**De Bauer's New Products: A Closer Look**

In this article, we'll delve deeper into De Bauer's new products and explore their features, power ratings, and potential design issues.

One of the first products to be discussed is the 490, which is officially rated for 450 watts. However, with the Fe model, users can push the power rating up to 600 watts. During testing, we found that the card was able to handle power levels anywhere from 600 to 660 watts. Additionally, the connector is rated for 720 Watts, although this is not a recommended power level.

It's worth noting that the connector's design has been questioned by some, with one tester even severing all but one pin and pushing the power through that single pin without it melting. This highlights the importance of proper cable management and securing the connections to prevent overheating issues.

The testing process also revealed some interesting results. For example, when we pushed the power levels to extreme values, we found that the card was able to handle significant amounts of power without issue. One particularly notable test involved pushing over 1,000 watts through a Sun Moon SM 8800 and SM 220 load tester. This highlights the potential for these cards to handle high-power applications.

However, it's also worth noting that not all tests are successful. In one instance, we found that a loose connection was enough to cause the card to melt. This underscores the importance of proper design and manufacturing processes to prevent issues like this from occurring.

**Power Ratings: A Closer Look**

One of the most significant aspects of De Bauer's new products is their power ratings. As mentioned earlier, the 490 is officially rated for 450 watts, although with the Fe model, users can push that rating up to 600 watts. During testing, we found that the card was able to handle power levels anywhere from 600 to 660 watts.

However, it's also worth noting that these power ratings are not always a reflection of the real-world performance. For example, when we pushed the power levels to extreme values, we found that the card was able to handle significant amounts of power without issue. This highlights the potential for these cards to handle high-power applications.

**The Importance of Cable Management**

One of the most significant design issues with De Bauer's new products is cable management. During testing, we found that even small issues with cable management could cause overheating problems. For example, when we severed all but one pin on the connector and pushed the power through that single pin without issue.

This highlights the importance of proper cable management to prevent overheating issues like this from occurring. It's also worth noting that not all tests are successful, and a loose connection can be enough to cause the card to melt.

**Design Oversight**

The design oversight in De Bauer's new products is another area that requires attention. As mentioned earlier, the connector's design has been questioned by some, with one tester even severing all but one pin and pushing the power through that single pin without issue.

This highlights a larger issue with the design process, where sometimes only minor issues are caught during testing. This can lead to significant problems down the line, as seen in this case.

**Conclusion**

De Bauer's new products offer some impressive features and power ratings. However, they also highlight some significant design issues that require attention. From cable management to power ratings, there are several areas where these products need improvement.

In our next article, we'll be taking a closer look at the thermal conductivity of liquid metal plating. Stay tuned for more in-depth analysis of De Bauer's new products and what they have to offer.

