Gadgetory

this test was inspired by or what I thought was because a lot of the coverage online of these Noctua air coolers for thread ripper basically drew a sort of general conclusion that these knocks what air coolers are better than the expensive high-end liquid coolers that use smaller cold plates not like this one but like the ace attack ones so we set forth to see if that was true and how it held up because just looking at the numbers it kind of makes sense to some extent with a larger cold plate but then at the same time liquid should make up a lot of that gap so we wanted to test coolers that were purpose-built for thread Ripper against coolers that are not purpose-built for thread Ripper before getting to that this video is brought to you by synergy the software that lets you share a keyboard and mouse between multiple systems if you have limited desk space and multiple computers to command synergy removes the need for separate peripherals or a KVM and works as over the network software use our link below to get 50% off the home or pro version with ssl so here's what this boils down to if boils down to the phrase the Noctua coolers are better and there's a few things here I want to point out one as you'll see in a moment it's not impossible that they are better but more importantly what does better mean and this is getting into territory where we basically derived some of our testing methods out of the phrase X cooler is better than Y cooler because ultimately you can be better in raw thermal performance but at the same time it's not hard to just put a faster louder fan on basically anything to make it thermally superior to some of its competition that doesn't mean it's better as a product though so there's also better in terms of value price proposition there's better in terms of noise normalized performance this is a big one so if we set these coolers in this cooler and the crack in X 62 all 240 DBA output so the fans when spinning produce about 40 decibels of noise over the room noise level about 26 which one has the best thermal performance at a noise normalized 40 DBA that's the other aspect of the word better when it comes to cooling performance then of course wrong acoustic performance as well at the high end so that's what we're looking at we're looking at those things today and the testing approach here we will be expanding this as you all know we have liquid cooler and air cooler reviews for LGA 2011 or 20 X X whatever was sockets already for years now and those charts are big and we've worked on them forever but thread rippers new so doing some specific cooler test for thread Ripper because it is somewhat of a special platform including with this inner mac slick tech 240 TR for CLC probably is a better name on that but otherwise it's an interesting product and the Noctua Nhu 14s and Noctua NH u 12s coolers and then we have the kraken x 62 as a baseline with the round ace attack plate that really only covers the dies but not any of the outer surface area of the IHS we previously showed that the extra coverage does actually help quite a bit but also to be fair to those ace attack coolers because threader per includes a bracket for them they'll work they'll keep it within spec for sure but it doesn't mean that they're the most effective you really just mostly want to use them if you lack another if you already own one and you just want to put the bracket to use so for testing here we're doing a different approach than normally somewhat where we've got blender and we've got prime95 for our workloads and the difference here is that we're logging power consumption down the EPS 12-volt rails the entire time and that's because prime and blender can both to some extent power and load cycle and that we're controlling for that we're making sure that the performance is consistent we're all still in 40 DB a noise normalized tests we've got manual coverage spread so this is basically taking a graduated syringe filling it to a known amount of thermal paste applying it to the cooler to have the same amount every time and then taking just like a little plastic shovel and manually covering the IHS with the paste and then we've also got the same thermal paste for all coolers we've got power controlled for we're overclocked to 4 gigahertz at one point 3 5 volts and it's important because you don't want to use auto anything because it's not trustworthy for a test like this and we're also using TDI for the measurement so we're not using tea control because it shouldn't exist TDI is the more or less the actual temperature of all those 36 sensors within thread rapport a bridge together as opposed to tea control which is all those sensors averaged together with the 27 degrees Celsius modifier for whatever reason so reason TDI and I think that's most of it for the methodology now we're focusing on the X 62 for the ACE attack representation it's one of the best ACE attack made cool is out there it's an NZXT branded cooler that will be tested alongside this enter max unit which has a full coverage plate big difference there and then we've got knocked a for the air coolers we'll be expanding other thread report cooler reviews in the near future but this will get us started first to address the concern of outputting roughly equal power between tests we clamped the EPS 12-volt cables during all tests for all coolers then compared power consumption between runs 595 load cycles at somewhat predictable intervals shown in this overtime chart and we can synchronize our data averaging the only pull data from the peak cycle the heaviest power cycle produces the most heat as you'd expect so we're averaging from that height as seen in