Video size: 1280 X 720853 X 480640 X 360
Show player controls
Watch our first contact frame review where we detailed the problem (and why these exist) here: clip-share.net/video/Ysb25vsNBQI/video.htmlThe best way to support our work is through our store: store.gamersnexus.net/
@mypastlife I would personally. Just to be on the safe side.
Is this something I should have for my 13th gen then? Before I put all this stuff together?
Does this work on new 13th gen intel procs? i have Z690 mobo and i7 13700KF, can i use thermalright frame?
on 9/9/22 on amazon the thermal grizzly contact frame is 69.99 and the thermalright contact frame is $13.98 for the red and $12.90 for the black... and in your own testing it performed 1-2C cooler than the thermal grizzly. Didnt you say in the opening segment that if you are using say an SFF build its very important to get every C out that you can? Then when this averaged almost 2C cooler in your tests you went meh its basically the same. Well which is it? Are we fighting for every inch or flying high and carpet bombing at 500mph?
COPPER IHS. AND COPPER CONTACT FRAME. future of temperature control.
Thank goodness there are ways we can fix the problems caused by these multi-billion dollar companies, heaven knows they can't afford this kind of extra cost.
The ILM is much less prone to user error destroying their CPU. Screwing them in wouldn’t scale to all their customers.
if we dont see an improved retention mechanism for next gen chipsets for lga1700 its gonna be a big L for intel
@aceofspades CPU IHS is also convex...before it bends. But yes, the flat cold plate of the Arctic does illustrate the extreme end of the issue, but it's far from the only CPU cooler with a flat cold plate. Most open loop water blocks are flat, too. The issue is reduced with a convex cold plate that can sink into the newly formed valley caused by the ILM tabs only contacting 2 points on the sides, which is why not everyone has problems.
Eat the Rich- XD
@H311FR0G lol, you really just need to move on. Experience != knowledge, and this is evident here. I never called into question your experience, people can have years of experience and still lack knowledge within their field. You seem to think because you have this experience that you're highly knowledgeable, but this is a large misunderstanding.As for my experience, if you want to get into a measuring contest. I avoided mentioning it, as again, experience is not knowledge. I have experience in network and security administration, database administration, full-stack general and web application development (language experience ranging from assembly to C/C++/C#/PHP/Python...etc), electrical engineering, computer engineering (kind of related these days), with experience in many other fields. I'm quite big into cross-discipline. I originally pursued a mixed discipline like mechatronics, but found that while I can certainly do mechanical engineering, it wasn't a strength of mine. "Not something you can easily do EVEN WITH engineering"I've built hundreds of computers, and serviced them at just about every level as well. Also have done electronics repair and reverse engineering. Don't think I'm just a desk guy that fiddles with high level scripts or something."yet I, with as you put it "No Engineering knowledge whatsoever" have done so numerous times."Even the experience you outlined is not engineering experience, and still doesn't show engineering knowledge. Repair is a skill, sure, but it doesn't require engineering knowledge, and I have seen many individuals repair things without understanding the why of the part, or the specifications, or the placement...etc. In my experience actually, most that do repair are lacking the engineering, which isn't anything against them, it is still a valuable skill. I belong to a few different repair communities/groups online, and I found the extra engineering detail in my answers often times is unnecessary for them as most repairs are identification and replace, the extra details are just excess information/noise. So, again, I commend you on your repair skills, but that isn't engineering. Confusing these two items will likely lead to a larger assumption of knowledge than actually exists. A capacitor for example has a specific capacitance, can interact with inductors (unintended or intentional like a resonance circuit), resistors (affecting its RC constant), have its own resistance (ESR) and inductance (ESL), interact with parasitic inductance/resistance within other parts of the circuit, physical position of its placement relative to other circuits...etc, all to produce different behavior and form many different configurations. All of such configurations are then chosen for specific purposes, like specific voltage fluctuations, noise filtering, DC blocking...etc. This is just for a single capacitor in a circuit. For repair, you rarely have to consider this level of information."As to the accusations of Conspiracy, I said nor implied any such thing stating that Intel clock speeds are Artificially inflated"Maybe conspiracy is the wrong word here. You're at least implying deceit on Intel's part, though you're simply ignorant of the reasons behind the decisions and so assign malice where there need not be."You don't have to know a thing about the actual engineering of the device you're programming it to emulate by the software you are pushing through it to make it work, just the numerical specifications of the hardware you are instructing it to emulate"I don't think you understand that FPGA's are using NAND gates, aka, 'universal' gates, which allow them to form any configuration of any of the other logic gates. The FPGA's are programmed either using higher level logic that outlines what you want it to do, AS WELL AS the specific logic gate structure you want. Meaning, you're in fact basically designing the chip hardware... this is why you use what is called hardware description languages to outline this, like Verilog or VHDL. The exact logic gate structure you're defining corresponding exactly with known transistor configurations that form those logic gates, which correspond exactly with physical transistors. Since you lack knowledge in this area you don't understand that this is how modern day CPUs are defined as well. The CPU components (IP blocks), and glue logic are defined in a hardware description language... this is actually why CPUs can be developed with an FPGA. You can literally run an ARM CPU design on an FPGA, though I'd imagine hardware emulation is also used with the big guys like AMD/Intel/Apple.The main difference between an FPGA and a chip that is fabbed later is that the chip to be fabbed goes through an additional physical design stage. FPGAs are also used to physically test chip designs ahead of time... because FPGA programming is effectively chip design minus the physical layout. Heck, FPGAs are even used in production environments in place of chip fabbing in instances where it is more costly due to the overhead."Thus it is Artificially Inflated, the numbers making it appear 125% faster despite being the same speed due to the increased latency of the longer pipeline. "... again, your lack of knowledge here is hurting your understanding. In chip design you can use your transistor budget in multiple ways. If you have a given die area and you're on a specific lithographic process defining transistors per area, you thus have a certain transistor budget. Pipelining can increase latency but increase throughput, but it also can increase clock speed compared to a non-pipelined circuit as you can decrease the logic path depth... it's a complex topic that I can't fully describe the pros/cons here as it depends on the circuit and what is being pipelined, like if you can fully parallelize a logic circuit and have low depth, that is better than a higher depth circuit which is likely to be