"WEBVTTKind: captionsLanguage: enhey everyone I am joined by Roman AKA D Bower so the topics for discussion today it's been a few years since we've gotten to do some of these videos and they're always really fun because we don't plan anything and it just ends up being the audience gets to watch us hang out and that's and that's what it is so this is stuff you're doing yes where do you want to start with I would start with uh the am5 direct D stuff so that's something I show briefly on my channel already especially like this so we started looking into the the am5 Heats better even though AMD says it's perfect and fine and 95° is awesome it's great um I kind of do not agree um that's why we looked into like making an upgrade heat better so that was this one before that this video is brought to you by D cool and the new zero dark series of AK 620 and ak400 CPU coolers we previously reviewed the ak6 20 and ak400 and found them to be among a new crop of extremely competitive coolers for the price the new zero dark and zero dark plus variations move out to a blackout color design with blackout FDB fans the heat sinks otherwise have the same characteristics as those that we tested previously and found to be well performing just with a fresh new look learn more at the link in the description below and uh we actually had a different version prior to that which was a little bit smaller but um this kind of is like a a mix out of a contact frame and heats better so it like pushes down the CPU into the socket with a like right amount of Force if you tighten it down to the board so it's kind of easy to use so sockets down on the CPU like that yeah direct ey yeah you basically M this yeah and then uh is the Z height does being the same yeah the Z height is actually a little bit lower it's like one 1.1 mm lower okay I'm guessing that's part of the performance game yeah yeah so we thought about keeping it the same because um originally when we made our direct di frame we had a lot of feedback from you guys that um it's performing well on custom water blocks like the Corsair block or like the EK block but that you cannot use it on monol blocks because monob blocks are like they have their set sea height right right and also aios were a big problem because an AO is designed that you have a heat spread that already spread the Heat and then because the pump is so slow and no flow rate so big hot spot would cause bad temperatures and that's what we saw we had people using our direct diap frame that basically had no improvement and the thing is originally we made the direct di frame so it can be used on like uh custom water cooling it's okay the table is the only thing that car H it's very thick metal the thing is about the contact frame you know um I totally agree with your review that actually ours is quite complicated to to to mount the thing is when we originally made it we made it just for L and2 overclocking guys because we we never expected that this would become a mainstream product that's also why we like didn't keep the like the cost aspect in mind that much because we're like I mean they don't care if they pay like 40 or 30 talking people who buy $400 L2 Pils right and they go through hundreds of dollars of liquid nitrogen and and they would like to be able to like fine tune pressure so they can check different memory clocks and stuff whereas the the average user like the thermal right um solution was actually better like or like easier to use for the normal consumer we made an update ver version actually so the if you buy the 13th gen frame now um I'm not actually not sure if that's already launched but sure that part of the video no that's totally fine totally fine uh but if that comes to the market I'm not sure if it's already in the market but we may we made it like like two months or three months ago this is like this this is like I I'm too success ful I I don't know if my products are even available where is all this money coming from that's certainly not the case I know where all my money is going into machines and stuff yeah well you're this is something we can do at a different time but to build might as well build on it because we're in one of the stepen Roman Mander through absolutely no plan whatsoever conversation but one of your things you were working on you were telling me is uh trying to build out factories in Germany is that right so we actually we acquired a new new building uh it has like 2,000 something square meters okay I don't know what meters are but anyway yeah it's the the proper unit okay it it's big big okay that I understand a lot of rooms so uh we have space right now for about like 60 people okay uh and like like right now we have about 30 people working and I want to hire like 10 15 more this year um because yeah we were moving the the pace production to to Germany um I mean I don't want to discuss politics but uh I mean globally you know you can figure that that things are getting spicy or could get spicy and it's getting also for us it's um it's getting more and more complicated just even to if we Import and Export like raw materials it's pretty strange like uh if just for just an example if I export an metal alloy from China then there's no problem but if I want to export the the raw materials uhhuh that's a big problem makes makes no sense okay well we yeah we deal with similar stuff and I one of the things you start eliminating like so cost goes up in other places but you can at least eliminate say ocean Freight or something like that plus the lead time and a kind of boring topic for everyone but I think it's interesting but yeah I think so too yeah yeah so so that's so anyway what he's saying is by his thermal P cuz he needs to make a building out of it uh and we'll hopefully do a tour once you're we talked about that so that could be very interesting we're actually reserving that for you to not show everything up front because we have a lot of very interesting machines so you can see you can come see how like past is developed like how we make it um like how we fill it into syringes like uh like automated machines and stuff so uh that's very interesting also like an analysis stuff is also quite interesting um like looking into Material Science and that'll be yeah I'm really excited for that yeah so that's that's where the money goes to like buying all these machines how about getting back to this stuff so we're talking about this plate um do you want to get into the water cooling thing right away I mean yeah I mean what what I find like personally quite interesting is um actually those surfaces you can see that this inner surface is like this portion right here looks different from the rest this is because like this is done with a normal End Mill like a normal steel mill okay and this is done with um a Diamond Mill Okay so it's like um um in Germany it's called mkd it's like one just one piece of diamond so it's extremely hard it has a very defined um Edge for cutting so does the machine just have basically multiple bits or Mills in it or it's different machines yeah it can no it's the same machine but you can change the bits and it's the same for this one even though I mean this has been used quite a lot of times but you can see that the surface is extremely uh like a mirror it's like a mirror finish so the the surface quality you can achieve with diamond Milling it's amazing and that the tool we had to make or like we had to go to a company and tell them we need a tool that can have those diamond cuts in it and have this size and I think just for the tool we paid €7,000 just for a tool to make this surface and then people are asking me why this is 3030 and is that is that for one machine basically one tool one tool for one machine yeah um luckily we can reuse it for like some other surfaces but how how much does uh manufacturing time versus mat materal cost play into the for us it's uh perfect 50/50 okay so when when I do the calculation um of this piece U let's say that the copper part maybe is like €7 then I will also have like7 EUR of manufacturing cost okay that's pretty much