the charts the test aligns well from past to pass validating that our testing was conducted under the same power load each time there's some variance and you should also expect to see about a 4% power consumption reduction from leakage for every 10 degrees Celsius that the CPU drops but these numbers are largely the same that's what we want the same is true for blender which produces a highly predictable heat load and power load throughout testing we were able to align the data to validate accuracy averaging again from the peak values the first test is prime95 torturing the CPU at around 310 Watts load with our 4 gigahertz overclock in 1.3 5 volts and using 100% fan and pump speeds the enter mac SelectTech 240 millimeter TR for cooler performs the best of the lot when disregarding noise plotting an average TDI of fifty point seven degrees celsius over a Meads with the 10-second peak at around 51 degrees Celsius delta T this positions the cooler significantly ahead of the NZXT crack next 62 with a more limited Coldplay contact area but a bigger radiator which is at 63 degrees Celsius over ambient where's cooler in the test was Dean aqua Nhu 12 as a 120 millimeter smaller version of knocked was higher performing nagu 14s the 120 millimeter option runs an average TDI of 66 C over ambiens achieving similar temperatures to the $160 excellently to cooler but priced at $70 USD the idle temperatures largely follow the low temperatures in terms of positioning as you can see in the chart but overall there's not a big difference at this low load level the knocked over 140 millimeter option performs notice ly better than the 120 millimeter option it makes sense the 120 has both a smaller fan and a smaller heatsink as you can see and so that makes sense in terms of performance versus the liquid coolers the 140 option maintains 59 degrees Celsius over ambient and that's still roughly 8 degrees warmer than the full coverage enter max 240 millimeter cooler but it's 3 to 4 degrees cooler than the ASA tech X 62 so just to be clear this one's a couple degrees cooler than the X 62 it's not tremendous likely you've seen some tests online but it is definitely there and it's measurable and it's outside of error so there's a couple degrees better than the X 62 and it's it's quite a few degrees worse than this one but we're not noise normalized yet so we'll see how that changes once we do account for the noise levels because this is quite loud compared to that when both are at max RPMs makes sense it has two fans and they're smaller as far as the ACE tech unit it's sort of spinning its wheels on the pavement this is an issue where ASA tech schoolers with those smaller cold plates they have the potential to cool it's got a bigger radiator it's got bigger fans they're half decent fans it's got the potential but it can't apply that power to the road there's no torque there so that's the problem of this smaller cold plate and hopefully a stack if thread ripper is popular enough for them to produce something for hopefully they actually make a cold plate that's threaded per size to compete a one big thing I question here until we tear this animatics went down I don't know if the actual liquid flows through this entire cold plate area or if it's just exposed cold plate because it's entirely possible that performance could be even better if the channeling design of the cold plate is specifically made for thread Ripper and it's likely that if you for example centralize your pump in your liquid in sort of the normal area you would for a cooler but you just expanded the cold plate like they've done it's possible there's still even performance on the table even though it's objectively the best in that first test normalizing for noise at 40 decibels should give us a better indication of whether the liquid coolers hold an advantage that matters it's easy to win a test when unconstrained a noise just put a louder faster fan on it but for these tests knock to us coolers we're already operating at about 40 decibels when Max that's their baseline performance so we didn't need to rerun them the other coolers though did undergo lower fan speed testing to fit the noise profile at 40 DBA with prime95 the kraken f 60 to only increases from its previous baseline by about 1 degrees Celsius we're at around 64 C over ambient now versus the previous 63 C not a big difference and ultimately doesn't really change the fact that the cooler which is otherwise good has become limited by its own cold plate it's still acceptable insofar as you could buy one and use it if necessary but there are more efficient than cheaper options that would better warrant to purchase the only cooler warmer than this is the NH u 12s which again is a 120 millimeter small air cooler which operates at around 66 degrees Celsius over ambient not that much warmer but a bit the Noctua Nhu 14s tr4 unit keeps the overclocked to 1950 axe to around 59 C over ambiens making it about five degrees cooler than the X 62 at about seven degrees cooler than the 120 millimeter Noctua cooler at our Maxis lick tech 240 TR for cooler leads the pack at fifty four point eight C over ambiens reduce units previous eight to nine degree lead over the nagas 140 millimeter cooler to a lead of four degrees Celsius animatics cooler technically performs the best here once again even noise normalized but is also priced at 130 you the versus the $80 price of the Nhu 14s there are other cons to the knock to a cooler of course but we'll get to those momentarily for a lower thermal load we now look at blender with the fans at max rpm the kraken X 62 maintains a 51.6 degree temperature keeping it still in the 2 to 3 C range above the