simple button press to pressure check multiple times against the volume to ensure it's within safe operating parameters to prevent leaks that means leaving some air as a buffer but there's some left anyway when disconnecting all the hoses the infusion machine and the vacuum seal another manufacturer told us that its coolers typically have about one cubic centimeter of air in stock conditions for the average cooler one supplier further told us that it has a loop which is actually 100 filled with liquid for a laptop product but that this significantly increases cost and it is not feasible for mass-market single-unit discrete coolers as for permeation that's dependent on a few things mostly the temperature of the water in the loop the materials used in the loop and the length of the tubing in the loop things like clear acrylic or clear plastic will are much more permeable than working with something like epdm or something like that for your materials and we'll talk about the material engineering in a few moments so over time your loop will slowly lose water volume to permeation it's going to you'll end up with a lower volume level if you could see into the tank like you can through hours that we've made here and most manufacturers currently spec for a life of at least five years of ideal operation with potentially some life beyond that in sub-optimal operation many units do live beyond this like some of the ones we've had here running for six years at this point but that's assuming that the cooler was properly used and installed and not too highly thermally abused so let's get into some of the testing information and layouts and then talk about the rest first with the cooler incorrectly mounted it's very likely that the customer will experience one of two things one is that there could be poor thermal performance as air bubbles settle in the pump block eventually leading to an rma or replacement before it would actually be needed if used properly the other one is an avenue where the cooler can get replaced for acoustic reasons so as far as permeation goes the first issue this can also speed up if running suboptimally due to air bubbles accumulating in the pump or otherwise impeding water flow and as for the acoustic issues most of our commenters who've complained of pump wine or that water feature trickling noise have found that flipping the cooler the right way around resolves many of those issues people who don't do this may end up prematurely replacing the part when it'd be fine if mounted more appropriately permeation rate increases with longer tubes higher coolant temperature and with thicker tubes and is also highly dependent on the material used in the pump block and the tubes and we'll talk about that aspect after we get through the testing most loops are designed to run under 60 degrees water temperature like the acetech products and keep in mind that you can run a 9900k at 95 degrees t-junction all day long and still be in the 30s for liquid temperatures so hitting 60 for the water temperature is very uncommon but it becomes more likely as the improperly installed loops age the plastics get soft at about 7 dc for acetec products where the clamping force will start to mushroom out the plastics when they are softened under those higher temperatures coolant pumps can operate about 5 degrees higher and a hotter cpu or gpu will accelerate permeation as will frequent thermal cycling or constant uptime under high load if you rarely turn the system off and also have the radiator improperly mounted the issues will compound and could lead to early cooler death where it might survive otherwise if it were mounted properly to begin with let's go through the problems and then we'll get back to the science of manufacturing these products our first experiment will demonstrate a common proper installation choice top mounted tubes down from a water circulation perspective you'll basically never go wrong with a top mount with the tubes down it's not always the coolest to run exhaust instead of intake but at least the flow will be correct in this orientation the pump is below the radiator and the air will sit at the very top of the loop as it was designed it won't get sucked into the tubes because flow isn't fast or violent enough in the inbound tanks to cause that in our clip here you can see where we change to this orientation that the inbound tank rapidly fills while the outbound tank drains the entirety of the air presence in the loop is now visible above the outbound tank well above the barb so our loop is running healthy it could use some more water but this could also easily run for a few years like this if under operating conditions that aren't abusive in fact to prove this point this particular evga hybrid cooler is about five years old now and it's had about two years of use before we stored it for a couple of years when we drained this for this project of cutting it all open it had about 100 milliliters of water in it and that's pretty much what these ship with plus or minus about 10 milliliters but it was also mounted properly for its entire in-use life and it wasn't abused thermally looking at the cold plate during this orientation you can see the water is expelled from the pump by centrifugal force and maintains a high speed the water then re-enters the outlet and goes back to the inlet tank it'd typically go through microfins in this process but those were removed for visibility remember what these clips look like as we progress through the others because these ones represent how it all should look internally getting into the problem mounts then air will always try to find its way to the top of the loop starting with the newest improper installation that we've seen spread far and wide with the nr200p mounting a pump at the top of the loop is one of the quickest ways to cause noise poor cooling performance and potentially early