like a rule of thumb for us and you have all the stuff that's harder to account for like uh providing better customer support packaging um possible overhead for r or whatever like yeah yeah that makes sense the the worst for us is to calculate development cost oh it's impossible I mean the you know what I found Works uh I value my time at $ Z per hour and then all of the calculations are easy yeah we as long as you do it that way yeah we we actually you go well I hope I sell it long enough to profit and then that'll be what my time was worth yeah like right now we do the same we just calculate the Machining cost right and like the material and the tools and everything and like completely neglect development yeah which is like from like yeah it works when you're the guy doing a lot of it I think when you become a big company or if if there are big companies that have to hire teams for it it's much more calculated I I hope I will never reach that state I also don't want to reach that state I I I feel comfortable having up to like 50 people maybe that's more than me than I would feel comfortable with but you grow into it yeah right yeah but that is that is important though if you want to scale yeah so you know yeah so cooling stuff yeah um what I wanted to show is basically uh also some of these prototypes so these were some of the the early samples we made of the the am5 direct cooler and um the way we do it is usually just trial and error because you can theoretically simulate it but from my experience it's a lot easier and faster that you just make like 20 designs okay and you just try all of them sure and then it it takes you like a week maybe well also if simulate it you're going to have to do real testing anyway at some point so you might as well just do the testing so we just uh we made like a lot of different versions of that and those are like two extremes like that is uh this looks much finer yes so that's like a 0.5 mm cut that's like 0.2 mm cut so it's um you can go down to like 0.15 but you don't want to do it it's like very critical process yeah uh because if you cut like I don't know like 40 of those lanes and then your your saw breaks M that's amazing yeah do you do you have I guess you would have to basically throw away that block like melt it down or whatever yeah okay so that's why that's why there is a trade-off between having like the best cooling yeah but also keeping in mind that people not everybody uses high pump speed that's one thing because you if you want to use that one like very fine one with like 100% pump speed it's going to perform the best it's going to perform better than this one but we did not use it for for mass production because I I was testing a 25% uh pump speed to make sure that even like a quiet us it can still push past the impedance yeah what which what did you end up going with either of these two or different one in between so we um we went with 0.5 mm uh like cut okay and the distance between is 0.28 wow okay yeah quite quite specific but then also it's not only this but it's also the height that matters so especially on Direct D the distance between this surface and like the bottom of the cut is usually quite a lot thicker than on a on a normal water block because you you have to have something like a heat spatter uh otherwise the the thermal density is just too much okay it's too much it was too much last night I said this is too much this morning I said this is too much and now I realize it's too much and you will perform worse got it um let's talk about so on these I'm sure someone's already typed a comment you can see the you can see the imprint we'll call it from a previous liquid metal application and we're talking about going to something like say a nickel plating can you help explain to me again I haven't covered this topic in years but why does this happen on say this one and and why do you end up better off over here so I mean the the liquid metal basically contains at least of two uh two metal components it's indium you can add tin to kind of G stret yeah yeah ginan is like the a trade name of a specific composition that has like the lowest melting point that's used like in thermometers okay um but uh yeah tin is actually worse for the thermal conductivity and um topic for later yeah but if you use like Gallum and indium Gallum is actually quite corrosive metal okay it's it's corrosive especially to aluminium uh it's doing this grain boundary diffusion it's called so it's going through the aluminium grains and then it's basically falling apart okay um it's different on on the the copper it's not going into the grains but it's forming an alloy on on the top so it's it's not directly corrosion but the thing is because it's like forming an alloy the composition will be different of the remaining liquid metal that's why after a long time period it will look like it dried out but it actually formed like an alloy with the copper and then the remaining uh liquid metal is a different composition and can be hard at room temperature that's why it will look like it righted out right so then do you avoid that entirely with nickel plated not entirely because even on the nickel if you like if you take guess the IHS is are ni plated but if you if you lose use like an sem and do a cross-section um then you will see that even on the nickel it will form an alloy like on the very top but only to a very small degree very very small degree that's why usually this is used like as a diffusion barrier it's also why applied under an IHS uh and I don't know how it interacts with like modern goldplated stuff but uh but under at least nickel plated IHS to a die um in The Limited Test we've done we've seen pretty good longevity without necessarily needing to reapply uh I don't know if gold is worth going over if that changes at all the the gold um is usually done um during the soling process because the indium sticks very well to it and it actually does the same on Gallum so it also helps to stick very well to it but then again um it forms Alloys so it changes things again it may look like it's oxidizing and uh that's something you don't want actually that's maybe a good um way to transition to this sure because the this this the pad this this is the pad um Qui going to open this one so it's like a graphine based pad uh graphite base pads have been around for for a long time already yeah what was the one I looked at was that a icy diamond yeah icy diamond and um ice yeah yeah we looked at Icy diamonds and and we looked at uh there were a couple others and then in that Roundup we eventually did the carbonaut pad which itself is also different but yeah so this one um it's graphine and graphine graphine and graphite are not that much different except for that graphine is like a it's like a 2d uh layer of um of graphite okay and it has like an amazing thermal conductivity in this in this plane our direction right and the the graphite pads are similar so they have this like insane thermal conductivity in X and Y directions like on a table you would put it on it would like perfectly even this out which is nice for like a phone for example if You' like have to spread the heat to the side but for CPU Cooler actually doesn't make any sense because you have to go in Z Direction the best example I had of a IC Diamond pad when I I eventually after we posted our video years ago came into contact with an engineer uh for McDonald's digital signage he emailed us and he said um uh so he his his job is to design thermal solutions for uh outside digital signage like screens things that are in the sun all day and display a menu and um so the explanation I got was that pads that distribute the heat load horizontally uh sorry Asus is coming for me yeah we we hijacked the Asus meeting room quickly and uh seems like they want to have it back but they're not getting it back look atal filming me getting dragged out by what he was telling me is that they use the pads that uh specifically that spread horizontally efficiently cuz uh they need to distribute the heat on as wide an area as possible because they're