The ability of a material to conduct heat increases exponentially with distance. A pad makes sense when you are closing an air gap, because 3mm of pad is a whole lot better than 3mm of air, but when you put that 3mm of pad in between two metal surfaces, that 3mm pad is infinitely wose than a tiny thickness of thermal paste. 3mm of pad is also going to be a whole lot worse than the microscopic valleys in the two metal surfaces of air. Also, vrms reach high temperatures, but they don't output that much heat compared to a cpu. That overclocked 8700k is going to be well over its rated 95w of heat at stock settings, a competent vrm on a high end board would be outputting closer to 40w (Gamer's Nexus often releases videos with buildzoid with vrm analysis where the heat output is calculated). That 40w is also going to be spread out over caps, chokes, and mosfets, so the heat is spread over a much, much larger surface area (2x the heat spreader easily on many of the overbuilt high end ones). The cpu is putting out more than twice as much heat in half the area, so it is obviously much harder to cool.)
This is correct! In this case distance is the key. The thermal resistance of a system (heat source + paste/pad + cooler) is exponentially proportional to the distance that the heat need to “travel”.
Tomorrow Im going to be testing the pads with different coolers on a r5 1600 and an 8700k for work. I might post a google doc somewhere for it since its public information.
@@modmen. Ironically, if NordVPN kept logs, they would have been able to find, isolate, and determine the scope of the exploit both faster and more completely.
Thermal pads are great at dissipating smaller amounts of heat, such as that from VRM's. They do get hot, but it happens slowly. You could not keep a computer running without CPU cooling, but without the VRM's cooled, the computer would stay on for quite a while. While thermal pads aren't as conductive as thermal paste, they do better than just plain air, so they are great for filling space, to provide some sort of contact with a cooler if they wouldn't already, as well as I would imagine, decrease, or better disperse the force exerted on the VRM's as to not cause damage. In short, traditional thermal pads do work, but they don't work as fast, so only use them on slower heating items.
I bet you used some Ram Cooling Pads with like 1-2W / mk. Like you said they just cant transfer the Heat fast enough. Maybe repeat the Test with something like the Arctic Cooling Thermal Pad (6w/mk), or one of the the Alphacool Eisschicht thermal pads (11-17W/mK).
since you wanna know when to use thermal compound vs. pads, I'll tell you! thermal compound/grease is used for components that require quick heat transfer and experience rapid changes in temperature. a CPU can jump 30C in an instant, and that energy needs to move FAST. thermal compound spreads thin enough that there's barely any of it between the IHS and cooler's surfaces, and mostly fills in the gaps to keep air from insulating the heat. thermal pads are used for components which are slow to heat, such as VRMs -- it takes a long time [relative to CPU temp changes] for the VRM of your GPU to get hot, and the pads are able to transfer the heat over that extended period. pads act like insulators [like in your CPU test] when there's a BIG dump of heat all at once because they need time to move the heat. thermal PASTE is the same as compound/grease, but it's an adhesive [AKA glue], and would be used for permanently applying a heatsink to a component, such as, for example, a tiny heatsink on a Raspberry Pi, which will probably never need to be replaced or removed in most cases.
What if I use kryonaut as the thermal paste between cpu heatsink contact and 6mW/k thermal pad on the heat pipes which go near or touch the back of my laptop?
@@rohitbaviskar7748 you actually run into thermal resistance due to air trapped between them. Thermal paste or pads push out air and ensure maximum surface contact. Even if you have to conductors that are as close to perfectly flat as we could manage, there may be a tiny air gap (you can never have a truly perfectly flat surface) or they might not be aligned to lay perfectly together
@@tejaspadhye I wouldn't use pads *and* paste, that might actually reduce thermal conductivity since you're adding material. But having a pad between a section of heat pipe and the case might help transfer heat a little... Can't hurt to try
@@DoctorX17 I'm not using them together between heatsink and cpu die, but where the heatpipe comes in contact with the case, the case has a copper sheet typo thing attached just above the heatpipe. I'm hesistant and asking you this cuz the pad might melt and leave behind a mess and my laptop's still in warranty
Thermal pads are used on VRMs, memory chips and such for a couple reasons: 1) those parts are less prone to heat damage than a CPU itself 2) more importantly: they need to dissipate less power for unit of surface that a CPU so they do not need the highest thermal conductivity 3) thermal pads are solid, while thermal paste is, well, a paste. It's much easier to cut and apply a pad rather than putting say 8 spots of thermal paste on an 8-phase VRM, so they are much more easily applicable than paste and, as said, they do not need maximum performance. Another point in favour of pads is that they adapt to different thicknesses, like an 8 phase VRM, without problems of not making contact (if they put paste on a slightly lower chip it would not make a good contact while a pad can still touch)
BackInBlack_4896 Good arguments, I would just add two things : - good paste application is much more difficult than placing a pad - the cost might be part of the good answer
Honestly, I believe I speak for more than just myself when I say that if you tried to de-lid the 8700K (because of it intense heat levels) and tried to cool it, and also make it a "How to", people would be very interested! Like I myself want to build my system with an 8700K, and I'm using a custom cooler from EK, and different Thermal paste, but, let's be honest, it's an area heater.