replacement of the cooler it's not always necessarily early death of the cooling product because pumps like ace attacks for example use magnetic levitation and therefore don't technically need a lubricant from the liquid to stay alive but the damage can be caused elsewhere in the loop from running at temperatures that soften the plastic if you don't notice the air pooling in the pump keep in mind that a user's tolerance as well for high temperature will be lower due to reduced performance so it's often assumed by the user that the pump died even if it technically didn't it might be fine and just need to be rotated to get the air bubble out but from a user's perspective this normally constitutes a failure and they'll just rma it and call it dead tubes up radiator down and pump up top is what we're starting with this means that the air bubbles will rise to the pump in our side by side you can see how a pump at the lowest point in the loop is going to behave when it compares to a pump at the top of the loop there's high velocity of water entering when installed properly while liquid doesn't quite reach the pump when it's mounted improperly we didn't simulate much permeation here because you can see the entirety of the air bubble in the shot of the plate in combination with manufacturing tolerance running a cooler like this for a while could easily get you to the permeation levels equivalent to the liquid level tested even if yours is filled higher than this one you're still going to run into some issues over time it will cause endurance issues as it ages and it will cause acoustic and thermal performance issues immediately but especially acoustic issues air can get forced through the loop eventually as well especially if you tap it and it'll briefly get stuck at the top of the tanks before going back into the bars this air will then prevent proper flow if it's sitting between the tank and the barb especially when it jams the outgoing tank that said for the most part it'll just end up getting sucked back through the pump because the pump is at the highest point in the loop either one of these is bad the former one prevents water from going up the tubes but it's likely to be resolved shortly and on its own the latter causes immediate issues though with pump wine aging and performance as for the pump you can see the yellow pump in the middle of the cold plate barely getting any water the cpu will run hot while idle but not hot enough to trip tj maxx and it might trip tj maxx and throttle under load but that's still a performance issue and it's still something that a user would call debt worst case scenario though is that it's working just well enough it's got just enough water going through there to keep the cpu under tj maxx but not well enough to keep the water really properly circulating in this instance what happens is the plastics will begin to soften internally and deform and over time the pump will in fact be something you could call dead technically it would still work but the unit overall would no longer be functioning as designed so that wouldn't be even a manufacturing flaw it'd be a user flaw at that point from mounting it improperly the next one is just sort of embarrassing mostly because manufacturers do this all the time when showing off their cases and their coolers nzxt for example is most guilty of this and it's a added embarrassment here of being the technical manufacturer of the coolers that it then shows installed incorrectly when marketing its cases thanks to manufacturers doing this because it looks pretty and thanks to power supply shrouds requiring this configuration in many instances people often mount their coolers tubes up we always try to point them tubes down if possible and if it doesn't fit that way we mount it in the top of the case instead in this one we have to consider the location of the pump the pump is likely to be about one and a half to two inches below the tank which is a good thing we don't want the pump at the top of the loop anyway if the pump is sufficiently below the pump will be filled but air bubbles will occasionally get sucked through the outlet tube and into the pump that's because the air accumulates right next to the barbs but ideally we'd have it flipped so that the air accumulates in the other tank that isn't attached to the barbs and the tubes the air bubbles that get pulled through will circulate but this is the reason a lot of users complain about that water trickling or gurgling or bubbling noise this is really more about the noise than anything else in most instances it does cause performance issues as well but it's primarily a noise problem so if you've complained about pump noise check for this orientation you can also get pump wine on some pumps when air is sucked through them and that might last past the period of the air getting pulled through in theory a small amount of air won't get pulled through because it's not a high enough speed to get sucked into the tubes but as the air cavity grows once it gets down to where the barb actually connects to the tank it can produce that bubbling and trickling noise that you sometimes hear as the air cycles through the loop finally when front mounted and with the tubes down this air bubble issue is no longer a problem ideally the pump is still below the top of the loop because the other tank should now be at the top of the loop if that's the case we don't need to worry about air accumulating in the pump because it won't and we also don't need to worry about it getting pulled through the barbs because that's not where the air is anymore it's not at the top in the tank that's not connected to them this top tank is actually designed in most radiators to capture air anyway so the air will pool in the tank and sit in such a way that it's no longer a consideration in the loop you won't experience any problems after the first few minutes