normally using large pieces of metal across say a 4 foot sorry let's say like 1 and A2 meter by 1 meter sign um so you have all this surface area but paste wasn't feasible for a lot of these applications because it just dries out so anyway that was one of the use cases but for a CPU this doesn't really work as well because you don't have that much surface area and to get vertical fast I guess that's kind of where you were going with it yeah exactly so um this is um made for us by a Swedish company okay um we've been working with uh for the previous year and um they are extremely experienced in uh just this very specific type of product um they've I think they've been researching this for like 10 years and um so what they basically did is um they stacked a huge amount of those 2D planer gra graphine sheets okay and then they they uh stood them up yeah like 90 degree and then they cut them cool and like this baking and cutting process is uh extremely difficult especially because it's um I mean this is like 0 2% uh 0.2 mm uh thick so it's like percent is the Imperial unit so yes it's extremely thin it's also very fragile actually that's the only downside that um if you don't know this product and you don't know how to handle it so if I like wiggle it around a little bit it will just simply break apart right okay um so you have to be kind of careful the best application what I would recommend this for is either that you want to use and forget like you apply it and it has a very good performance it's in line with good thermal pastes um it will not beat liquid metal simply because of the the height it's not possible but it's still very good and we tested this on even those direct die coolers and you lose like 6 to 7° over liquid metal okay which I think is like the perfect trade-off like okay you lose 6° it's still like 20° Celsius lower than the stock so that's the thing right it's like liquid metal is all you're losing six deges off of something that is already a massive Improvement so um and you will not have any hassle like liquid metal uh but it sounds like because of the relative fragility of it it's not the type of thing you should you should be like I wonder if my my pad is still okay and then pull the cooler off and look um do carbon not a little bit of that where it's it's it's like reusable yeah to an extent uh but once you start clamping down a lot the thing is um most of these pads theoretically so let's say you compress them you use them if you heat them up especially the the graphite pets do this very well so you use like a heat gun you heat them up to like 300° cus and they basically pop up so you can reduce them interesting okay so that's for graphite pads for this this does not work um if you use it once and it's getting compressed you press it in those tiny like micro scratches or something right um it will stay in there and if you like open the cooler yeah you would have to use a new patter it's not reusable well it's reusable but you will lose temperature right if you do it right and but from my perspective that's amazing like we tested this for year with like liquid with liquid nitrogen it works extremely well extr overclock Community they don't have cracking anymore if they use it and so the it sounds like the objective is fill a spot the mark between liquid metal and paste with uh using your own phrasing but like set and forget type of thing yeah um that makes sense especially for the liquid metal application if you used it I know that it works most of the time but sometimes depending what kind of materials you use it with you might be unlucky you might have have something that looks like corrosion and to avoid all of that you can simply go for this put it on there and it will perform extremely well because what I absolutely I will I would love to say that again don't use paste on any kind of direct application okay the power density is way too high for for a normal paste and you will have that something that looks like the the pump out effect I was going to ask yes okay and we covered that in an old video I'll link it below yeah so you can watch that and to avoid this this is the perfect alternative yeah you can't really pump it out with the heat Cycles so yeah St the same uh how about moving over to this stuff what's going on here so we have in general two designs that is the the am5 direct die cooler which is um called micro and that's the like the high-end RGB version so it has an acrylic piece that also distributes the the water and it's mounted like that so it replaces the the AMD Sam okay you clamp it down with those two screw uh those four screws you can screw it down all the way until the the cooler hits the the main board so it's quite easy to use and after that you simply put this back on and you're done so very simple very easy to use also um quite tiny actually yeah so Bas I actually saw the video on that one uh but direct die water block I guess is kind of the approach yeah exactly and then there is this like uh L lower cost version which is this one it's just using a p top instead of the acrylic and the aluminium one and it's obviously also going to going to be nickel plated that's just because this was one of the prototypes right and what we did not show yet is uh this version which is going to be the Intel uh direct die water Block it's the same thing so basically we mimicked the the imprint of the IHS okay to push the CPU down into the socket and have good contact in the center at the same time so it's the same thing as with the A4 block just Intel adaption even even this one should fit on here nice so that was a risky move I'm glad it paid off yeah me too because it's me who designed it I mean in between obviously you make mistakes like one dimension one millimeter off and you're like great I have to do everything again oh also just uh having multiple different samples and revisions and keeping track of them yeah you know what's the best on this one I made a mistake that I uh flipped the inner uh the inner uh structure of the of the you can see the has like those cutouts those two for the smds and when I made my first one I made the mistake that I like rotated this by 180° like by mistake when I put the cooler onto my CPU I like crashed all the all the s smds and the CPU was instantly dead I was like yeah nice thank you that that's the part of the cost we we were talking about earlier the the design and development cost you didn't account for yeah destroying products to test them so we're adapting this into nondirect is well so we are now currently manufacturing a piece that's a little bit higher and it will replace the ilm and it's going to be like a contact frame with water block at the same time okay so for people who want to use it on Intel they can just move away the ilm they don't have to buy a contact frame which is bad for me m but um they can buy this water block which will be like 90 90 $90 roughly sure and instead of having to buy a contact frame as well just get this one product and does it all together okay yeah got it how about uh this version of wire view yeah so we showed some versions before and I think we never showed this it's just ready right now okay and especially with all the 12vt high power I'm not sure if you heard about it uh no what is it oh I think uh might go up in Flames okay there's this famous YouTuber who made a very nice video about it did he make a catch on fire yeah oh interesting yeah it was very nice to watch yeah we made this adapter so for people who don't have a PSU that natively can support 12vt high power and who don't want to use some weird Nvidia adapter because they I mean they usually work they don't catch uh on fire but they just look ugly because they hang down on your card with like four pins and with this you can just PL it into the card and like three eight pins going to the what I usually find um also intertaining to talk about is the power ratings yeah because this is rated for 600 uh let's start with the card so the 490 I think officially is a 450 watt like on paper yeah stock and then if you have the Fe model at least it