Jayz, Please make a water cooling system where one tube is connected to a water tap and then the outflowing tube from block is flowing into the sink so there is no loop and water is never heating up! I've never seen it done and haven't found anyone doing it, would be cool to see cold tap water cool a cpu!
There might be a reason its not done. And the reason might be that theres so many that dont have acces to clean water, so it would be quiet hypocritical to do. But thats just my opinion.
I really get the idea, but honestly it is terrible. For starters it will get costly quick and is ridiculously wastefull. Tapwater isnt free and running a faucet whenever your pc is on will get expensive.
Gold isn't all that good at thermal conductivity and is also electrically conductive. The only advantage gold would have is that it wouldn't corrode. Gold leaf would also be very difficult to apply without getting it everywhere.
@@LordMoebius Gold leaf sticks in place pretty well, in my experience. It's certainly less messy than paste, which seems to be the main issue with thermal paste. It's also easy to apply. Just put a little pressure on your sizing, and the gold leaf will adhere just fine. Not sure why you think Gold isn't a good thermal conductor, it's one of the best metals for thermal conductivity. In fact, aside from some odd alloys, only Silver and Copper are better thermal conductors, but while both can also be used in a leaf application, they are thicker and will corrode, unlike Gold. The only thing better than those three is Diamond, which obviously wouldn't be applicable here without some serious technological development.
@@berserkasaurusrex4233 gold leaf has a tendency to flake during the application process especially when applying it on an intricate or small surface and you also need to use adhesive to hold it in place. So pretty sure it wouldn't be the best for thermal conductivity. And yea silver and copper are better conductors at a cheaper price point compared to gold. The only advantage you'd get is that gold wouldn't corrode but it's sorta irrelevant in this context so you're just paying for a material that performs worse than cheaper materials.
Thermal pads do make sense in GPU blocks if you consider them as spacers. It would be very costly I imagine to design a cooling block that is correctly spaced for all the chips on the card and it is much easier to use an X thickness pad to achieve contact. Another problem is that some pastes can short the chip's legs if it runs down (most are so small that it is difficult to push the correct tiny amount needed from the paste tube and even spreafing can make a mess of things). While some memory chips on GPUs tend to run rather hot, those temps are nowhere near enough the GPU's temp without a heatsink (they're probably in the region of 50-60C uncooled which tbh is not a high temp as far as chips go.
Late to the party! In engineering we choose thermal pads when you need to close gaps where you do not have control of component tolerances or a wide gap in component height that needs cooling. It’s a lot better than nothing and vary greatly in “thermal resistance”
I rebuilt my daughter's EVGA 1080 that had the thermal problems using the EVGA thermal pad kit, but I used Thermal Grizzly Kryonaut instead of their TIM. Dropped temps by 8 degrees C. Thermal pads only work on lower wattage heat sources as the pads have a lower amount of heat transfer rate. This is what accounts for the variance in various brands of TIMs as well; they all have different heat transfer rates. Kryonaut has one of the highest transfer rates out of the other TIMs. You have to go to a liquid metal for the better heat transfer rates. When looking at any cooling solution look at the amount of heat it will transfer given the application. Also, if you lap the surfaces of the CPU lid and the cooler plate to make them smoother (and make sure they are absolutely flat (I've seen some cooler plates that aren't)) , this will also increase heat transfer rates as there is more metal to metal contact.
Also because they're electrically insulating. You get some thermal transfer without risking shorts when in close proximity to 100+ Amps. Thermal pads are rated in terms of watts/area and they're a fraction as efficient as thermal paste. If you had a 3ft square heatspreader, then a big thermal pad would be able to transfer all that heat. A pad that small is probably only rated to transfer a few watts of heat, not the 300 watts or so from an overclocked 6 core.
