Gadgetory

we already have a dozen or so content pieces showing that deleting can improve thermal performance of Intel CPU significantly but we've always put the stock Intel heat spreader back in place today we're trying a $20 accessory it's a cnc machined copper IHS that we bought from rocket cool which purportedly increases service area by a claimed 15 percent and smooths out the points of contact Intel stock IHS is a nickel plated copper block but it's smaller and exposed surface area than the rocket cool alternative at fourteen to twenty dollars it seemed worth trying out today we're looking to see if there's any meaningful difference in thermal performance for the Intel i7 8700 K when using the stock versus the aftermarket copper IHS before getting to that this video is sponsored by Dollar Shave Club makers of the new five dollar shower shave starter set the kit includes the Dollar Shave Club executive razor bearing the EPS of any high-quality tool and also includes reloadable cartridges the five dollar kit includes everything you see on the screen now like the body wash and shave butter and can be refilled for a few dollars a month this deal is available for $5 exclusively using our link below or dollarshaveclub.com slash gamers nexus this is pretty simple on the surface of it to get everyone up to speed and integrated heat spreader by intel that sort of silver looking one is actually copper it's just nickel plated and it's not the shiny kind either so it is a copper IHS or integrated heat spreader this is a copper IHS it's not nickel plated it doesn't really matter a lot what matters more here is that this has a slightly larger surface area which is done just by kind of the Intel IHS has some more lower down dead zones around the outer edges of it it's made to fit basically all the Intel CPUs of the same socket type so it's not like they custom tune it this is custom tune 2 for LGA 1150 RIA so it's ultimately surface area that matters because that's increasing the amount of contact area between the IHS and the cooling element on top of it or the heatsink I should say the cold plate so that's what we care about replacing it's pretty easy it's you can watch any of our d-lighting videos it's the same process you've popped the CPU into a tea litter you deal it it and then you pull the IH ace off and start to replace it with this one in the process of doing so there are a couple things to watch out for like the alignment of these ridges on the outer edges there is a guide that's pretty good you need to basically make sure that the arrow is aligned with the arrow and the D litter and then you just mount this on to it the only downside so there's a couple things to watch out for here reel it in or resealing a component whenever you reseal an Intel CPU there's a really good chance that you just killed all of your D lighting performance improvement because from doing this several times now for multiple youtubers as well I can tell you ideally you do not reseal it with silicone adhesive you leave the thing deleted yes you put the IHS back on however when you put it back on we recommend that you just sort of you drop the cpu into the socket without ever moving it any other way you hold the CPU and the IHS like this in transit and don't let it slide around but also don't reseal it so the downside to that is you can't really move the CPU in any kind of normal way to be really careful about it so if there's no sliding you don't have the liquid metal run over onto components on ties on these or you don't have the liquid metal just loose contact in general however the upside is that you don't have a sealant around the edges that will increase the sort of Z height between or the gap between the IHS and the die that's what tills performance and that's actually on an Intel CPU more of what kills performance in some cases than the thermal paste alone so it's you want to avoid silicone adhesive resealing so for our testing we tested without resealing the cpu i dropped it into the socket with the liquid metal applied how I wanted it and that was just the bare substrate with a die and I put the liquid metal on this side it's kind of I mean it's it's basically stained now that's not actually liquid anymore but I put the liquid metal on this side and then drop that just on top of the CPU right I wanted it as the instructions recommended latched it down and then put the cooler on that's what I would recommend because when I did try resealing this because it is easier to transport it but it just it kills the thermals and you introduce a lot of variants it's not worth it just just do it without the sealant so let's get into some of the data here test methods we use the liquid metal conduct or not so this is a thermal grizzly conductor not that we used between the die and the heat spreader on both the Intel and the aftermarket solution we ran three tests and for each of those three tests three tests per i HS so six total times an additional three tests so like eighteen and so we had basically three sets of frequencies and voltages tested three times each so three iterations and then in between each of those we reapplied the liquid metal and the thermal paste and then averaged all the data throw it on anything that looks bad because there's a lot of human element involved here where application of thermal pay is on top of the IHS can very application of liquid metal under the IHS can vary so hence running three different iterations of them or test passes and averaging and it looked pretty good after all that so fairly consistent overall we established a plus or minus about point seven degrees run Tehran variants with those reapplications and additional information on testing will be in the article linked in the description below as always but we can get as the charts now so this chart is really what we ultimately ended up with that five gigahertz and one point four two volts we end up generating between two hundred watts and 220 watts at the EPS 12 volt cables that configuration isn't on here it's just in between the other two so depending on the test that's where you're at for wattage 200 to 220 watts and they give or take couple watts of course for the two tested voltages here stock temperature monitoring showed 49 point five degrees Celsius over ambient for baseline average of all core temperatures and that's also averaged after a 30 minute burnin mark on our X 62 with max fan speeds which allows the radiator to soak and hit equilibrium the peak ten-second high was 51 point six degrees Celsius with liquid temperature at thirteen point eight degrees Celsius for the rocket cool IHS we measured forty four point five degrees for the same test or forty eight point two across the peak high and liquid was roughly equivalent within margin of error these differences between the temperatures are just outside of error margins and we can only confidently say that they are outside of the margins of error because we ran the test enough times to establish what kind of variance we had which was not a whole lot so with each one of these using a new liquid metal and paste application we end up