of run time where the bubbles might shuffle around a bit as the loop is first powered on but those first few minutes you'll hear it after that though things should smooth out and the only thing you have to worry about is long-term permeation if you're abusing the cooler thermally and that might still be a few years from now there's one huge reminder for this orientation that do not mount the pump in such a way that it's at the top of the loop with the barbs at the bottom because then it hasn't fixed anything if you intentionally position your radiator lower on the front panel for some reason rather than higher it's possible that the pump ends up at the top of the loop at which point we're dealing with all the same concerns as we showed earlier when the pump is located at the top now that we've shown all this it's worth expanding on some of the science behind the cooler manufacturing so tubing material it's commonly commonly fep epdm or a teflon or ptfe epdm is ethylene propylene diene monomer material and basically a rubber there's also ptfe which is a teflon lines tube you typically see it in a product like this like cooler masters where you can identify it usually by a corrugated tube rather than a rubber one there's a teflon lining inside of this and that helps with preventing permeations one of the best approaches to it actually but there's a significant downside which is that if you bend these tubes sufficiently you can crack the teflon lining and that means that it will actually permeate much worse performance goes down in general and uh it's a flaw of that particular material that weighs against its upside for some people as for fep that's not as common anymore but that's fluorinated ethylene propylene acetec provided gamersnexus the following information on its material choices quote the rubber that we use is a special epdm mix in the mix is typically iir or isobuteline isoprene and epdm the permeation rate of this material is greatly dependent on the mix between the iir and epdm but also the internal mix of the epdm or how much flame inhibitor there is in it etc as such epdm can vary greatly in its permeability and supplier customer relationships are very important as it ensures the expected quality levels and permeation rates which for obvious reasons is critical in aios our r d and quality department referred to the manufacturers of epdm typically safeguarding their recipes as comparable in ways to coca-cola and its recipe the mix that we use is ace attack only as an epdm mix that is specifically designed to deliver the performance we're looking for to meet our rigid quality and control program the cooler master uses fep epdm and the ptfe solutions for its tubing according to testing that we received in documentation from one of the suppliers off record fvp evaporation rate is about 0. 001 to 002 grams per day under ideal operating conditions this would accelerate if you're in sub-optimal conditions further it noted that epdm has an evaporation rate of about 0. 004 to 0.
007 grams per day and it's testing under ideal conditions anything featuring acrylic or transparent plastics will undergo significantly worse and accelerated evaporation and permeation over time and most chambers use a pps and glass fiber to minimize the liquid absorption but those that don't may end up in a situation where their plastic is less heat resistant and can mushroom under the tension of the cooler when we asked around two of the cooling suppliers told us that coolers operating with a liquid temperature of about 20 degrees celsius for their life can mostly be expected to live forever now that's from a permeation standpoint obviously the coolers can still die from other things like mechanical failures over time but the liquid permeation is not going to reach dangerous levels under those conditions in that lifespan regarding waterfall effects internally the flow rate is generally not enough that you end up with waterfall effects being a concern but the liquid slows down once it hits the outbound tank in the cooler because it's already gone through all of the tubes within the radiator so this is what can cause air bubbles to form and eventually be sucked through and back into the pump as we showed in one of our experiments we'll quote arctic on this one as well because it provided the following guidance for clearing initial install pump noise issues quote indeed during shipping the air is somewhat split all around the place once you plug the pump in the first time it will for sure make noise the noise will settle after minutes of running it i would recommend even a few hours to have the final result you might also help by stopping the pump a few seconds and restarting it this is because due to centrifugal force the air in the pump tends to stick to the pump axis and it will eventually go away but stop start cycles will help speed up that process anyway that's assuming the cooler is actually mounted properly so anyway long story short make sure you don't mount your liquid cooler like this with the radiator at the bottom and the pump somewhere way up top and really really try to avoid mounting it like this we understand that sometimes you don't have quite the tube length to mount it tubes down in the front and still reach the cpu cooler but try your hardest gpus can get in the way stuff like that but it will ensure the best longevity of the device and also the lowest noise hopefully that helps answer all the questions in a definitive way so we don't have to approach this topic again share this video to make sure people stop installing their liquid coolers incorrectly because a lot of people sadly end up wasting money on them and buying something else because they think that the cooler is bad when it was just installed in a way which is not optimal so that's it for this one if you want to support us directly you can go to store. camerasaccess.