Allison Pell 100Amps? Are u sure what u saying? Isnt that "100 A"? Maybe im missing something but... U got like i limit of 12 A in the wall, or it will cut out the power...
amps times volts is watts. if you are pulling 100 amps at 1 volt, then you're drawing 100 watts from your power supply. your power supply manages how to pull the volts and amps so that you're not drawing too much power. if you have two electrodes and there's a hundred volt potential between them, you aren't drawing power because there aren't any amps (no electricity is flowing). if you link the two electrodes, current will pass, meaning amps will start existing. the voltage will drop, depending on the resistance of your wire. if your wire has a high resistance, you won't draw as much as if it didn't resist at all (a short circuit). in a short circuit, you have no electrical potential (there are no longer two potentials because there is no two distinct electrodes) but you have a high current, meaning a shit ton of electrons are passing through your wire. this heats it up (pretty much because of friction) and probably melts it. I'm dubious that there are 100 amps going through a PCB because if you want to have more amps going through a wire, you need a fatter wire. welding wires use twelve amps and they can easily be a centimeter in diameter (welding is short-circuiting metal to heat it up to the point where it melts).
VRMs may pull out about 60W of heat, but the heat is spread across a way bigger area than on a CPU/GPU. oh, and VRMs aint flat, there are solderings etc arount it which also get contacted. and since most thermal pads are non-conductive, you can't short out a FET and get 12V into the CPU. under LN2 you see way more heat output, but since the CPU/GPU is sub-0, most parts of the Board around it are frozen however so heatsinks usually just heat up the Board, cuz they are contacting something which is not at - 50°C or smth.
Hi Thermal pad work excellent where the source like Vram, GPU, Ram or chip heat in these chips heat gradually increase so the transfer pads work well. But like CPU where heat generate suddenly Pads fails. So it's better we use Thermal paste that evenly spread heat and removes spikes of heat that can damage CPU cores. Its better to use 6W/mk to 13W/mk for CPU. But check does the heatsink or water block sits properly on CPU surface. Note if not fitted properly on CPU NO Thermal can pass heat to the heatsink. Don't apply too much that will also ineffective. Apply the proper layer of CPU . It's only to fill microscopic holes of CPU surface and heatsink..
4:46 I took apart my two PS2 Slims and found plastic on the thermal pads. Taking them apart is easy and I recommend any other PS2 owners check their system!
@JayzTwoCents That type of thermal pad isn't for what you think. its meant to go on things like ram to spread heat through the chips in the case of vrm and normal ram. This is not for cooling but when the silicon heats and cools when in use so a hot spot could cause that part to expand while the rest is still stiff causing the sodder and silicon to warp damaging connections in sodder. This is the same for the case with heat spreaders on normal ram sticks.
I've used ceramic thermal paste ( some generic stuff for audio components , and Arctic white) to adhere some old zalman ram sinks onto a 6850 with a zalman vf1000 led heatsink "back in the day" ran it like that for a few years and they never came off/ fell off untill I changed the paste....just throwing it out there, I think it actually did better then pads
Take the leaf blower idea, but add in an intercooler from a junkyard turbocharged engine and get a second leaf blower to blow air across it and chill the charged air.
Hey Jay, I’ve got a dumb idea/suggestion for a video. It more of a curious experiment. Here it is: Delidding a CPU to replace the TIM between the IHS and the die with a graphite thermal pad. 🤷🏽♂️
The pads that everyone has been using are made out of graphite and are way better than the generic ram pads you were using. Kinda no comparison. Nice vid thought.
Themo pads are used where there are large gaps between heat spreaders and the module. VRMs have typically large gaps and there is higher heat transfer through a thermo pad than through the air. In general, conduction transfers more energy efficiently than convection. (solid vs air) Thermo pads are also used where you cannot feasibly put paste. It would be impractical to cover VRMs in paste enough where it actually contacts the heat spreader. Also, the paste would ooze out of the card eventually due to gravity.
I wouldn’t have known about the pads if you didn’t make the video. And can’t wait to see the cpu pads in action. It’s doesn’t matter who made the video first without you making the video then where’s my excitement in my life. Thanks jay.
The thermal pad predominantly helps to resolve alignment issues. Like if the plane of the cpu isn’t parallel to the cooler, only parts of the cooler will touch the cpu and not a lot of heat can transfer. How the pad does this, is that it is squishy and metal isn’t with the pressures involved. To clear up, I wasn’t talking about a specific material of thermal pad, but rather just the general idea
If not stated previously the surfaces are micron smooth -must use tissue paper not paper towel to clean surfaces and alcohol to clean out micro-pores first. when spreading paste it should be thin, thin, thin and worked back and forward several times so it "gets in" to those pores.
Thermal pads are cleaner and allow for a greater space to be crossed, if you look at some pads on cards they are thicker or thinner for optimal contact with the gpu/heatsink, less metal machining and cleaner. They are also used in testing applications, as in hey lets see if this cpu boots as fast as possible then onto the next. OH AND they work well in situations where something is warmer than expected due to bad silicon or design, like the recent fiasco with evga? sending thermal pads to 1080 buyers.
Makes sense, and you kind of answered your own question. The thermal paste is used to cover micro cracks and displace the air, air is a poor heat transfer medium. The thermal pad just creates pockets of air underneath and in the cracks plus introduces an additional thick layer to the heat sink that creates additional heat transfer resistance.