allows you to go up to officially 600 watts uh and our testing we were seeing anywhere from if it's firar power virus you 600 to 660 watts or so uh and then obviously any additional overclocking I if you're modding the BIOS or something it's it's kind of different scenario but yeah and I think the connector is rated for 720 Watts if I remember correctly I don't remember if it was at 9 amps per pin we could do the math or we could just not but uh I think it was 720 watt but then usually the Nvidia connector has four * 8 Pin M which is total Overkill because the eight pin they are usually used for 150 watt but the actual connector rating is so much higher yeah uh I mean we extreme overclocked cards that I don't know they were pulling like uh 1,500 watt out of like three pins yeah uh at three well so on this one the interesting thing with the 12 high power stuff um when we did some initial testing and we were the only ones who did this but we severed I think it was all of the pins except for one at one point and we pushed all the power through one pin and it still wouldn't melt when it was mounted properly uh and then we also did a test on a sun moon SM 8800 it's like load testing unit and an SM 220 and I don't remember the how how much power we end up pulling through it it was like over 1,000 Watts through through the cable and it was completely fine so there some of the initial testing we did uh which led eventually to us figuring out how to at least melt it in one scenario which was the loose contact problem and the loose contact problem I think feeds into the design oversight we talked about in the original video where it's like if you mount this thing perfectly to a load tester and run whatever crazy wattage through it it's fine or if you snap pins it's fine but as soon as you go and manage your cable behind the case and it Wiggles a little bit like that's where it's that mix of problems so I think it's also quite hard sometimes I think people forget about it but it's um it's always something different like making a product like developing testing something and then having actual mass production like it's obviously the the thing the like user is testing but that's the way it is because if you make a connector like this you have certain tolerances on a connector then you have like accessories cables that have certain tolerances to them and then imagine you make like 1 million of that cable and 1 million of that connector and then you will find out if there was a problem or not I mean it's different if you make let's let's say you tested 1,000 pieces you might not even find that's kind of a lot like for testing I mean so it's not sufficient obviously but a lot in terms of maybe Nvidia goes that high but I sincerely doubt most like cable companies I doubt they test a th yeah of course not and I don't get me wrong I don't want to defend Nvidia right here before you kill me uh on the connector I'm not a big fan of that connector to to make it clear yeah and I'm not either and I I think people do lose the plot with that yeah but but still I can see that sometimes those things cannot be caught in like the development process you will only see it uh during mass production sometimes that's it for this one it's an Abrupt end here because we after this immediately branched into the thermal conductivity rant which is already live on the channel so you check that out but otherwise that wraps up some of De Bauer's new products and we have several of those in housee now so we're planning to work with wi view it looks actually really cool there's some neat features to it so we'll check all that out and as always thanks for watching this completely unscripted video with Roman where we just bander through a bunch of engineering topics really educational for us to do these I love doing these types of videos It's Great learning about things like uh liquid metal plating other different types of metals and all that stuff subscribe for more go over to Roman Channel at der Bower he has an English Channel as well and we'll see you all next timehey everyone I am joined by Roman AKA D Bower so the topics for discussion today it's been a few years since we've gotten to do some of these videos and they're always really fun because we don't plan anything and it just ends up being the audience gets to watch us hang out and that's and that's what it is so this is stuff you're doing yes where do you want to start with I would start with uh the am5 direct D stuff so that's something I show briefly on my channel already especially like this so we started looking into the the am5 Heats better even though AMD says it's perfect and fine and 95° is awesome it's great um I kind of do not agree um that's why we looked into like making an upgrade heat better so that was this one before that this video is brought to you by D cool and the new zero dark series of AK 620 and ak400 CPU coolers we previously reviewed the ak6 20 and ak400 and found them to be among a new crop of extremely competitive coolers for the price the new zero dark and zero dark plus variations move out to a blackout color design with blackout FDB fans the heat sinks otherwise have the same characteristics as those that we tested previously and found to be well performing just with a fresh new look learn more at the link in the description below and uh we actually had a different version prior to that which was a little bit smaller but um this kind of is like a a mix out of a contact frame and heats better so it like pushes down the CPU into the socket with a like right amount of Force if you tighten it down to the board so it's kind of easy to use so sockets down on the CPU like that yeah direct ey yeah you basically M this yeah and then uh is the Z height does being the same yeah the Z height is actually a little bit lower it's like one 1.1 mm lower okay I'm guessing that's part of the performance game yeah yeah so we thought about keeping it the same because um originally when we made our direct di frame we had a lot of feedback from you guys that um it's performing well on custom water blocks like the Corsair block or like the EK block but that you cannot use it on monol blocks because monob blocks are like they have their set sea height right right and also aios were a big problem because an AO is designed that you have a heat spread that already spread the Heat and then because the pump is so slow and no flow rate so big hot spot would cause bad temperatures and that's what we saw we had people using our direct diap frame that basically had no improvement and the thing is originally we made the direct di frame so it can be used on like uh custom water cooling it's okay the table is the only thing that car H it's very thick metal the thing is about the contact frame you know um I totally agree with your review that actually ours is quite complicated to to to mount the thing is when we originally made it we made it just for L and2 overclocking guys because we we never expected that this would become a mainstream product that's also why we like didn't keep the like the cost aspect in mind that much because we're like I mean they don't care if they pay like 40 or 30 talking people who buy $400 L2 Pils right and they go through hundreds of dollars of liquid nitrogen and and they would like to be able to like fine tune pressure so they can check different memory clocks and stuff whereas the the average user like the thermal right um solution was actually better like or like easier to use for the normal consumer we made an update ver version actually so the if you buy the 13th gen frame now um I'm not actually not sure if that's already launched but sure that part of the video no that's totally fine totally fine uh but if that comes to the market I'm not sure if it's already in the market but we may we made it like like two months or three months ago this is like this this is like I I'm too success ful I I don't know if my products are even available where is all this money coming from that's certainly not the case I know where all my money is going into machines and stuff yeah well