Gadgetory

everyone welcome to another episode of sgn as always if you have questions you can leave them in the comment section below or if you're a patreon back you can go to patreon.com/scishow sexist comes out directly and then join our discord where we haven't asked GN discord for patreon backers so this week we've got a couple of topics one the opening topic is on knock - as fans the thing that blew up on build a PC and then we'll be getting into the viewer and reader questions before getting to that we have a brand new shirt design so that's this one right here this is on the GN store you go to gamers Nexus squarespace.com to grab one of those we have them in teal and in grey with cotton and tri-blend and they were kind of debuted on our first stream which happened last night as of filming so that'll be actually last night for you as well um so pretty cool stuff we're going to do more streams in the future I don't have a schedule for you it's just kind of going to be when I want to take something apart like this and we'll go from there we'll try to announce it on Twitter but that's about it alright so for the episode let's open with Noctua the build the PC thread that blew up if you didn't see it was basically that there are manufacturing differences between the fans made in China and the fans made in Taiwan by Noctua so the a f-14s I think were the ones that were discussed there and I theoretically would also apply to the 120-millimeter fans so I have some thoughts on this we are actually getting a couple sets of these fans in from China and Taiwan to test I will tell you right now what my kind of hypothesis is leading into it from an industry standpoint and then we'll see if the testing disproves any of it so first of all the differences pointed out in the reddit thread were primarily color there's a slight color variation between the standard ugly nock to a fan and the slightly different ugly nock to a fan between the China and the Taiwan factories and then beyond that there were some differences pointed out in the what would consist of your output from the tooling or the mold and those were small features like a slightly different depth profile for some of the if it's that are cut into the side of the fan or that are on the fan blades and slightly different changes in the the outer four screw hole areas so let's first talk about tooling manufacturing and molds for making different products when we've toured a few factories in Taiwan ads in China and the thing that is ultimately the companies make tooling which for cases anyway tooling it consists of large pieces of steel that you use to create the chassis for a fan you would have tooling you would have molds or you'd injection mold stuff things like that there's a tolerance even at the same factory one tool to the next or one molds the next because they age so the source the source mold or the source tool if you're doing case tooling eventually wears out and has to be replaced and when you replace it hopefully your tolerances are pretty tight but there might be some changes though you would hope that they're isolated to a point that you're not going to introduce some kind of issue to the product so that's number one that means that there can be changes even from the same factory but maybe not as big as the one seen here that's not to say they are big changes but they are certainly bigger than the very small things you might see normally the next thing to discuss is with the the move from one factory to another you're either going to be cloning tools or making them in another place so that's a factor to account for and then there's tooling the age so as they get used and worn down it may be the case that one factory just happens to produce a whole lot more fans than the other so maybe the tool made that was responsible responsible for making the fan that was shown in the pictures was older or newer than the one next to it so those are things to keep in mind next thing to keep in mind is a lot of the stuff that Noctua boasts for their fans and all of the other fan companies too is just kind of marketing it doesn't really do anything of note that's not to say the fans aren't good they certainly do their job but also a lot of those small features if if you were to take putty and fill in all those tiny holes on the inner frame of the fan I would bet there'd be a pretty small difference I'm not positive we'll test hit but I would bet there'd be pretty small difference and even less so when you have the the holes or the divots on each fan chassis but one is slightly different than the other one but they're still present I think that's going to be a small change ultimately this whole Noctua thing looks to me like it's one of the usual industry discussions that pops up every couple years happens pretty frequently where a manufacturer changes supplier or factory how people notice and then everyone is like why are you selling me a product that's worse than what was originally designed now sometimes that can be the case but not always there's a reason for example that Logitech doesn't talk about or confirm or disclose their supplier or source for their switches in the mice so we can ask the question as simple as you say you have 50 million clicks switches are there's Omron and they'll say we're not going to answer that and as they've pointed out to me it's not because they're trying to be difficult or because they don't know or because they think we can't find out it's because they don't want to go on record if supply changes and that's true for most of these companies and that goes to per fans as well so if there's a supply side change or a factory change they try to make it happen silently and then theoretically the new product should match the quality of the old one the difference here is with Knox oil you're basically looking at color so that's sort of annoying yes the color of one