That crunch in the toothpaste isn't the "Whiting Crystals". Toothpaste is an abrasive. It's microparticles in tasty soap. Basically silicon and soap. It cleans your teeth and whitens your teeth by literally scrubbing and scraping them. The whitening crystals (lol) are usually just things like peroxides and light bleaches in crystalline form.
I have a Turris Omnia router, which has a small aluminum cube as a cooler on top of Armada CPU, I changed the thermal pad on the CPU to a thinner one and added another on top of the cube, so it started transfering heat to the chassis (which is metal). Long story short - average temperature dropped by 30 degree. So it works in certain situations.
TIM is meant to fill gaps, it has worse conductivity than metal, but better than air, so you want it to be as thin as possible. VRM components have much higher variability in height than a heat spreader would, so much that without a thermal pad, they probably don't even contact the heatsink. As for temporary solutions, I'd be very curious to see how a drop of water(distilled) performs. Obviously you'd need a flat surface to mount the heatsink, but once mounted, surface tension should keep it in place long enough to do some testing.
To answer your question it has to do with heat capacity vs thermal conductivity. The pads have a low capacity but ok conductivity. Dumping too many joules into the pad, more than it was designed to handle, will saturate it and create an insulating effect. It's like the gauge of copper wire, kind of. Its conductivity isn't changed by the thickness of wire, but can transfer more heat the thicker it is. The pads and paste have different physical characteristics and properties, but the concepts are the same.
Random, crazy idea to try: without any fans attached, submerge the radiator part of a water cooling loop inside a bucket of ice water (variations could include dry ice and other crazy stuff) and see how well it cools (or goes gloriously wrong). You could also try potentially lowering the temperature of the water using salt (like ice cream salt for those homemade ice cream makers). Give it the ol' ice bucket challenge this Summer!
The thing about thermal pads is: with a cpu they get saturated with heat and they can't transfer anymore heat, but on something like vrm or ram they have a lot more area to transfer the very little heat (in comparison to a cpu) and don't get saturated. It's like spilling water and cleaning up with a sponge and then seeing a flooded house and trying to clean it up with the same sponge
Hey, you should mention that thermal pads are not meant for high psi CPU mounts. They're bulk gap fillers. Under a high psi CPU mount, the copper IHS and the copper cold plate is being elastically deformed by pressure to reduce the need for gap fills. With thermal paste, the remaining patches of poor contact is filled to prevent an air gap. If you use a thermal pad, it prevents all metal-metal contact and increases thermal resistance over just bare copper.
The reason and the application where those pads are used, is for when you have a single heat spreader or cooling block for multiple chips because no matter how exact the mounting of the chips is, there will still be differences in height between them. If you put a flat cooling block on top, it will make contact with some, but have significant gap (well, significant in terms of heat transfer. Like 0.01mm) to others. So while you would get better cooling on some chips, you would get virtually no cooling on other chips. By using these pads that are a bit thicker, you can even that out, and since they are a tiny bit squishy, the ones on the thinner chips will just have a little bit less pressure on it which doesn't affect cooling. The reason it becomes worse than nothing for the CPU, is because compared to direct touch, it actually is worse since no paste or pad, while you will have air gaps, you will also have direct touch to SOME degree and these pads, while superior to air, are far FAR worse than directly touching, even if that direct touch is over a much smaller area in total. And it also works as an electrical isolator (be warned, the CPU version you talked about at the end, is different and IS CONDUCTIVE) but that isolation is more of an added bonus, not the reason they're used.
How about extreme polishing on both the water block and the heat spreader on the CPU? Smoother, super flat surface should help transfer the heat more efficiently without the need for paste. DO IT! Please :)
do you have millions to do it ? No ? That's why we can't. When you polish something, you're just making thinner scratch/dents in the material. For a PERFECTLY flat surface, you need an ultra-precise machine, with an ultra-expensive cutting tool. It can easily cost hundreds of thousands of bucks.
Thermal Pads are actually transferring a bit of heat away from VRMs/RAM/Coil and Stuff.. The special thing is, they can bridge the gap inbetween the Cooling plate and the Parts on the PCBs. We use Thermal Pads at work also. The Parts on PCBs are not always the same height after soldering, and to catch all the tolerances, thermal pads are used. When PCBs for Spacecrafts are built, the cooling plate is actually fabricated after the build of each individual PCB. The PCB is measured in 3D to micrometers, yep, really! and then the Cooling Plates are milled to these measured Values. This makes the gaps between the Parts on the PCBs and the cooling Plate really small, so we can use very thin Thermal pads. The Thinner the Pad, the better the heat is transferred. There are also HUGE differences between some Thermal pads, the are som cheap really crappy ones, and there are some good ones. The ones we use for Spacecrafts are awesome but expensive as HELL.