pipelined. There are also such things are instruction pipelining... this is all covered in the instructions per clock metric (sort of, more so this and the interactions with other components required to execute an instruction, but I'm simplifying). What you seem to miss is that a simplified total computational power = (instructions per clock) * (clock rate/frequency).In chip design this depends entirely on your goals with your transistor budget. There is nothing artificial here, and optimizing either way is 100% valid. You may use a more pipelined approach for a specific IP block and then a less pipelined approach for another, if you view the other as having more important speed implications, it's all an engineering tradeoff game. What really matters if how many computations your system does in 1 second, which is covered by the total computational power mentioned above. Intel has worse instructions per clock compared to AMD's, yet, Intel can have higher per thread performance simply because their clocks are real and are higher than AMD's, the math simply just works out that way. The downside is power scales with frequency, so Intel CPUs are more power hungry. Again, all an engineering trade off."Thus it is Artificially Inflated, the numbers making it appear 125% faster despite being the same speed due to the increased latency of the longer pipeline"Again, not artificially inflated, just a different design approach. If someone is comparing clock frequency to determine computational power, that is on that individual for having an incorrect understanding... it certainly isn't on the engineers who are simply engineering the device. Clock frequency is only comparable within the same architecture, and any proper technical user understands this. Don't assign blame to the engineers or a company when this is a known incorrect method of comparison among technical users."And even more so after pointing out that Intel has only led in real world performance or marketing during periods with no competition to speak of. "And I don't necessarily agree with this, though Intel did get lazy when competition is less."I love how your overall tone changed after your self admitted insulting reply was hidden"I had actually edited my one post before I realized it was hidden, to be a bit less insulting. I never outright went for low blows, but I was too harsh in my wording. Are people not allowed to get heated in discussions and pull back? Why not?"Yet you still managed to throw a 15 foot wall of text back at it within an hour. "So... I'm not allowed to be watching youtube and seeing a reply notification? Further, I'm a fast typer, so a wall of text like this one only takes me about 10 minutes to type. " Stick to your FPGAs guy. You don't have to know a thing about the actual engineering of the device you're programming "... again, just shows your lack of knowledge in this area. You also assumed I have no physical design experience, even though I only claimed to be "more experience with FPGA," more is a hint. " Good luck in life."While I have been sharp-tongued periodically, I have never made it personal. Stating you lack knowledge isn't stating you as an individual are of less worth or some such nonsense. Try not to take such statements personally. Rather, it is evident you lack knowledge in this area, which I tried pointing out in futility."The only time I'll waste on further arguments from you is hitting the thumbs down button, even if you get it right. "Alrighty, lol.
I find it really trustworthy of GN that they are willing to almost outright say that an older/current sponsors product might not be the best bang for the buck, and even recommend the competitor, this shows the integrity a channel has, I can't say many other channels always behave that way, yet another huge win for Steve and GN's credibility.
@DeSinc washer mod doesn't fix intels flawed mounting system. The issue is caused by trying to apply even pressure using a little lever.Washer mod applies less pressure to the cpu, it's not a fix.
@Katherine Silens Roman, like most tech tubers is chill. But it's not all good will. He knows his reputation brings in more revenue & what he can gain from anything like you said.Still respect and enjoy his products & content.
@Katherine Silens no one's taking an L here. You either get something made in China on bulk orders or something routed in derbauers garage (almost). They pretty much only sell through like caseking or overclockers anyway its not like they're looking to get some huge volume, otherwise they'd actually scale it up.
Also props to der8auer, the guy behind the Thermal Grizzly product line. He doesn't have a problem with this and would probably agree with taking the L for customers who don't care about the quality difference, while most traditional companies in the space would throw an absolute hissy fit with sample threats, blacklisting, and bribery.
For me it was as simple as finding the thing. Thermal Grizzly's version was out of stock literally everywhere I looked, and in a lot of places had a non-insignificant markup over MSRP. The only way I could find one available was from retailers in Germany itself, with a higher price and on top of that the cost of international shipping and other associated fees. Or you could just grab one of the Thermalright ones off of eBay with plenty of stock for 12 bucks.
Same here, first i waited for the tg to be in stock but then just bought the thermright frame on ebay for 13€ and I'm totally happy with the results .
Yeah I'm in the US I can only find them for like $80 right now.
@MrJethroB Then why are you here, lol?
@MrJethroB AMD is just better ™
@Mame A All my gear is AMD
The difference in machine precision between TG and TR and the fact that they gave essentially the same results is a testament to the effectiveness of thermal paste in filling the gaps.
Thermalright has been killing it. Wish you'd take a look at their coolers they seem to be on par with top end air coolers while priced way cheaper.
Thermalright *is* a top-end cooler-- they pretty much invented the category. Literally every modern "high end air cooler" you see on the market is a clone of Thermalright's products from ~2000-something.
Can't remember the last time something so simple could be so easy and cheap, and still have a more-than-trivial improvement.
socket 1151 and 1151v2 with thin processor pcb, thermalright also fixed that with free plastic cover, that must be placed between processor and socket brackets, to make less cpu bending under cooler pressure
@Ceil Yurie Around the late 90s/early 00s CPUs were packaged on their own PCB (printed circuit board) and said PCB would slot into the motherboard like a modern video card. The pencil unlock was after AMD and Intel moved back to a socketed CPU and both camps locked their CPUs so they couldn't be overclocked. Why give your customer free performance if you could charge them for a higher sku instead? But AMD foolishly left the circuit they purposely blew to lock the chip exposed. Drawing on the CPU with a pencil the graphite would rebridge the connection and unlock the CPU for overclocking.In those days a 20-30% overclock was not unheard of so overclocking was insanely popular. Since then AMD and intel have realized they can overclock their own CPUs via "boosting" them and charge you for their full performance. This is why overclocking gives such anemic results today without extreme cooling and voltages.
@Ceil Yurie Google it... as well as your grammar. SMDH! 🤣🤣
@Zodwraith I remember unlocking them with a pencil lol
@Skorp Being born leads to death in 100% of cases, I prefer to focus on the in-between glass half full
Thermal Grizzly wasn't the first one to release this, Thermalright was. So it's disingenuous to say that this is "continuing what Thermal Grizzly started" when the Thermalright product was on the market months prior.