Nhu 14s enter max Luke Tec 240 TR for cooler keeps the lowest temperature as previously at 43 point two degrees Celsius with the Nhu 12s at 60 C blender at 40 decibels also shows familiar scaling we're at fifty 4.5 cm x 62 with the Nhu 14s at 49 C as it was already operating at 40 DBA with the enter max 240 cooler a few degrees cooler than the 140 millimeter knock to a cooler but it's otherwise unimpressive in its lead it is technically superior it's just sort of a shallow victory that said liquid is always going to outperform air and otherwise similar or identical conditions so there's no surprise here it's just better at moving heat around the difference comes in how the cold blades made and whether the liquid is actually enabled to leverage its advantage or whether it's left again spinning on the pavement then a 212 ass is rather unimpressive and should probably be skipped in most cases as the U 14 s is significantly better at a similar price there are a couple of other factors to consider of course for example clearance this cooler is a bit larger and even by knocks wasone admission it's just not going to work with every motherboard if you're trying to accommodate a top slot PCIe device so for our testing when we tested this one on the ACS Zenith motherboard we had to take the video card and move it down a couple slots to the next by 16 slot because the top one just didn't work it would collide with this and even if it kind of makes it kind of clears be careful because you don't want a bridge contacts on the back of the cards with these fins that would be bad so that's a downside potentially you can use a lower PCIe slot if you have to you can do some manoeuvring with these but not a lot keep that in mind as for the rest though it comes down to price and performance technically speaking in the most sort of pedantic way possible this cooler is better it's better noise normalized it's better outright at Full Tilt but it's 130 ish dollars and this is something like 80 or somewhere around there so that's certainly a consideration there's a consideration of how much do you need the extra couple degrees in most instances we're seeing noise normalized about a four to five degree celcius improvement max for this cooler 4 to 5 degrees isn't a ton if that matters to you you probably know who you are and you're probably buying something better than this anyway they have a 360 option as well we'll test eventually but it's a good cooler for something that's got a threader precise plate it's also enter max which means that if other companies decide to get into the game here and produce their own full-coverage copper plates or liquid coolers we would expect that there's potential for more improvement there enter max is first to market with this which means that they're probably cutting corners somewhere I mean I would expect to see things improve with time that's just how product development works even from mentor max so you'd expect that the future iterations would improve more but as it stands today it's really not that distant of a fight between this and and this one the Nhu 12s I'd say skip entirely unless you're in a situation where you you really need that clearance for that top PCIe slot and you also need an air cooler for some reason which is also a consideration for liquid because if you're in a really tight case and doing something with a smaller board which there's not a lot of this thread Ripper because the sockets so big but that's a potential use case for liquid there's a smaller footprint so keep that in mind but there's your data though as far as my thoughts on this so 1 into it like I said expecting that the something like an X 62 which is one of the best days of tech coolers would perform more or less equally to this cooler if not slightly better because it is a good cooler but the fana is that quote the cold plate in the limited contact area really comes back to bite a Sutekh in ways that surprised me a bit because what happens is this one ends up leading by a couple degrees now it's not as big of a lead as we've seen some other places but it's still somewhere in the 1/2 four degree range depending on the test and the noise levels things like that once he normalized for noise the Kraken loses a little bit more ground but not a lot so that that was genuinely a bit surprising to me just how much the cold blade matters we kind of tested it on knocks or previously between knocks with coolers but not against liquid and that's interesting I mean that's basically that's where stuff like this companies like NR max who aren't really tier 1 brands are actually going to have a chance at making some sales here because they were early to thread Ripper with a closed loop product that actually does perform objectively better than the smaller circular cold plates of vasotec products that have been on the market a while now so maybe this little kit gains attack a bit and wake them up to make a full-coverage plate but it depends on if they think there's volume there to actually profit on it thread Ripper is still kind of developing as a market it's definitely the biggest upset that we've seen so far in terms of knocking Intel around but even at these with the lead that thread Ripper has in some scenarios it's like $1,000 part for the high end so it's gonna be lower volume even if they're selling more than Intel might be in that particular segment so keep that in mind that's all for this one you can go to patreon.com/scishow stew help out as always you can subscribe for more and you should because we'll have more of this covered up soon and store doc gamers access dotnet to pick up a shirt like this one I'll see you all next time