I built a dope gameing / content creating pc and accidently forgot the "WARNING!!! PEAL OFF BEFORE USE!!!" sticker on on the cpu block of the masterliquid 240. I used it for a half year with an overclocked 7700K and complaind a lot because the temps i got with a prety good AIO. Than i decided to delid the cpu and when the liquid metal arived and i went to take apart the whole pc i recognised the f*cking sticker i just replaced the thermalpaste and the idle temps moved from the high 40s to the low 20s.......
Thermal pad makes sense in a situation where you have lot of different heat sources, in that situation efficiency doesnt matter as much as being able to transfer heat to any surface that can be cooled
If you dont want to use Thermal Paste on your CPU, I recommend Graphite Pads. Innovation Cooling (The maker of IC Diamond) sells them at Amazon. They are quite good and they are reuseable.
dear Jay. if you would have consulted the technical properties of these thermal pads, you would have noticed the very little thermal conductivity of the compound. the reason it is used is because it is the best solution for a heat conductive material that is elastic. because the heat sink isn't as firmly connected to the ram/vram and it needs an elastic connection. the only possible better solution to thermal paste or liquid metal is to weld the cooler onto the surface of the processor. that is the only way i know without a thermal barrier between two materials. thank you very much for all your testing and even dumb ideas. they are entertaining and sometimes shone light onto a matter from a very different angle. keep it up.
Power MOSFETs generally only dissipate a couple of watts. They get hot because they are often not actively cooled, and so generally have a high thermal resistance to the air (on the order of 50°C/W, meaning each Watt of power output raises the steady-state temperature by 50°C). Thermal interface materials have thermal conductivities expressed in Watts per meter-Kelvin. The weird unit comes from the fact that thermal conductance (the inverse of thermal resistance) is proportional to the area and inversely proportional to the thickness. A cheap 1W/m.K 0.5mm x 10mm x 10mm thermal pad will have a thermal resistance of about 5°C/W, meaning if a single MOSFET dissipated 5W, the difference in temperature between the MOSFET and the heatsink would reach about 25°C. Meanwhile, a CPU is capable of dissipating 100W or more of heat. A 0.5mm x 25.4mm x 25.4mm 1W/m.K thermal pad would result in a steady-state temperature difference of about 80°C between CPU and heatsink at 100W. The best silicone thermal pads have thermal conductivities of about 5W/m.K, resulting in a steady-state temperature difference of about 16°C between CPU and heatsink (less if compressed) at 100W. The IC Graphite thermal pad has a thermal conductivity of about 35W/m.K, which assuming the pad is about 0.1mm thick 30mm x 30mm pad would give a thermal resistance of about 0.003°C/W, resulting in a temperature difference between CPU and heatsink of about 0.3°C at 100W.
heat energy transfer relies on the temperature gradient and thermal conductivity. If you can accept a high temperature gradient, you can have the same heat energy transfer with a bad thermal conductor. But if you a re limited to 100°C on the chip, you need a better conductor for better performance. Pads are in the end just cheaper and are able to hold small cooler in place without screws and such. You can use them if you want to oc you raspberry pi or smt like that.
To answer your question jay i believe thermal pads typically make sense in areas where your heart source and sink are farther away since pads can be made thicker than paste and transfer better than air
Thanks for the video. Regarding your question about the idea of thermal pad and how it works, I should say that when heat (or any other things like sound, light, motion, electromagnetic waves, etc) conducted through a medium everything depends on the conductance capacity of that medium and as long as the medium is homogeneous (no abrupt change in the material) everything is fine. However, when heat reach to an abrupt change like passing from one medium to another one the challenge happens. Exactly in the junction you have an abrupt homogeneity which means that the conductance capacity changes abruptly because of two reasons: first is that the junction surface is not microscopically perfectly flat and there will be thin layer of air there (air is a heat isolator) the second reason is the difference between heat conductance capacity of the two medium (the more difference the worst the case). To envision it assume that you are on the roof of a building and want to go to the nearby building if the height of two buildings are the same (conductance capacity) then you can easily go on the roof of the nearby building but if the heights are different you cannot go and you need a ladder (conductance matching) which helps you to smoothly pass the heterogeneous junction. You can model the microscopic air-gap with the space between two buildings. This is what thermal paste or pad plays in the junction, providing a ladder for the heat to easily pass the junction for both removing the microscopic air-gap and conductance difference. The specification that the thermal conductance should have is that it should have a conductance capacity that matches to both mediums.