@Gvazdas I am not talking about the method, I am talking about the end product. This is a low tech product. It does not need the precision the expensive vendor claims it has. It's a piece of metal with no function on its own. As the video shows, wether it is made cheaply in China or expensivly in Germany (with machines that were probably made in China) makes no noteworthy difference. With enough patience you could probably train a monkey to do the steps necessary to produce these things.
@Die Finsternis only somebody with 0 industry experience can claim that precise machining is "low tech"
@Die Finsternis Your saying craftmanship is silly ? "producing low tech in a high tech country" I'm talking about how others perceive snobbishness, not necessarily Germans themselves.
@Dac33Nr It's not so much about snobbishness as Germany actually having inklings of reasonable labor laws. They can't sell them at lower price while keeping a reasonable profit. (which doesn't mean you should spend 40$ instead of 5$ given the chance - producing low tech in a high tech country is just silly and borders at irresponsibility)
I agree, no need for Mercedes level snobbishness with price as long as it's good enough to give even pressure.
I don't really know how these got thought up honestly when four washers under the official intel bracket is actually safer and easier to apply and gives the same results. I'd love if you showed the washer mod compared to these frames. It's really simple, four 1mm thick nylon washers under each corner of the official intel bracket, which has been shown to get the same thermal reduction as these brackets. The advantage of this method is the pressure is always totally even. Screwing the intel bracket down harder doesn't change how it's positioned on the board, the bracket always sits the same way on the board regardless of screw thread tightness, unlike this bracket design which entirely relies on the pressure being very even.
I think Derbauer already went into why these frames are more advantageous vs the washer mod in a video on his channel.
Exploit master has spoken o7
I have been using the Thermalright for a few weeks now. Lowered my temperatures by 8-10 degrees Celsius. I strongly recommend saving money and going for the Thermalright. Using this with a 12700K.
Yep, 12700k, 8-10c drops. For about £8 and 15 mins work.
my temperatures > the CPU temperature 🤒
@Blocked Pot Nope, I stopped as soon as I felt medium resistance.
@Vlad Cristea hope you didn’t over tighten.
@Blocked Pot I installed it yesterday, honestly a bit disappointing in my case, maybe like 2C difference, though I noticed the random temp spikes to 50+ during normal use are almost non existent now.
10C is a gamechanging decrease in temps. It can be the difference that allows the 12900k to be run at full load on air, in warm climates without throttling.
@John K. if u can install aio this be ez
@mcl48YT 10C is disproportionately huge value for a such a small, cheap product. 10C is the difference between a cheap single tower air cooler and a 360 AIO, and this little shit costs like 10 bucks. If this is even half as good as seen here, it's well worth the (negligible) cost and effort to get.
@blueone I think he means for the possibility that mobo companies to include this frame as pre-installed. But as Steve says it literally costs a couple of bucks to make it. Not sure why is he so mad about it.
@mcl48YT tf are you talking about? Then just don't buy it, you're not paying more unless you CHOOSE to, and you choosing to or not has 0 impact on other consumers 🤡
@mcl48YT What nonsense are you talking about? Are you ok in the head?
The dumbest part of this issue is that Intel could have fixed it for even less -- probably $1 more for a better ILM, since it would still be high-volume-production stamped steel. They could even spec it to the boardmakers for the Z690 only.
@r Negoro Go back to the end of my original comment. I said they could make it required for the Z690 only. Also, even if a lesser application means the difference is only 3 degrees, that still means a quieter and longer-lived fan.
@concinnus 7-10 degrees enormous but most of us buy 12400 anyways , and we dont buy the overclockable asus rog board. I bought 12400, using stock cooler - and the stock coolers are actually a huge pain in the ass to install properly to get optimal temperatures.
@Drunk Husband Design is theirs.
@Scott Grammer I actually looked it up and Intel hasn't made mobos themselves since 2013, so that's why I thought that.
@Drunk Husband It may be that Intel was putting their name on mobos manufactured by others, but I have definitely owned Intel-branded mobos. They were known for two traits: Absolutely no overclocking abilities whatsoever, and superb stability if used with a decent power supply.
Considering the extremely low torque evolved in mounting the contact frame, would it harm the instillation to use locktight or super glue? PCs get moved around for maintenance and cleaning. With a cooler bearing down on the IHS (especially heavier Air coolers) the shifting weight from handling the PC might cause the IHS to rock under the contact frame. There are also other forms of vibration such as having a personal fan on the desk next to the pc. Now this is more likely to cause nominal board flex rather then backing out a torque screw, but I worked decades in industrial machinery. Everything moves! I have had bolts back out of static machines.Whether that movement is perceptible or not depends on how fast it is happening. Expansion and contraction is movement, vibration is movement, shifting orientation is movement. These minor forms of movement (at very low torques) could back out a screw. How would that effect the test results?
I was utterly baffled that ThermalRight sorta disappeared into obscurity after they produced the XP-120 cooler in the early 2000's. They were the Noctua of their time.
They never stopped being awesome when it comes to their cooling solutions, they just don't market as much as other brands, which is a shame, since I have been using Thermalright solutions since the early 2000s, they had some innovative solutions back then and they still do, but without all the fanfare and jacked up prices.
@Kevin the Caveman Yes. The Odyssey pads are excellent and TF8/TFX are excellent TIM especially for bare die application like GPU and laptops. Using both on my 2xGigabyte 3090 OC reduced mining temps to 50'C Core and 88'C Memory Junction, a great result. In games the Memory only gets to 65''C- 68'C max, really good for GDDR6X.
@sp33d4l0l still rocking the Silver Arrow. They sent me a free mounting kit for the new socket, pretty damn nice of them. Silver Arrow still works flawlessly and I bought it for the 2500k!
They never stopped making solid coolers (I've recommended their stuff many times) but for some reason they just kinda stopped receiving much coverage after their Silver Arrow series.They're right up there with Scythe if you're after value for money, in my opinion.
Isn't thermalright making some of the best thermal paste and pads around though?
I think people are side tracked with TG vs TR differences, but lets remember the fact that intel skimped(?) on this $4.5 part on a $800 cpu
I'm amazed how poorly Intel did, and that 3rd parties have solved it for them so cheaply. Thank you for comparing these products so carefully!
for pressure testing: I think you can calculate the pressure area so it can be used as a numerical approach and calculate a "percentage of coverage" rather than a visual judgment.