you're this is something we can do at a different time but to build might as well build on it because we're in one of the stepen Roman Mander through absolutely no plan whatsoever conversation but one of your things you were working on you were telling me is uh trying to build out factories in Germany is that right so we actually we acquired a new new building uh it has like 2,000 something square meters okay I don't know what meters are but anyway yeah it's the the proper unit okay it it's big big okay that I understand a lot of rooms so uh we have space right now for about like 60 people okay uh and like like right now we have about 30 people working and I want to hire like 10 15 more this year um because yeah we were moving the the pace production to to Germany um I mean I don't want to discuss politics but uh I mean globally you know you can figure that that things are getting spicy or could get spicy and it's getting also for us it's um it's getting more and more complicated just even to if we Import and Export like raw materials it's pretty strange like uh if just for just an example if I export an metal alloy from China then there's no problem but if I want to export the the raw materials uhhuh that's a big problem makes makes no sense okay well we yeah we deal with similar stuff and I one of the things you start eliminating like so cost goes up in other places but you can at least eliminate say ocean Freight or something like that plus the lead time and a kind of boring topic for everyone but I think it's interesting but yeah I think so too yeah yeah so so that's so anyway what he's saying is by his thermal P cuz he needs to make a building out of it uh and we'll hopefully do a tour once you're we talked about that so that could be very interesting we're actually reserving that for you to not show everything up front because we have a lot of very interesting machines so you can see you can come see how like past is developed like how we make it um like how we fill it into syringes like uh like automated machines and stuff so uh that's very interesting also like an analysis stuff is also quite interesting um like looking into Material Science and that'll be yeah I'm really excited for that yeah so that's that's where the money goes to like buying all these machines how about getting back to this stuff so we're talking about this plate um do you want to get into the water cooling thing right away I mean yeah I mean what what I find like personally quite interesting is um actually those surfaces you can see that this inner surface is like this portion right here looks different from the rest this is because like this is done with a normal End Mill like a normal steel mill okay and this is done with um a Diamond Mill Okay so it's like um um in Germany it's called mkd it's like one just one piece of diamond so it's extremely hard it has a very defined um Edge for cutting so does the machine just have basically multiple bits or Mills in it or it's different machines yeah it can no it's the same machine but you can change the bits and it's the same for this one even though I mean this has been used quite a lot of times but you can see that the surface is extremely uh like a mirror it's like a mirror finish so the the surface quality you can achieve with diamond Milling it's amazing and that the tool we had to make or like we had to go to a company and tell them we need a tool that can have those diamond cuts in it and have this size and I think just for the tool we paid €7,000 just for a tool to make this surface and then people are asking me why this is 3030 and is that is that for one machine basically one tool one tool for one machine yeah um luckily we can reuse it for like some other surfaces but how how much does uh manufacturing time versus mat materal cost play into the for us it's uh perfect 50/50 okay so when when I do the calculation um of this piece U let's say that the copper part maybe is like €7 then I will also have like7 EUR of manufacturing cost okay that's pretty much like a rule of thumb for us and you have all the stuff that's harder to account for like uh providing better customer support packaging um possible overhead for r or whatever like yeah yeah that makes sense the the worst for us is to calculate development cost oh it's impossible I mean the you know what I found Works uh I value my time at $ Z per hour and then all of the calculations are easy yeah we as long as you do it that way yeah we we actually you go well I hope I sell it long enough to profit and then that'll be what my time was worth yeah like right now we do the same we just calculate the Machining cost right and like the material and the tools and everything and like completely neglect development yeah which is like from like yeah it works when you're the guy doing a lot of it I think when you become a big company or if if there are big companies that have to hire teams for it it's much more calculated I I hope I will never reach that state I also don't want to reach that state I I I feel comfortable having up to like 50 people maybe that's more than me than I would feel comfortable with but you grow into it yeah right yeah but that is that is important though if you want to scale yeah so you know yeah so cooling stuff yeah um what I wanted to show is basically uh also some of these prototypes so these were some of the the early samples we made of the the am5 direct cooler and um the way we do it is usually just trial and error because you can theoretically simulate it but from my experience it's a lot easier and faster that you just make like 20 designs okay and you just try all of them sure and then it it takes you like a week maybe well also if simulate it you're going to have to do real testing anyway at some point so you might as well just do the testing so we just uh we made like a lot of different versions of that and those are like two extremes like that is uh this looks much finer yes so that's like a 0.5 mm cut that's like 0.2 mm cut so it's um you can go down to like 0.15 but you don't want to do it it's like very critical process yeah uh because if you cut like I don't know like 40 of those lanes and then your your saw breaks M that's amazing yeah do you do you have I guess you would have to basically throw away that block like melt it down or whatever yeah okay so that's why that's why there is a trade-off between having like the best cooling yeah but also keeping in mind that people not everybody uses high pump speed that's one thing because you if you want to use that one like very fine one with like 100% pump speed it's going to perform the best it's going to perform better than this one but we did not use it for for mass production because I I was testing a 25% uh pump speed to make sure that even like a quiet us it can still push past the impedance yeah what which what did you end up going with either of these two or different one in between so we um we went with 0.5 mm uh like cut okay and the distance between is 0.28 wow okay yeah quite quite specific but then also it's not only this but it's also the height that matters so especially on Direct D the distance between this surface and like the bottom of the cut is usually quite a lot thicker than on a on a normal water block because you you have to have something like a heat spatter uh otherwise the the thermal density is just too much okay it's too much it was too much last night I said this is too much this morning I said this is too much and now I realize it's too much and you will perform worse got it um let's talk about so on these I'm sure someone's already typed a comment you can see the you can see the imprint we'll call it from a previous liquid metal application and we're talking about going to something like say a nickel plating can you help explain to me again I haven't covered this topic in years but why does this happen on say this one and and why do you end up better off over here so I mean the the liquid metal basically contains at least of two uh two metal components it's indium you can add tin to kind of G stret yeah yeah ginan is