fan is different than the other if you're trying to color match your system with Noctua fans then then that would be kind of annoying but we'll see I have some coming in we just finally for the first time ever got in touch with Knox what I'm going to try and get them to just send me a bunch of fans from each factory so we can do a large sample size test and either I'll be wrong and there are actually large differences or or not there might be small differences if they want something a larger sample size we're also getting our own from readers viewers and just buying them online so we should have samples from hopefully the official source Noctua and then retail sources that way there's no concern of did they just send you special samples that were painted differently or something ok so I think that's not but basically the knocks within it seems to me like it's potentially a little overblown in terms of coverage and concern and things like that but there could definitely be some manufacturing quality differences generally moving one factory to the nut to another you have those it could be possible that the China or the taiwan-based Factory is producing plastics or using a mold that is in fact lower quality but but let's just kind of wait and we'll see what the testing shows if any difference at all ultimately the differences between them are pretty insignificant from the looks of it but we'll find out next next real question this one comes from YouTube it looks like I think so it comes from Kaidu who says oh yeah this was easy this was on that had GPU quad GPU anther machine so this count was kind of amusing a idea said regarding the GPUs that we had on the shelf we have a shelf of GPUs because we use them for work said why don't you use them for mining waste of card just sitting there so a couple things they don't just sit there the content that people watch on this channel does in fact require video cards and then also I think the definition of waste is maybe a little different between us I wouldn't want to use cards 100 percent load mining and then have to use that same card for thermal testing later because who the hell knows what kind of abuse that went through under a hundred percent load non-stop it may no longer represent a real user scenario or it made something may burn out or some other issue could pop up or there's some difference between stock that makes it no longer representative and that makes it a bad test no we don't abuse the cards for mining played with it for a bit for that how do manufacturers feel that mining video but that was about it so I would not consider that a waste as a side point there was a comment replied to that one I think that I ended up deleting for a lot of reasons but it said someone else or someone else made a comment saying that there was a certain level of arrogance of having a lot of video cards on a shelf just as a reminder generally speaking you don't walk into Lowe's or Home Depot and look around go wow these people are really full of themselves look at all of this hardware they have here man that's I I can't shop here there's too much arrogance coming from the drills and the hammers on the shelves I'm going to have to go somewhere else so yeah we actually do need the hardware for work that's why we have it this leads into the next question which was what do you do with old hardware so pretty good transition there for once the old hardware answer this in the video Commons as well but there's a few different options so things that are smaller GPUs video card CPUs we try to keep basically forever where as long as as reasonable video cards we're starting to run out of space for those so that is a problem but CPUs we have plenty of space for memory plenty of space for so we try to keep those for a few reasons one is regression testing so a couple examples of when we would pull an old piece of hardware off the shelf would include the i7 930 revisit where the 22 Connery K revisit where you basically grab an old product and say I wonder how this does today that's pretty good content it gets good views it's certainly worth keeping and for something like an i7 930 just speaking very plainly we're going to make more money on the content creation for a revisit on that than we would for selling it for 30 50 whatever dollars it sells for these days and also it's just kind of cool to hang on to that stuff for other regression reasons like what if someone discovers some major flaw or issue with these things there were recently some hyper threading bug discussions pertaining to older processors and it's good to have that stuff around so that you can test it if it comes down to needing to test it so we hand out to it for those reasons and comparative analysis what if becomes out with the competitor to Company C which I'm using because those letters don't start with IA or n if Company B comes out with a competitor 2 company sees last-gen mid-range product we want to have that last gen mid-range product to make sure we can test it in modern era new drivers at all and see how it works so that's what we try to do some stuff like cases are just impossible to keep because they're huge and you hang on to the stuff that's either new and might you might need it or is particularly interesting and like for example the anti HTS 340 non-elite that's an old case now but I keep it because it's I mean we liked it and it was very popular so it's good to hold on to because what if in the future we want to do how does today's I'm the exceeds that's 340 compared to whatever competitor comes out four or five years later whatever it is but eventually you have to get rid of cases they're too big so we do have a local high school that actually we have a connection there where we donate some of the older cases to the school so they can use them for the computer engineering class some stuff gets sold or given to friends but that's