I love that you tell right in the beginning how much experience you have in testing that specific type of gear. You guys do your absolute best to be as transparent and objective as possible and acknowledge that, while they are rare, even someone with as much experience as you guys can still make mistakes every once in a while. Your professionalism and objectivity makes you easily the best computer hardware channel on Clip-Share, keep up the amazing work!
The content and information itself is as good as always, but I really have to say I'm enjoying the more slick production style you guys have been moving towards lately. Very clean and crisp, but also more friendly and inviting with more of the team being on camera. 😀
I was leaning toward Thermal Grizzly before watching this review. The main thing that swayed me toward the Thermalright was not the cost, but rather the mounting process being more straightforward (since I do not own a torque wrench), with no real trade-offs in performance.
Really enjoy seeing the other presenters in recent videos, it's nice to see the different personalities in the team!
Mike is very charismatic
Especially ones with short hair.
The amount of dedicated work you guys put in is just amazing.
The "Thanks Steve." I'm dead! XD Really nice review as always. This is the kind of stuff I love. Comparing the same product at different price points and drilling down on any differences, why they exist, and what they mean for the consumer. You guys are awesome!
It would’ve been neat to see the flex or bend tolerances between the two and maybe between a few of the cheaper ones. Though I guess the impression testing might show that to a degree.
I'd lay a painter tape over the plate's holes, then make a hole on the tape where the screw hole is.Mark the tape with lines around each hole, just like thermal grizzly.Mark the screw.Install it like thermal grizzly's (using the lines as a pressure distribution marker & pseudo torque counter)Take off the painter's tape.
ever since i've discovered this channel, i've loved every upload no matter the length. Steve is such a great commentator and is very knowledgeable, could listen to him talk about computers for hours
@Potatoes I thoroughly enjoy it.
@elijzh Yeah on the technical side GN always wins they'll present you the thing as if you already have a very good grip on IT
@Larry Kounlavong I don't really enjoy much of jayz videos, can easily get any information he is giving out from a different creator in a much more concise method. And not much of a fan of the humor he has either
Agreed… doesn’t stretch irrelevant content like jaytwocentz in my opinion cause that guy can go on and on about something that can be explained simply with a few sentences.I still watch his videos though… just more selective what I watch however.
Thank you for the kind words!
Nice. I've bought Thermalright products in the past and they've been decent. I don't know why Intel didn't just use the two lever system like the 2011 though. It would have resolved this issue.
I'm really hoping to see a future where motherboard manufacturers are introducing this sort of improved design on their own. Or at least include a kit for their higher end boards if there is some sort of Intel requirement to use the original ILM.
Great to know the Thermalright is a viable budget option. Now all I wonder is if there's any benefit to a frame at all if we're using a cooler with a convex cold plate (i.e. Asetek) as opposed to a flat one.
I’m blown away from the temperature reduction. I can not understand how a heat interface so far from the core can be this bad.
Thermalright is an old company and has had solid products for a long time. I have no issues trusting them. Currently rocking a 10 yo AXP100 in a sff case.
@Kepe Thermalright kept with the times, their current coolers are just as good as Nocuta only 50% cheaper. They also pioneered all this stuff, including the spring clips for fans which Noctua copied a month later. For some reason Thermalright doesn't market itself, that's all about them being obscure nowdays, don't know why, but the quality and results are still there.
@Bearded Frog Hehe, CPU cooling has come a long way from those days. Still remember the Swiftech heatsinks, ones which had a heavy copper base and just a ton of thin steel bolts screwed into the base to increase surface area? Swiftech MCX462-V for example. There was some pretty crazy stuff on the market back then.I remember having one of the first tower coolers with heatpipes, it was an Aerocool tower that fit a 92mm fan. It was cooling an Athlon XP 3000+.EDIT: I think it was actually an Aerocool HT-101 with an 80mm fan. It was pretty much the first heatpipe tower cooler on the market. There are still reviews online which make quite interesting reading.
I still remember the Thermalright SLK-800 and SLK-900u copper heatsink back in the old Athlon XP days. 90% sure I still have a 900U in my parts storage somewhere. A CPU heatsink that was massive at the time, and made entirely out of copper. They were the best of the best outside of water cooling which was still pretty underdeveloped at the time.Normally you'd pop a Vantec Tornado fan on it for heavy overclocking. People who complain about fan noise now have never heard one of those lol.
Thermalright was THE heatsink company back in the day before Noctua was a thing. They've been kinda quiet for a long time and their products aren't really available at least here in Finland anymore. Maybe they've just pulled back (partially) from some western markets? They do put out new products but I've rarely heard of them in western tech media, except for the Assassin King 120 which has been tested by GN for example.
Idk what happened to them. The TRU 120 was THE STANDARD for a very long time. I had a Venemous X that cooled an AMD FX series for a decade.
I bought one of these brackets, I have to say I'm pretty pleased with it! My CPU no longer thermal throttles and I'm at a 5.4ghz to 5.5ghz overclock! I wasn't aware of the torque issue at first, so I just went by feel based on how surprisingly loose the screws were at stock. I had no problems, my PC booted and the RAM was all there. I did get a torque screwdriver later on, just to be sure though.
Good to see Thermalright making something good again.Their Macho CPU cooler was one of the original huge coolers and seemed to have set the stage for the nh-d15, deepcool assassin III, and dark rock pro 4
I prefer the bottoming install for the thermalright frame as well. TG's 1/4 turn spec is awfully close to not compressing the pins past intel's .2mm contact height variance. (depending on screw thread pitch) If you have a couple pins sitting .2mm low, and you only compress them .19mm, you may end up with marginal or no contact on the low pins.
I strongly believe that motherboard manufacturers should develop similar solution, ideally perfectly adapted to their mobo.
"They have matched a far more expensive competitor on the market" I found this to be the case with thermal paste as well. Noctua NT-H2 performs identically (give or take 1% at best) to Kryonaut, but it's significantly cheaper. I usually buy medium-sized tubes that will last multiple installs because it has a strong shelf life, multiple years
I am so old school, I have thermal paste that is probably 15 years old. Took apart the laptop I had recently and replaced it with that thermal paste, It still preforms like it should, maybe even a little cooler than stock. It's almost like all thermal pastes are the same, no matter the brand.
@Wyatt O'Shea the paste in the tube I got had settled/separated a bit. I had to stir it up once it was on the CPU. I don't think it will dry out but, you may need to stir it with a toothpick as it ages.