like the a trade name of a specific composition that has like the lowest melting point that's used like in thermometers okay um but uh yeah tin is actually worse for the thermal conductivity and um topic for later yeah but if you use like Gallum and indium Gallum is actually quite corrosive metal okay it's it's corrosive especially to aluminium uh it's doing this grain boundary diffusion it's called so it's going through the aluminium grains and then it's basically falling apart okay um it's different on on the the copper it's not going into the grains but it's forming an alloy on on the top so it's it's not directly corrosion but the thing is because it's like forming an alloy the composition will be different of the remaining liquid metal that's why after a long time period it will look like it dried out but it actually formed like an alloy with the copper and then the remaining uh liquid metal is a different composition and can be hard at room temperature that's why it will look like it righted out right so then do you avoid that entirely with nickel plated not entirely because even on the nickel if you like if you take guess the IHS is are ni plated but if you if you lose use like an sem and do a cross-section um then you will see that even on the nickel it will form an alloy like on the very top but only to a very small degree very very small degree that's why usually this is used like as a diffusion barrier it's also why applied under an IHS uh and I don't know how it interacts with like modern goldplated stuff but uh but under at least nickel plated IHS to a die um in The Limited Test we've done we've seen pretty good longevity without necessarily needing to reapply uh I don't know if gold is worth going over if that changes at all the the gold um is usually done um during the soling process because the indium sticks very well to it and it actually does the same on Gallum so it also helps to stick very well to it but then again um it forms Alloys so it changes things again it may look like it's oxidizing and uh that's something you don't want actually that's maybe a good um way to transition to this sure because the this this the pad this this is the pad um Qui going to open this one so it's like a graphine based pad uh graphite base pads have been around for for a long time already yeah what was the one I looked at was that a icy diamond yeah icy diamond and um ice yeah yeah we looked at Icy diamonds and and we looked at uh there were a couple others and then in that Roundup we eventually did the carbonaut pad which itself is also different but yeah so this one um it's graphine and graphine graphine and graphite are not that much different except for that graphine is like a it's like a 2d uh layer of um of graphite okay and it has like an amazing thermal conductivity in this in this plane our direction right and the the graphite pads are similar so they have this like insane thermal conductivity in X and Y directions like on a table you would put it on it would like perfectly even this out which is nice for like a phone for example if You' like have to spread the heat to the side but for CPU Cooler actually doesn't make any sense because you have to go in Z Direction the best example I had of a IC Diamond pad when I I eventually after we posted our video years ago came into contact with an engineer uh for McDonald's digital signage he emailed us and he said um uh so he his his job is to design thermal solutions for uh outside digital signage like screens things that are in the sun all day and display a menu and um so the explanation I got was that pads that distribute the heat load horizontally uh sorry Asus is coming for me yeah we we hijacked the Asus meeting room quickly and uh seems like they want to have it back but they're not getting it back look atal filming me getting dragged out by what he was telling me is that they use the pads that uh specifically that spread horizontally efficiently cuz uh they need to distribute the heat on as wide an area as possible because they're normally using large pieces of metal across say a 4 foot sorry let's say like 1 and A2 meter by 1 meter sign um so you have all this surface area but paste wasn't feasible for a lot of these applications because it just dries out so anyway that was one of the use cases but for a CPU this doesn't really work as well because you don't have that much surface area and to get vertical fast I guess that's kind of where you were going with it yeah exactly so um this is um made for us by a Swedish company okay um we've been working with uh for the previous year and um they are extremely experienced in uh just this very specific type of product um they've I think they've been researching this for like 10 years and um so what they basically did is um they stacked a huge amount of those 2D planer gra graphine sheets okay and then they they uh stood them up yeah like 90 degree and then they cut them cool and like this baking and cutting process is uh extremely difficult especially because it's um I mean this is like 0 2% uh 0.2 mm uh thick so it's like percent is the Imperial unit so yes it's extremely thin it's also very fragile actually that's the only downside that um if you don't know this product and you don't know how to handle it so if I like wiggle it around a little bit it will just simply break apart right okay um so you have to be kind of careful the best application what I would recommend this for is either that you want to use and forget like you apply it and it has a very good performance it's in line with good thermal pastes um it will not beat liquid metal simply because of the the height it's not possible but it's still very good and we tested this on even those direct die coolers and you lose like 6 to 7° over liquid metal okay which I think is like the perfect trade-off like okay you lose 6° it's still like 20° Celsius lower than the stock so that's the thing right it's like liquid metal is all you're losing six deges off of something that is already a massive Improvement so um and you will not have any hassle like liquid metal uh but it sounds like because of the relative fragility of it it's not the type of thing you should you should be like I wonder if my my pad is still okay and then pull the cooler off and look um do carbon not a little bit of that where it's it's it's like reusable yeah to an extent uh but once you start clamping down a lot the thing is um most of these pads theoretically so let's say you compress them you use them if you heat them up especially the the graphite pets do this very well so you use like a heat gun you heat them up to like 300° cus and they basically pop up so you can reduce them interesting okay so that's for graphite pads for this this does not work um if you use it once and it's getting compressed you press it in those tiny like micro scratches or something right um it will stay in there and if you like open the cooler yeah you would have to use a new patter it's not reusable well it's reusable but you will lose temperature right if you do it right and but from my perspective that's amazing like we tested this for year with like liquid with liquid nitrogen it works extremely well extr overclock Community they don't have cracking anymore if they use it and so the it sounds like the objective is fill a spot the mark between liquid metal and paste with uh using your own phrasing but like set and forget type of thing yeah um that makes sense especially for the liquid metal application if you used it I know that it works most of the time but sometimes depending what kind of materials you use it with you might be unlucky you might have have something that looks like corrosion and to avoid all of that you can simply go for this put it on there and it will perform extremely well because what I absolutely I will I would love to say that again don't use paste on any kind of direct application okay the power density is way too high for for a normal paste and you will have that something that looks like the the pump out effect I was going to ask yes okay and we covered that in an old video