all pretty standard stuff I think for reviewers and then the smaller stuff CPUs we hang out to GPUs as long as we can what else even is there SSDs we actually use everyday and test benches so there's there's going to go anywhere but I think that answers that question for the most part the next one is a question that was for AMD I passed it along so Crimson asked this in some video it was one of our Vega videos maybe the hybrid modders something crimson asked does that Vega do 10 bit OpenGL because only fire pro and quadrant cards can do it currently and video may give you ten bit DirectX but no application with ten bit support uses it so I sent that quote to Andy and the representative got back couple days later and said good news 10-bit color is supported in OpenGL on the Vega Frontier edition did not comment on our activator of course you'd expect that but I think that hence just had pretty cleanly next one this is from Zeta who is on discord in the patreon group Zeta said is there a reference motherboard design like with graphics cards oh wait and on I misread one of the words mother bet powered mother reference mother Bower design so Zeta we talked about this in chat but for everyone else this was a more popular thing to talk about a couple years ago so if anyone remembers the Intel boards that they made the actual Intel branded boards they did the whole campaign that was a better together which I think AMD may have done as well at some point but they basically did the whole better together by an Intel board as an Intel CPU thing and some of those were more beefed up than others but this was around the era when we last looked into this discussion of motherboard reference designs the manufacturers did and still do supply a course back that says the boards for this processor this architecture need to have an A+ for pin or a single a pen or minimally a for pin or whatever it is for EVs trouble and then they also have suggestions for other components use maybe the RM layout suggestions they'll have the spec for the PCH or the chipset the spec for the CPU and how many lanes it can support and all of that goes into basically laying out this is your max PCIe slot count that you can reasonably use this is your max SATA slot count before you have to get your own third-party controller and send stuff like that so there's sort of a reference design I don't know the extent of that today I'm actually not sure on if the board vendors still use a lot of that stuff as heavily but when Intel made their own boards it was more common to talk about that kind of thing or hear about that kind of thing at conventions and booths where they'd say based on Intel's whatever reference board that they were using at the time and this is also where you run into questions of what does Intel use what does AMD use when they're building the CPUs and obviously no bother boards exist yet generally make their own and test with those but a next question is tau tau says all the pre-built this is a youtube I think although this might be a discord I don't know anymore all the pre-built liquid-cooled AIB partner cards I have our dual slot cards with AC LC cooling the GPU and maybe the vram with Reverend style fan cooling the BRM and other non GPU components why haven't we seen a single slot aiv partner liquid cooled cards using full covered blocks and rad mounted pumps similar to what we get with the ek predator so I think a lot of that answer is it's cheap and relatively easy to source a c LC from a Sutekh or from does anyone else even do them on GPUs I don't know basically coolermaster Corsair coolermaster EVGA are the people who do liquid cooling gigabyte does as well and the cooler for Andy and I think last I checked for gigabyte came from cooler master have to check my own content on the I actually it is cool master and then the coarser one on MSI is an ASA tech supplied cooler EVGA is an ASA tech supplied cooler I don't know that cool it does anything right now in that space but either way it's quick and easy to throw one of those on there cool the GP with that and those pumps are big so they're not going to really fit single slot which is the main answer to that question as for why not get into a full covered block well at that point you're actually designing something and that takes money time things like that I was going to say effort but that seems dismissive but basically just money and time and money and time are not something that a lot of these companies will put into a closed-loop solution because at that point you're really most the way to an open loop solution anyway and also you start running into having to custom design a pump or something like that unless you can source it from someone else ATX not going to supply you a pump that'll work with that kind of open loop layout unless you really start going down the custom path at which point your cost goes up a whole lot lead time to produce it goes up a whole lot maybe it's no longer relevant once it's actually done the card is aged but it's mostly time and money like most things in the industry and then at that point you're just again competing with open-loop anyway and some of these companies do make open-loop cards or blocks so at that point they're just competing with themselves next question is AJ Thomas who says those from YouTube what happens if a PC fails when does error etc is workload saved by each frame or must you start all over so this is from again the quad GPU video where we were run using blender to render an animation and the the answer here is it depends how you render it so we render in PNG s basically each frame is rounded out as a single file PNG file they're pretty big sometimes 30 or 40 megabytes something like that you end up with hundreds of number thousands of them depending on how long the animation is you drop it into something like premiere and then create the video that way