@Stephen De Tomasi I have used it and it does a fine job. You get a LOT for very little money. The paste I got had 'separated' a little bit and I had to take a toothpick and stir it (the dab on the CPU) back into a homogeneous consistency. I have no idea what the shelf life is as I've always just stirred it up as needed. lol. Also, I've never used it on a CPU with a TDP rating higher than 95 watts. So, I have no idea if it could keep up with a 12900K. But yeah, it's good stuff for the money.
I. Prefer noctuas as well
thermal grizzly is a smaller compagnie than thermalrigh
Thermalright is a well known for great cooling. I've used the Ultra-120 Extreme back in 2006 and it was great. Kept my old Overclocked Conroe Core2duo really cool even without a fan.
As far as price difference is concerned, you also need to figure in the cost of changing out the equipment used to create such a low variance in depth from unit to unit. It's not as simple as making an adjustment every 100 units to make sure you're still within 40 microns. Sometimes entire molds need to be changed out after a certain number of uses because they lose a few nanometers of surface area after each injection. I seriously doubt the factory in China that's pumping out 10,000 units a day is switching out their equipment every 100 brackets.
I think the difference in thicknesses has no impact on the torque you need to apply on the screws. You will need to turn the screws in less on a thicker plate and in more on a thinner plate, but the amount of pressure you need to apply is the same, so the torque should be equal. It's the same as if you would use screws of different lengths.
Looking at the cost of the frame only, the difference between the two brands is large.But you won't buy it without a CPU. And as a percentage of the platform cost, the difference is *a lot* smaller.
It actually seems like there's more room for Thermal Grizzly's solution to go wrong because if I understand Mike correctly, Thermalright's solution makes physical contact with the board, so there is a hard limit on how far it can go down, so it's harder to screw it down "too far". Whereas TG's solution is relying on specific turns of the screw. Somebody not paying attention or yolo-ing it is much more likely screw it down too much. The same can be said about TR to an extent, but the "hard wall of resistance" at least gives some feedback while installing it that says "maybe you should stop screwing it in now."But the fact that one of them isn't even readily available for purchase kind of makes the choice moot given they have near identical performance. I saw a TG earlier on Amazon (from a third party seller) that was $64 (that doesn't seem to be there anymore) compared to $15 for TR. On eBay, it's $50 + $50 shipping from Germany for TG, and $11 with free shipping for TR. If my choice was $35 and $15 and they were readily available, I'd go for TG, but since the actual choice at the moment is $100 for TG and $11 for TR, that's a much harder sell.
@JC Thanks! I can see underside of my MOBO. I will try to swap wo detached the MOBO from Tower.
@Marcelo Luz if you can see the underside of your mobo where the cpu is, you should be ok.
@Marcelo Luz I didn’t remove my motherboard from the case when I installed mine. 👍
@rawednylme could please tell me if it's possible to install the piece into the tower without removing the motherboard?
@rawednylme Yeah, I was looking to put a 12600k in a living room PC so I don't need OC level cooling, but I also didn't want the bendy CPU, so that's why I was looking into it. Good to know that it's a relatively quick install.
I really appreciate your team testing this, I wanted the TG one, but it had such a long lead time and I wanted it before a finished my build, so I got the TM one.. I did follow the same TG install process and put my own witness marks on it... Glad to see it seems to had been a good choice!
Total height variance may not change the mounting pressure. It is the height from the bottom of the frame and up to the bottom of the hole where the screws are located that matters. You could even have 10 identical height frames, but have 10 different mounting pressures depending on how precise the holes are made.
In Canada, Thermal Grizzly contact frame was priced at $100 CAD (~$75 USD) and was only available from a specialty vendor that would take weeks to ship, Thermalright is available on Amazon for $20 CAD (~$15 USD) available next day shipping. Easiest decision ever after seeing the results of this reviewEdit: just installed it on my 12600K and saw a 9 C drop in max core temp when doing an XTU 30 min stress test, so consistent with GNs results. However, interestingly my core to core delta slightly increased from 16C on stock to 17C on thermalright, although on stock max temp was 93C so I think there was some 'soft' throttling going on that may have caused temps to bounce off before hitting 100 C hard throttle
The LGA 2011 v3 dual lever locking system was pretty great, so it's not like Intel had no choice but to compromise in the current way to save money when they could have adapted that for negligible extra cost.
Great to see such informative, objective, and well-researched content being put out. I can only hope some of the engineers at Intel are GN enthusiasts and can use your findings to make a case to their colleagues about the need for a better implementation of the CPU sockets in future.
thank you for waching and commentinhit up i have somthing for
Thanks for the review Steve. I bought one of the Thermalright contact frames before even hearing about the Thermal Grizzly. Good to see that the performance is even a touch better. I paid $15 for mine and figured you couldn't beat it at that price.
I have the thermalright version. At first install, I tightened it down similar to what's mentioned in this video, and my machine wouldn't POST. I had trial and error loosening each corner a quarter turn to get the machine to POST and be stable. It was very finicky and frustrating, so be patient.
@Paul Brookfield Honestly, my opinion is that the amount of tightening is motherboard mfg/pn dependent. MY build had issues with POST when there was 4 sticks of RAM with the standard intel frame right out of the box. I was able to get it to POST with only 2 sticks. Then I had issues with it running XMP on 2 sticks unless the oem intel frame was placed "just right". I tried the washer mod, and also tried changing PSU, RAM, SSD, GPU... I literally tried everything. I saw the Thermalright frame and thought I'd try it as a last ditch effort. With that background, the first time I put the Thermalright frame on was not a success. Because I was at the end of my rope and about to RMA the Motherboard (within Amazon's 30 day window), I just kept screwing with it until it finally just worked. It took me a considerable amount of time and thermal paste doing the trial and error to get the right frame pressure. Bottoming out like you said, resulted in no POST for me. Its currently running 100% stable on XMP with 4 sticks now. I have a Asus z690m prime. everything was built in June, installed the frame on in early July.