No, though you could if you want for personal preferences. Thermal solutions carry eg. 1w per square cm surface, at 0.1mm solution thickness, at a temp differential of 10C, or 2W at 20C, linearly. Depending on the surface available for transferring X watts of heat at a predetermined differential, you can use different compounds or methods. A CPU puts out around 10-15 watts of heat per square cm, so you want paste that can transfer eg. 10W/cm2/10C (150W CPU / (10W x 10cm2 surface) x 10C = 15C differential); VRAM and FETs only put out 0.5 - 2W per cm2, so in this case you can use potentially cheaper, easier to handle, less messy and reusable pads that can transfer eg. 1W/cm2/10C (2W FET / 1W x 10C = 20C differential). If conductivity requirements aren't particularly demanding, whether you want the potential benefits from pads or not, the choice is yours.
I've had great results with graphite thermal pads less than .02 inches thick. Thickness and material composition counts. Silicone pads are 1-2 W/mK rated.
Use a thick layer of thermal paste on memory chips and VRM instead of a thermal pad, got it! Btw, you have to try homemade peach marmalade, it's the best household item to use as a thermal paste substitute.
VRMs output much less heat and diffuse it in a very different way. You don't even need VRM heatsinks even with an overclocked 7980XE as long as you have good airflow over the MOSFETs. You obviously can't cool a CPU die with airflow alone. Thermal pads work, just much less effective than thermal paste, and they cannot conduct heat very quickly, hence the naked setup outperforming the thermal pad. There is a reusable thermal pad for CPUs now that is actually a revolutionary product as it's effectively solid thermal paste. I would have liked to see that tested in this video vs. the stuff that you shouldn't use, but I'm weird in that I like useful content.
Have you considered pumping chilled air into a case rather than using liquid cooling? It's a thought that some may say runs the risk of getting condensation on your hardware, but unless you pump chilled air in after you power down the system or you make the hardware colder than room temp (something that would be crazy to do anyway) you should have no trouble with condensation. At least the science of it all says that the hardware has to be colder than the air it's in to cause such a problem. I've considering using a solid state cooler to chill the air that is being drawn into my laptop as a means to help it run cooler, and that's where my question comes from.
Thermal pads work when the surface area is large when compared to the amount of heat being generated. On a GPU and RAM the surface area is large and as a result the insulating property has less effect. But on a CPU the heat is concentrated in a small surface and therefore the insulating properties prevents a larger portion of heat from transferring. The pad essentially creates a more even transfer of heat on the microscopic level but due to its insulating effects it also creates a heat transfer bottleneck. The bottleneck can be lessened by a greater surface area. So basically, thermal pads work fine when the demand for heat transfer compared to surface area of the transfer surfaces is low.
One comment suggested Gold foil, made me look at some charts and it seems gold and aluminum performs roughly the same. Both are easily available , So please JayzTwoCents , test it. (aluminum foil & gold leaf) Another thing you could test is the heat transfer goop, you can buy for engines. Its some kind of copper stuff, and I remember seeing it some time ago, wondering if that would work. ( off to the hardware store we go! )
the only thing I can think of for why the pads are used for VRM and such is because the surface area is smaller on those, so the pads help add more material for the heat to be absorbed into rather than help dissipate the heat into something else.
what i know of thermal pads on lesser components is that they are made to handle a higher temp. so they need less thermal dissipation but without it they still get to hot. the thermal conductivity is better at higher temperatures differences. those components can work fine at 80-90 degrees. the VRMs dont put out as much energy they are just designed to handle more heat.
Reasons for Thermal Pads: 1. If there is a moving connection between parts, which thermal pads can buffer. 2. If you have too get thermal connection without electrical conductivity. 3. If there is no airflow on very uneven surfaces. Greetings from Germany ;-)
Pads are intended for use where you need to transfer heat with less pressure applied, and an area where paste wouldn't be thick enough . As in, the gap from the component to the sink is too big. I remember back in the day when a lot of OEM's used pads on CPU's, urk.
Thermal pads are actually useful when you have a gap to bridge. it can be because the component surface is away from the cooler and you can not get the 2 surfaces to touch for some reason. Although insulating, they insulate far less than the air that would be there without pads. Another interesting application is when you get a seriously uneven surface where the thermal paste would not fill the gaps reliably enough. To be noted though is that thermal paste is actually the same... it insulates but just far less than the air that is there without paste. We see less of an impact as, you said it right, the layer is way thinner than with a pad.
How about a "radiator free" water loop with like a 25 gallon Rubbermaid tub to see how Long it will take for the water to exceed it's cooling capacity. Also use temperature fittings to monitor the water temp. It's something I've been wanting to try, but I currently lack the resources at the moment.
You should try gold leaf. It has excellent thermal properties and it is pretty soft, especially 24ct gold. It would take a little experimenting to get the right thickness, and to apply it with zero creases, but it should work well. Ironically, even 24ct gold is cheaper than many of the TIMs out there (gives you a little perspective on how expensive some of these compounds are).