looking pretty even across the board scoring about the same from run to run this helps ensure that we're using the same application technique each time we're all seasoned graduated syringe for the thermal paste to make sure that it's an even amount each time the result is a slight temperature improvement with the rocket cool solution roughly in the range of three to five degrees Celsius so not at all bad for a copper brick that you replace the Intel one with and at one point four four volts and five point one gigahertz we saw nearly identical Delta's to the previous the stock IHS reported fifty six point eight degrees or 60 degrees peak high temperature and the rocket cool IHS resulted in a fifty one point eight degree average or 56 point five degrees for the ten second high the improvement again is a range of about three to five degrees depending on how loose you are with your error margins and liquid temperatures are about the same as for temperature over time the IH has changed doesn't meaningfully impact or elongate the heating period in our test setup the rocket cool kit is routinely shown below the stock IHS by a few degrees but there's no actual change the ramp up or the ramp down time for temperature at least not one that exits error margins one more reason that the custom copper IHS helps here is less about the extra surface area and there is more because if you look at the Intel one it's got that lowered surface around the edges and the custom one takes the whole service area but it's less about that and also about the flatness of the surface the intel heat spreader stock is actually very slightly convex and the asa tech cold plate some of the other cold plates on the market are we concave this is to somewhat account for the bend in the Intel IHS but not everyone does it and also you've got two modes of manufacturing tolerance to deal with there where now you're relying on both Intel and the cold plate manufacturer to align perfectly with their concave and convex bends this is is flatter it removes that it's actually basically perfectly flat and ultimately when you mount it all down anyway most the cold plates kind of become flat just from the mounting pressure to begin with so that's another big contributor aside from the extra service area as far as the copper versus the nickel plated copper that probably has very little to do with anything here it's all copper at the end of the day so what we're really looking at is the other factors mentioned overall that is pretty straightforward we got a couple degree improvement and honestly it's more than I was expecting for it's $14 without the real ed tool if you already have the d Lloyd kit it's a bit easier to get with but I had twenty dollars for the complete kit it's really not bad I was pleasantly surprised with how well this works now when I say how well it worked I mean relative to its price relative to the effort involved and relative to what it is piece of copper now it's a nicely machined piece of copper and rocky cool did a great job of it this isn't something that's going to change a lot for you in terms of your overclocking Headroom it's probably not going to change a lot you for you in terms of voltages it's all gonna be about the same as where it was before you might be able to very marginally reduce your fan speeds you might be able to very marginally reduce your cooler size but this is it's it's one of those enthusiast items that if you asked me is it worth buying straight out no no extra is tagged on to that at all not is it worth buying as an enthusiast is it worth fine for budget user be just straight ask me is it worth $20 I'd probably say no in terms of just I mean it doesn't it doesn't do a ton for you however from an enthusiast perspective in terms of you're already deleting the CPU already putting liquid metal on it I'm actually pretty happy with it so I would buy one but I would also buy not expecting a lot out of it you get a couple degrees what does that really do for you if you already know if you have a target in mind I need an extra couple degrees to get to this threshold that will numerically satisfy my OCD or something like that if you already have an objective then yes it's worth picking up I wouldn't buy it expecting a lot in terms of overclocking but it's never bad to run at a lower temperature either that's always pretty much a good thing in 99% of cases so yeah it's a pretty happy with it and I would recommend it if you're just kind of like you want a Saturday project you want to play around with computer parts for 20 bucks for an hour so it's pretty good project for that but outside of that it is a bit of work it's not hard but I did have to reapply it a couple times that when I first went through with it to make sure I'd aligned it correctly make sure the liquid metal wasn't sliding around underneath just like any dealin process I mean yeah don't expect great things but you can expect decent things out of it and I think other than that do keep in mind that if you try to validate your own thermals run to run like before-and-after it is pretty hard to actually run thermal test properly you'll want to use probably something like hardware info to log and do it over time don't just spot check it and then run averages and you'll also probably want like a current clamp on the EPS 12-volt cables so you can make sure these software is pushing the same amount of current into the CPU each time because running blender or prime or whatever pass a to pass B there can certainly be variants of sometimes 20 watts and that's enough to eliminate some of the advantages you might see or even you can make one look worse than the other if the power input is different so if you're trying to validate your own numbers make sure you control as much as possible and get something on there to read the power and then you should be good hardware info does a pretty good job at reading the wattage from most CPUs you can use that for pretty much everything in lieu of a current clamp so yeah overall very quick recap here is for $20 I do really like the idea of just a fun Saturday enthusiast project because it's so cheap that it's basically something you do because you haven't messed around with your computer in a while and you feel like that'd be fun there is not much more reason than that that I would buy it you know you might as well just buy a better cooler if you're really digging for temperature improvements and the deal is do a lot for you anyway but it's still admirable for you know a small thing like the Hat get a couple degrees out of it so that's it for this one pretty fun if you have suggestions for similar stuff like this like small computer component accessories enthusiast type products that you don't see a lot of tests for we like working with these it's fun to do it's not a big expense to buy it so let me know in the comments below what kinds of things you've seen that you'd like us to try out and subscribe for more as always and go to patreon.com/scishow self-directed go to store doc gamers nexus dotnet to pick up a shirt like this one or one of our other products I'll see you all next time