Interesting difference, in my case it was a first-time success. Well, 'first-time success' if discounting me installing the sodding CPU upside down first time because I'm a dipshit, thankfully I'd at least noticed the IHS text was upside down when it came to applying the thermal paste.I just had all screws down fully until the sudden huge increase in resistance caused by the frame fully contacting the motherboard. Very light touch used, finger-pressure only on basically a bit in an extender rather than a full driver. Anything pressure-sensitive like this nowadays, I follow the 'wheel lugnuts' methods of partially tightening each screw bit-by-bit, keeping the pressure very equal at all times until it's done; it likely makes a difference in these huge sockets.Also, if your build is slightly older than mine was (a couple of months) then perhaps your CPU itself picked up a slight bend from the stock IHS over time?
@jitHster cinnebench r23 went down from 88c ish to 81c ish. Gaming wise; DCS, GR Wildlands, Rocket League, iRacing etc I have yet to see it get over 65 to 68 so i can run quieter fan profiles. Idle is in the 26 to 30 range. 12700K, 280MM Actic Freezer and syy paste
Good that you managed getting it to post. What temp difference did you see?
Very well done, as always. Fantastic team. I usually lurk, in fact I'm running a 5950x, but this interests me, and I would GREATLY encourage everyone to purchase Thermal Grizzly. the type of company that creates good jobs.
I went with the Thermal Grizzly option since I have been dabbling with overclocking my CPU and RAM. I’m currently using the AIO Liquid Freezer II 420, although eventually I want to try to build my own custom loop for the first time. While I didn’t run any benchmarks, or write down any temps while in game, I was with the Intel ILM thermal throttling in Star Citizen, as well as seeing regular temps in the high 80’s to low 90’s. With thermal grizzly’s frame, I haven’t yet thermal throttled in game, running through any of the cities on the planet, and now most of my temps are low 80’s, to high 70’s with the one core so far only peaking to 88c.
As far as I remember, I saw news on the TR frame announced first before I saw anything on the TG frame. Looking it up, the TR frame was announced on April 21st, while the TG frame was on the April 24th (video on Derbauer's Clip-Share channel). So you cant really say one company copied another. I do prefer the simplistic look of the TG frame but I bought the TR frame due to availability and pricing. Got it on Aliexpress for $9CAD and it took less than a month to arrive. I just tighten the screws a few turns at a time in opposite corners until I felt a "stop". I didn't do anything fancy with 90° turns and such. The frame gave me a -8°C drop when testing with Cinebench R23. I use an Arctic Liquid Freezer II 420mm.Edit: I used the frame for my 12900k.
Through history, plenty of manufacturers copied Thermalright, so this wouldn't be a first.
Essentially the Thermal Grizzly price is a blatant rip off. Would have been far better if Steve had called them out over their pricing properly instead of trying to save them some face by vague comments about better tolerances for extreme overclocking.
Do you not understand higher labor costs making a product more expensive? There is a reason it costs more, just not one you care about as a consumer. Doesn't mean they're scamming you.
well if you look at it objectively the thermalright did measurably better in tests (more than 1 degree difference), costs a lot less and was subjectively easier to install. so i feel like the conclusion should go way more in favor of the thermalright.
And TR was first to market, not the other way around as Steve claims.
I used a red one on my step son's PC and you actually can see it a little from the top and bottom of a Corsair AIO liquid cooler. Sure, it's not a lot you see, but it isn't completely hidden. So it does matter to pick the color that best matches the rest of the case colors.
Can you see the "Intel 12" logo? I'm going to be running the same cooler h150i elite with a 13700k but seeing 12 gen logo would bother the petty person I am lol
Aliexpress shipping has been really good for me lately. Usually the items arrive within 3-4 weeks even though Aliexpress say over 1.5 month on the shipping estimate.
Good to know Thermalright is a viable option. I'd always prefer to support a company like Thermal Grizzly, but if it's going to continue to be unavailable in my area, I might need to go with the backup. Thanks for the solid review!
I got mine from Amazon for $19.99 CAD (and it came with thermal paste). I paired it with the NZXT Z73 360MM AIO, and actually used the paste that came with the bracket. It works amazing on the i9 12900KS, 25C when idle, 48C when in game, and the highest I got it was 78C on cinebench!
Great couple of videos and super useful. I think I'll be going with the Thermal Grizzly solution for the various marketplace, social, fitting reasons given in the video. At the end of the day $30 for a $2000+ build isn't going to break the bank.
For me it's availability that would decide it, right now the TR frame is $20AUD on Amazon, the TG is $59AUD and it's a pre order
not much considering how much the cpu cost
After testing my new i5-12400 system to make sure everything worked properly, I took it back apart and installed a frame. Definitely seeing lower temps so I can attest to its effectiveness as you proved. It really makes me want to take it back apart and lap the CPU (remembering the old AMD Duron days)...
I tested mine exactly one month before you (28 June), I pay $17 Ca on Amazon, it's not a rocket scientist thing to do, just remember how much torque you used to unscrew the original system and replicate the action knowing your working with aluminum not cast iron! Since I use a 65W 12 Gen Intel CPU, my new A4H20 WC ITX machine is amazingly quiet and cold! 😊
Once Thermalright gets these frames onto amazon at a discount price they are going to sell a bazillion of these!
I ordered mine off Ebay for $12 and they had free 4 day shipping lol. The one on amazon at the time had 2 week prime shipping and was like $17 I think lol I wasn't beat. This thing is legit though. Been super happy with it. Temps dropped an average of 6C on my 12700K. Zero issues installing.
GN's ever expanding content coverage and quality keeps getting better and better. Quality over quantity. Keep it up GN team.
For anyone running to grab one of these - the $4.35 one almost ends up coming out to $9.00 after shipping over doubles the price, you're probably better off grabbing one of the ones with free shipping for $7.**
I got one of these and installed it last weekend. It's working incredibly well on my 12900k. Hopefully Intel will fix the default frame, so these aren't needed in the long term for getting good thermals.
Great to see third-party solutions for cheapskate companies. I do wonder why there is not a proper stiff backplate solution. Anyone with a slight insight in mechanical clamping knows, that to put even pressure on a floppy surface, you need an equally stiff frame on both sides.