I got the graphite pad yesterday. Ordered after watching linus's review. It works well best used with a Water cooler. The graphite pad tend to keep the temp of each core steadier then thermal paste.
I'd like to see metal hard lines, or a scrap/ car themed PC build. I think it'd be cool to see you doing odd stuff with pieces that don't belong in a computer; gluing sheet metal blades on to fan hubs, custom ducting, using a giant car radiator and e-fan mounted on the side of a case, driving a whole car ac system to use as a chiller. I know that kind of stuff can get really complicated, but I, and I think many others on this channel would love to watch it
Necrobumping here ... It's probably already been said ... but the pads are there to prevent an airgap. The pads suck vs paste or direct contact, but they're there to fill in when you'd have an air gap.
There is normally a small dip of height in the center of the intel cpu aluminium lid. So when you put a thermal pad as is, you will have the thickness of the thermal pad PLUS the air gap that you have when no thermal interface is used. so you will have high temps. try cutting a smaller area of a pad to cover the air gap only. you can use some paste to fill the gap and a straight exacto blade to pass over and gather the excess to see the gap.
@JayzTwoCents They should probably design a thermal material that comes in like a foil form. You know, it melts when in contact with heat and turn into gel form but returns into soild form when it cools down. That would be nice.
"Don't try this!"
Of course I'm going to try this!
MWAHAHAHAHAHA
Lets all try this, is what I got from the video.
JaystwoCents Please can you make an Video about you can test this one? vs thermal paste :) IC Graphite "Thermal Pads" ?
It will surprise me if you cannot use this... more comment to follow...
Top ten pictures taken before disaster
"I hate thermal paste. It gets everywhere." -Jay Skywalker
Christopher Morgado the wild prequel meme always makes my day
Thank you for this. Very good meme
is it coarse and rough?
I was literally reading this as he said it
The quote is 'I don't like sand', Anakin 'hates' Obi-Wan. Boy, I sure hope somebody got fired for that blunder🤓
Jay... that is silicon pad... you need a graphite pad... which is what linus used.
yea, this isn't designed for heat transfer, this is designed to adhere a heatsink to....the graphite would perform worlds better (3x better W/mK)
but graphite pad only works for cpu right? it is bad for gpu right?
Don't see any reason why that would be the case, it just transfers heat.
i think i read some ppl say because it conducts electricity or something. i dont recall too much of it tho.
Yes he knows that’s what he said in the video.
The ability of a material to conduct heat increases exponentially with distance. A pad makes sense when you are closing an air gap, because 3mm of pad is a whole lot better than 3mm of air, but when you put that 3mm of pad in between two metal surfaces, that 3mm pad is infinitely wose than a tiny thickness of thermal paste. 3mm of pad is also going to be a whole lot worse than the microscopic valleys in the two metal surfaces of air.
Also, vrms reach high temperatures, but they don't output that much heat compared to a cpu. That overclocked 8700k is going to be well over its rated 95w of heat at stock settings, a competent vrm on a high end board would be outputting closer to 40w (Gamer's Nexus often releases videos with buildzoid with vrm analysis where the heat output is calculated). That 40w is also going to be spread out over caps, chokes, and mosfets, so the heat is spread over a much, much larger surface area (2x the heat spreader easily on many of the overbuilt high end ones). The cpu is putting out more than twice as much heat in half the area, so it is obviously much harder to cool.)
This is correct! In this case distance is the key. The thermal resistance of a system (heat source + paste/pad + cooler) is exponentially proportional to the distance that the heat need to “travel”.
NOOOO IT IS NOT EXPONENTIAL! IS PROPORTIONAL TO THE INVERSE OF THE DISTANCE
You are right a other thing is that the material get pressed very hard between two metal sides. it make the pad insulate
I think he just used a bad pad, My 3700K with the IC thermal pad and an NH-d15s doesnt go above 70c on all cores
@Phi6er good catch buddy LMFAO...
Anyone else notice how at 2:54 the pubg shortcut is named "Fortnite Hi-Res"... 😂😂
James Murray LMAO, all the thumbnails actually lol.
Other shortcuts are funny as well.
Keep Out haha yeah just looked 😂😂
Pubg doesn't even look good...
MeatNinja compared to what?
Jay, you fool. Everyone knows Colgate has superior thermals. For shame!
but menthol toothpaste would do best with it's coooooling taste.
But the crystals add a burst of flavor your CPU will never forget.
Gold leaf as a thermal transfer!
Damn that's a good idea :O Going to patent it and become a billionaire, thanx! :D
you cant afford gold, and even cant afford to patent it. and gold is conductive. :D
rlund3
That's stupid since silver conducts heat way better than gold.
its for the rich people u know they dont care as long as its expensive and something to show off with
good luck not tearing it when you mount the cooler
jay keeps toothpaste in his kitchen?