I'll link it below yeah so you can watch that and to avoid this this is the perfect alternative yeah you can't really pump it out with the heat Cycles so yeah St the same uh how about moving over to this stuff what's going on here so we have in general two designs that is the the am5 direct die cooler which is um called micro and that's the like the high-end RGB version so it has an acrylic piece that also distributes the the water and it's mounted like that so it replaces the the AMD Sam okay you clamp it down with those two screw uh those four screws you can screw it down all the way until the the cooler hits the the main board so it's quite easy to use and after that you simply put this back on and you're done so very simple very easy to use also um quite tiny actually yeah so Bas I actually saw the video on that one uh but direct die water block I guess is kind of the approach yeah exactly and then there is this like uh L lower cost version which is this one it's just using a p top instead of the acrylic and the aluminium one and it's obviously also going to going to be nickel plated that's just because this was one of the prototypes right and what we did not show yet is uh this version which is going to be the Intel uh direct die water Block it's the same thing so basically we mimicked the the imprint of the IHS okay to push the CPU down into the socket and have good contact in the center at the same time so it's the same thing as with the A4 block just Intel adaption even even this one should fit on here nice so that was a risky move I'm glad it paid off yeah me too because it's me who designed it I mean in between obviously you make mistakes like one dimension one millimeter off and you're like great I have to do everything again oh also just uh having multiple different samples and revisions and keeping track of them yeah you know what's the best on this one I made a mistake that I uh flipped the inner uh the inner uh structure of the of the you can see the has like those cutouts those two for the smds and when I made my first one I made the mistake that I like rotated this by 180° like by mistake when I put the cooler onto my CPU I like crashed all the all the s smds and the CPU was instantly dead I was like yeah nice thank you that that's the part of the cost we we were talking about earlier the the design and development cost you didn't account for yeah destroying products to test them so we're adapting this into nondirect is well so we are now currently manufacturing a piece that's a little bit higher and it will replace the ilm and it's going to be like a contact frame with water block at the same time okay so for people who want to use it on Intel they can just move away the ilm they don't have to buy a contact frame which is bad for me m but um they can buy this water block which will be like 90 90 $90 roughly sure and instead of having to buy a contact frame as well just get this one product and does it all together okay yeah got it how about uh this version of wire view yeah so we showed some versions before and I think we never showed this it's just ready right now okay and especially with all the 12vt high power I'm not sure if you heard about it uh no what is it oh I think uh might go up in Flames okay there's this famous YouTuber who made a very nice video about it did he make a catch on fire yeah oh interesting yeah it was very nice to watch yeah we made this adapter so for people who don't have a PSU that natively can support 12vt high power and who don't want to use some weird Nvidia adapter because they I mean they usually work they don't catch uh on fire but they just look ugly because they hang down on your card with like four pins and with this you can just PL it into the card and like three eight pins going to the what I usually find um also intertaining to talk about is the power ratings yeah because this is rated for 600 uh let's start with the card so the 490 I think officially is a 450 watt like on paper yeah stock and then if you have the Fe model at least it allows you to go up to officially 600 watts uh and our testing we were seeing anywhere from if it's firar power virus you 600 to 660 watts or so uh and then obviously any additional overclocking I if you're modding the BIOS or something it's it's kind of different scenario but yeah and I think the connector is rated for 720 Watts if I remember correctly I don't remember if it was at 9 amps per pin we could do the math or we could just not but uh I think it was 720 watt but then usually the Nvidia connector has four * 8 Pin M which is total Overkill because the eight pin they are usually used for 150 watt but the actual connector rating is so much higher yeah uh I mean we extreme overclocked cards that I don't know they were pulling like uh 1,500 watt out of like three pins yeah uh at three well so on this one the interesting thing with the 12 high power stuff um when we did some initial testing and we were the only ones who did this but we severed I think it was all of the pins except for one at one point and we pushed all the power through one pin and it still wouldn't melt when it was mounted properly uh and then we also did a test on a sun moon SM 8800 it's like load testing unit and an SM 220 and I don't remember the how how much power we end up pulling through it it was like over 1,000 Watts through through the cable and it was completely fine so there some of the initial testing we did uh which led eventually to us figuring out how to at least melt it in one scenario which was the loose contact problem and the loose contact problem I think feeds into the design oversight we talked about in the original video where it's like if you mount this thing perfectly to a load tester and run whatever crazy wattage through it it's fine or if you snap pins it's fine but as soon as you go and manage your cable behind the case and it Wiggles a little bit like that's where it's that mix of problems so I think it's also quite hard sometimes I think people forget about it but it's um it's always something different like making a product like developing testing something and then having actual mass production like it's obviously the the thing the like user is testing but that's the way it is because if you make a connector like this you have certain tolerances on a connector then you have like accessories cables that have certain tolerances to them and then imagine you make like 1 million of that cable and 1 million of that connector and then you will find out if there was a problem or not I mean it's different if you make let's let's say you tested 1,000 pieces you might not even find that's kind of a lot like for testing I mean so it's not sufficient obviously but a lot in terms of maybe Nvidia goes that high but I sincerely doubt most like cable companies I doubt they test a th yeah of course not and I don't get me wrong I don't want to defend Nvidia right here before you kill me uh on the connector I'm not a big fan of that connector to to make it clear yeah and I'm not either and I I think people do lose the plot with that yeah but but still I can see that sometimes those things cannot be caught in like the development process you will only see it uh during mass production sometimes that's it for this one it's an Abrupt end here because we after this immediately branched into the thermal conductivity rant which is already live on the channel so you check that out but otherwise that wraps up some of De Bauer's new products and we have several of those in housee now so we're planning to work with wi view it looks actually really cool there's some neat features to it so we'll check all that out and as always thanks for watching this completely unscripted video with Roman where we just bander through a bunch of engineering topics really educational for us to do these I love doing these types of videos It's Great learning about things like uh liquid metal plating other different types of metals and all that stuff subscribe for more go over to Roman Channel at der Bower he has an English Channel as well and we'll see you all next time\n"