For future content like this (even not more ILM replacements), it would be great to get torque figures from installs. Even if it’s just a ‘here’s what we got, ymmv’ kind of thing having a N*M number would be great. Good work as always!
with the thermal right solution torque does not matter at all. As you just screw it down until the parts touch. tightening more doesn't significantly change the pressure on the CPU as any extra load is just between motherboard the the thermal right frame.Thermal rights solution depends on the tolerances of the motherboard, socket and CPU. I am not sure if they did tests with different motherboards. The more I am thinking about it the more I like the thermal right solution. It has some significant benefits. Depending on the manufacturing and that shouldn't be hard to get right it will orient the CPU parallel to the motherboard. This helps with the cooler as the cooler is also screwed parallel to the motherboard. The Thermal Grizly solution only adjusts for the same torque on all sides but that might lead to the planes of the motherboard and the CPU top not being parallel even if perfectly tightening the screws. And its very easy to not tighten them perfectly.
Coming from an optics background, I sympathize with this. But the issue is, such small screws have very low torque values, on the order of 0.8 nm maximum torque or less. And based upon the procedure they give here and in the other video, they are actually under-torquing the bolts to meet the cpu's pressure requirement. At these low Nm values, torque measurements just aren't useful: conventional tools can't reliably make such small measurements, and the unconventional ones that can are expensive and not really available to end users.For small hardware like this, the best ways of assembling is clocking the amount of turn via reference marks, like they do in the Thermal Grizzly video, or simply turning it until the frame makes contact with the motherboard, like they do in this video.
@tep1003 Yes, that's why it would be useful to the end-user that *the reviewer* provides that which the manufacturer decided you weren't worthy of being provided.
I am assuming there were no torque specs given, since it was mentioned that the instructions were pictures.
This is genuinely very helpful. Currently running a 12700k with an NH-D15 and the thermals are decent, but I'd like to have a little more headroom and consistency between core temperatures. The thermal paste spread indicates the same uneven pressure you've observed. Gonna give the thermalright unit a go. It's sure as hell a lot cheaper than using a beefy aio.
I will NEVER get tired of the cuts to "Thanks Steve". Gold as always, including the solid reporting and testing!
I’ve been using the thermalright frame for a while on my 12700F bclk oc build. Same numbers as here, great bargain, no more thermal throttling even with 240aio at max tdp.
Always had good luck with Thermalright coolers. I had the 90mm and 120cm downdraft coolers and the Ultra 120. All very capable coolers. So it's good to see that they seem to be carrying on the quality I'd expect from them. By the way who punched Mike?
Thanks for this review. I had to buy this frame because the TG frame was out of stock on most sites, as of 08.12.2022, and the ones on Amazon are selling at $60.00. The Thermalright’s are around $15.00, so big difference here.
Love what you guys are doing. Keep up the great work!
When mounting the frame you should hold it down with your hand while snugging it and then torque in a kitty corner.Without holding it down it is too easy to get canted even if your only doing a few threads at a time as you instructed.
btw heads up guys, the included thermal paste TF7 was far more superior than NT-H1 in my own testing, I don't have controlled enviornment but it is noticeable better ---- with NT-H1 under CPU stress it will heat up my RAM a bit more through the PCB copper trace enough to cause instability, and it won't with TF7.
so Intel's next gen CPUs will be roughly 8° C cooler, good to know ahead of the marketing slides
@Victor Gil You've fallen for marketing hook line and sinker if you're not a shill. "next generation will totally" doesn't mean shit.What's wrong with "efficiency" cores is that they're stupid marketing doublespeak to pad advertised numbers. A cpu doesn't need "efficency" cores if it's actually efficient,all the cores should be efficient. All the cores aren't efficient because they just cranked up the power draw when they got caught with their pants down and walked all over by chiplet designs. They couldn't do speed and efficiency at the same time, actually adding more cores would be an inferno disaster, but they needed to advertise high core counts in their own chiplet designs like the competition. So they just added crappier cores to pretend it's just as good.
@ffwast What’s wrong with the efficiency cores? Raptor Lake will improve multi-core performance by like 40%+. That’s a massive year over year increase.
@Libra SD If you really want it to reflect Intel's philosophy it should be called an i5+, with another '+' added each generation.
@Igotspawned i think no, but the i5 12600k isn't an i5, is like a hypothetical "i6" because it has 4 extra E core
@ffwast only yesterday I noticed that all the i5-12x00 have 6 P-cores and only the i5-12600K and the KF version have the 4 E-cores ... practically they have inserted in the i5 series a processor that has nothing to do with all the other i5s, the K series has always had the only advantage of the oc but now they have made it a real intermediate step between i5 and i7 ... they could have called it i5 12600 Ti, it would have better reflected the company philosophy
I have one of these on my ITX build. I bought it specifically to have an even pressure on both sides of the cpu and in turn prevent my 12700kf from bending over time. I dont think it helps with the temps, but that could also be because im using an LGA 1155 standoff for my Corsair AIO. Once the proper LGA1700 standoffs arrive i am hoping for improvement on the temps. If not, well at least it wont bend. 31 degree Celsius on idle while maxing at 85 degrees on cinebench dont seem too bad anyway but if it can be improved, why not.
The mounting instructions are so good and well made, just as expected from your channel
Seems like the more important tolerance when trying to measure by clocking the screws with a reference isn't really the overall height of the clamp frame. Rather the distance between the recessed seat where the screw rests vs the underside of the frame where it contacts the cpu ihs. If those holes are countersunk at different levels one screw will have more free spins before it actually applies pressure. So counting 'turns' even once the threads grab essentially means squat.Not to mention the variation of thread differences. It's really just a guessing game using that method. Which is why torque values are used for critical applications. And then of course you're relying on the device used to be accurate and unvaried itself. A janky torque driver or meter is just as useless as doing it by eye or feel.
thanks for reviewing 12th gen contact frames. i'm really in need of one after 4 months of trying to get my own pc to work; i just want the cooling to work super well when all is said and done. sadly, it looks like every site in the USA is out of thermalgrizzly stock.
Mike is very quickly becoming one of my favorite GN crew to watch for. He's extremely funny and also presents information in a very relatable and clear way. Hoping to see more from him over time.
Please give us more Mike.
Mike did a beautiful, clear, logical, well flowing explanation of the process that gives all the theoretical info to people who might be uninitiated, without being too long or boring for the people who may know some of the stuff.
A small torque wrench would be a much more accurate and consistent way of installing the plate. But it would probably cost much more than the plate. But a small torque wrench is a great tool to own for all sorts of projects though.
for the price I was honestly stunned. I understand the machining processes each likely took. but a $4 item typically doesn't get anything but cosmetically accurate care.