where premiere will just export it as an mp4 whatever you want you could technically render it as a video file which I think I did once but the problem with that is like this user points out if you render a video file for something that's going to take you 30 hours to render you better hope that Windows doesn't update in the middle of it or something else like a power failure because then you have to start over so generally people render out single frames and then just put them all together and some Psalter at the end and create the video that way that was a good question next one is from inversion who says why are thermal pads preferred over their own paste this from discord for V RMS and MOSFETs on the GPU and/or CPU and is there any damage that can be caused if paste were used on the V arms and MOSFETs instead I guess so builds or jumped in there to talk about this one briefly and it's I'm more or less going to say the same thing a lot of it is because MOSFETs are irregularly sized and shaped they come in all different types of sizes and Heights and it's general I guess how densely packed the MOSFETs together so just easier to put it their own pad on there it'll work with pretty much everything and also the thickness matters because their own page you can really only get so thick before it just doesn't really work anymore and a lot of the times it Heights between the cooler and the MOSFETs is great enough that you'd want to use that we used like a point 5 or a 1.0 millimeter thermal pad or something just because it's for sure you're going to make contact and whatever extra there is what access will just compress on the service as you apply force so I think that's a lot of it and probably some kind of manufacturing reasons too but largely it's thermal paste isn't thick enough a lot of the time and then MOSFETs are all kinds of shapes and sizes so you just order all your thermal pads you cut them to some base size or a PCB and you throw them on there and that's why a lot of the time you'll see if they're all pads and put like bleeding over into places where they actually don't need to be any on some cards because they have bought the thermal pads for a different design cut them and they just happen to fit on this one as well so why not use it because it's cheaper to do that than to change your spec or change your supply or something to that effect next question is from tpv this one's kind of interesting TV said is there any information on how watching a youtube video at two times speed affects a channels profitability not sure how it's calculated figure eight my mess with metrics you know I was curious about this so the way YouTube does a lot of stuff now is watch time I think ad revenue ties into that as well so with watch time generally speaking the longer people watch videos the higher average watch time you have I think with YouTube red or something you get a bigger share of user X's YouTube red allowance per month I'm not actually sure how that works but to answer this question my guess is if you watch that design speed and you see the whole video I think YouTube just sees it as ten minute video or whatever I don't think it's easier that's five but I'm not positive on that I would I would hope and assume that it sees it based on video length and video seconds not real seconds and real video length because then that would that would just seem kind of kind of screwy if people are watching stuff at design speed it really suck but I don't think that's the case I'm pretty sure you just get paid based on the video like actual seconds in the video consumed not sure though but ultimately not a huge deal that question is our last question is from metric on who said this is also on a YouTube how many people actually work cameras Nexus Steve is obviously the host of the channel and most of the articles on the websites indeed written by him I don't know if someone else tests the hardware and he writes the articles so who else is there besides him a cameraman maybe someone else to run the test and the cat no no hate just honestly curious not sure why I'd be hateful so I yeah I host the stuff and then we've also got Andrew does the video work actually if you look in the description or the bottom of the articles it'll have like a byline where it's got like the author and then additional reporting and testing or camera works things like that but Andrew does the camera and editing a lot of it Keegan does some of the Cameron editing when we're travelling Patrick Lathan does a lot of the benchmark and Recife use these days he's written a couple of CPD reviews this year and some of the revisits like by 7:00 9:30 revisit article was written by him and then we take a lot of that data and put it into the video script as well and then Eric Hamilton helps collect news for the the news recaps we do every week and he also writes some of the articles and then we've got a couple of Ryan helps with some of the articles as well the weekly sales round ups on the website then we've got a couple of less frequent contributors so employees contractors contributors are all part of the company it's just a matter of how frequently is this stuff posted by that person and you know is it is a day behind the scenes type of job or is it something that's got their name on the article byline if they wrote the article but yeah but the core core teams basically Andrew Patrick and I'm working on stuff too and then we've got contractors helping out with all the other stuff so more people then it seems to be believed I think working on it but still pretty small team and I think that's all for this time so as always you leave questions below patreon.com slash gamers Nexus tops out directly and check out our new shirts again givers Nexus at squarespace.com links in the description below for everything else thank you for watching subscribe for more I'll see you all next time you