I heard from some Intel engineers today that work on the NUC extreme element and cases. And they had some things today about a socketed CPU versus a soldered BGA mobile solution. The pads are really bad a a thermal bottleneck, as there is notable resistance by just pressing the things together. it makes up as much as 15% of a difference compared to BGA soldering.
Any chance of testing adding a ton of Thermal paste between the CPU and the bracket and also between the bracket and the cooler to see if there is a slight temperature improvement? I know it probably won't be worth it because it will be a mess, but I'm curious.
I purchased the TR frame, installed before my first boot on my 12600kf so have nothing compare to, however certain it's doing better than the stock ILM. Idle at 24 dc and gaming doesn't go above 45, so plenty of OC room. Using a recycled custom WC loop, WAY better temperatures than my 7700k delided with TIM.
Great video, and obviously the proof is in the pudding that they work the same.But if you're going to start comparing precision physical measurements between products, I think you should read up more on metrology and good measuring practices. The measurement you compared actually has no bearing on the performance of the product, and those calipers are not the correct tool for taking such a small measurement. It's a rabbit hole and maybe it's not common enough for you to invest big money into the tools needed, but I wouldn't consider your measurements accurate enough to draw the conclusion that you did about the part's tolerances.
Thermalright has some CPU coolers that I used and like it. Glad to hear they are still doing good.
The TG frame will probably keep the tolerances throughout the production life. For Thermalright, maybe the next batch will not be as good. So everyone should buy now.
This was such an honest and refreshing review that approached the subject from several different directions.
Been waiting for this. I bought the thermalright version as the TG version wasn't released yet. I'm using it on a 12900k and it seems to work great so far! I purchased it from ebay and I believe it was 15 n change. Thanks for sharing 👍!!
Theoretically speaking, one could machine or have machined a custom acrylic block with good enough tolerances that the copper cold plate doesn't cover much more than the CPU lid itself, and than have all clear acrylic that allows the frame to be visible, but it would be a massive effort.
I would really like to see a review with these and a delidded with liquid metal vs a stock cpu to see how much it helps a delidded cpu.
“Thanks Steve” for making this video. I own a few of these thermalright frames and they work really well which you made very clear. Versus the thermal grizzly one at $35 is a terrible dollar per degree saving value.
@RichardFliehr Well said. And yes that is the simple answer to the overall topic. I was just responding to someone who said it's "greed" and all this other nonsense about them being manufactured in Germany vs Asia bla bla. That's where my comparisons came from.
@thegamefaqWho owns the copyright? Racist much?You obviously never owned a current bmw, it's a nice car for sure but unreliable.
@Pabz No it is the point because you said the BMW was of better quality, how can the Honda not be of better quality when its much more reliable. And I totally got what you were saying but it was flawed just on the fact you threw 'quality' in there, for these CPU frames there is no better quality because one was made in Germany and the other Taiwan, they are both quality frames, the TR actually performing marginally better for 1/5th the price. The simple answers is the TG frame is a ripoff when compared to the TR frame.
@rci2990 You're kidding me... I'm blown away at how this flew over multiple peoples heads. Such a basic common sense comparison and yall are looking way to deep into this. I could just copy and paste the comment above and reply it to you as well but you can go ahead and read what I replied to the other guy.^^^ Same goes for you.
@RichardFliehr Dude that's not the point lmao. You're looking way to deep into this. Obviously everyone knows a honda is more reliable of a vehicle. There's obviously more variance to a car then a fricken LGA 1700 contact frame. That's not the point the expression completely flew over your head. You're paying extra for the BMW because it has more features vs a Honda that elivate the driving experience but at the end of the day they both cars drive from A to B. Same thing with the Thermal Grizzly frame you're paying more for tighter tolerances, anodized aluminum, nice edges, etc. but at the end of the day the day both the TG and TR frame both give you the same results AKA driving from A to B. The comparison flew completely over your head obviously there's a lot more reasons one would choose a car. It's the base point I was making which stands. You can make this comparison with anything. A 5 dollar T shirt vs a 30 dollar t shirt. The 30 dollar T shirt is made of better material and quality but at the end of the day both shirts cover your body. They both give you the same end result. Get it now? I only used the BMW/ Honda example because one is basically from Germany and one is from Asia.
Great stuff! Good news for Intel owners. Wonder if there will be an equivalent doohickey for AMD CPUs?
Very informative review! Besides the thermal issues, how serious and bad is the motherboard bending? Is that really something to be worried about?
That's insane that third party companies need to repair Intels mistake
The problem is you can’t find the thermal grizzly frame anywhere right now, they kinda shot themselves in the foot with inventory. Thermalright option was very easy to get. The only place I could find a thermal grizzly frame was on eBay and they wanted about $100 with shipping
Would be interesting to see how the temps are for someone using significantly less care. ( starting 360 degree turns instead of 90)
Since the Thermalright ILM makes contact with the motherboard, I don't think it should make any difference whether you do 90 degree turns or 360 degree turns as long as each screw is at about the same torque at the end of your installment process. It's going to tighten down as far as it can go at all corners anyways. Of course if you tighten down one screw completely before tightening the others at all, something bad or weird could happen. I think the TR solution is more fool-proof than the TG one since you can't really over-tighten it unless you screw it down so tight you strip the threads or something like that..
My first installment I did a yolo installment where I stopped when I felt it was having to use significantly more force, it's very noticeable...Got a small decrease of a few temps. My 2nd installment I decided to all the way but without using any excessive force and see what happens and it didn't want to boot lulz.Now I'm using the 1mm washer method and recorded all my core temps with specific voltages over a 1 hour Cinebench run. I plan on trying the Thermalright bracket again for a more accurate comparison now that I wrote down all the core temps.Edit: For the heck of it I think I might also try to install the bracket on top of the 1mm nylon washers as well lol
Yes, I would like to see that too. Lets say you have no idea how to install one of these and the manual (in the case of the Thermalright one) doesn't offer you proper information or you're just one of those people who thinks "Hey how hard can it be to mount this, it's just 4 screws" it would be interesting to see how much difference a proper install and a "noob" install would make.
Thermal Grizzly might still be cheaper if you live in the EU, depending on how high your import customs fee is. I don't know, I haven't checked the prices. But whenever I buy anything outside of the EU I get an obscenely high import customs fee every time. Doesn't matter how small or cheap the item is. Thanks for the video.