IC graphite is the one Linus tested. I really want to see gamers nexus test it. They're the only one I expect to get accurate information from.
Nexus owns a 'the'?
Hmm...
YESYESYESYES!
LTT video was fine, but it would be really nice for gn to review it because their tests are really comprehensive
Tomorrow Im going to be testing the pads with different coolers on a r5 1600 and an 8700k for work. I might post a google doc somewhere for it since its public information.
PureBlue okay little man, let's calm down. Go back to the kids table and let the adults talk.
Jay: This video is sponsored by Nord VPN
Me: (Checks date) oooh...
Someone please explain
@@jz2187 This video was made before the Nord VPN security exploit.
@@dekoldrick All VPN's are exploitable ;)
@@christophervanzetta It's about how NordVpn handled the "exploit"
@@modmen. Ironically, if NordVPN kept logs, they would have been able to find, isolate, and determine the scope of the exploit both faster and more completely.
Thermal pads are great at dissipating smaller amounts of heat, such as that from VRM's. They do get hot, but it happens slowly. You could not keep a computer running without CPU cooling, but without the VRM's cooled, the computer would stay on for quite a while. While thermal pads aren't as conductive as thermal paste, they do better than just plain air, so they are great for filling space, to provide some sort of contact with a cooler if they wouldn't already, as well as I would imagine, decrease, or better disperse the force exerted on the VRM's as to not cause damage.
In short, traditional thermal pads do work, but they don't work as fast, so only use them on slower heating items.
I bet you used some Ram Cooling Pads with like 1-2W / mk.
Like you said they just cant transfer the Heat fast enough.
Maybe repeat the Test with something like the Arctic Cooling Thermal Pad (6w/mk),
or one of the the Alphacool Eisschicht thermal pads (11-17W/mK).
the ic graphit ones should have worked too
since you wanna know when to use thermal compound vs. pads, I'll tell you!
thermal compound/grease is used for components that require quick heat transfer and experience rapid changes in temperature. a CPU can jump 30C in an instant, and that energy needs to move FAST. thermal compound spreads thin enough that there's barely any of it between the IHS and cooler's surfaces, and mostly fills in the gaps to keep air from insulating the heat.
thermal pads are used for components which are slow to heat, such as VRMs -- it takes a long time [relative to CPU temp changes] for the VRM of your GPU to get hot, and the pads are able to transfer the heat over that extended period. pads act like insulators [like in your CPU test] when there's a BIG dump of heat all at once because they need time to move the heat.
thermal PASTE is the same as compound/grease, but it's an adhesive [AKA glue], and would be used for permanently applying a heatsink to a component, such as, for example, a tiny heatsink on a Raspberry Pi, which will probably never need to be replaced or removed in most cases.
How is a piece of conductor not better for this application? I mean instead of the thermal pad, a piece of conductor will conduct better heat.
What if I use kryonaut as the thermal paste between cpu heatsink contact and 6mW/k thermal pad on the heat pipes which go near or touch the back of my laptop?
@@rohitbaviskar7748 you actually run into thermal resistance due to air trapped between them. Thermal paste or pads push out air and ensure maximum surface contact. Even if you have to conductors that are as close to perfectly flat as we could manage, there may be a tiny air gap (you can never have a truly perfectly flat surface) or they might not be aligned to lay perfectly together
@@tejaspadhye I wouldn't use pads *and* paste, that might actually reduce thermal conductivity since you're adding material. But having a pad between a section of heat pipe and the case might help transfer heat a little... Can't hurt to try
@@DoctorX17 I'm not using them together between heatsink and cpu die, but where the heatpipe comes in contact with the case, the case has a copper sheet typo thing attached just above the heatpipe. I'm hesistant and asking you this cuz the pad might melt and leave behind a mess and my laptop's still in warranty
Ther should not be any adhesive on thermal pads.
Thermal pads are used on VRMs, memory chips and such for a couple reasons:
1) those parts are less prone to heat damage than a CPU itself
2) more importantly: they need to dissipate less power for unit of surface that a CPU so they do not need the highest thermal conductivity
3) thermal pads are solid, while thermal paste is, well, a paste. It's much easier to cut and apply a pad rather than putting say 8 spots of thermal paste on an 8-phase VRM, so they are much more easily applicable than paste and, as said, they do not need maximum performance.
Another point in favour of pads is that they adapt to different thicknesses, like an 8 phase VRM, without problems of not making contact (if they put paste on a slightly lower chip it would not make a good contact while a pad can still touch)
BackInBlack_4896 Good arguments, I would just add two things :
- good paste application is much more difficult than placing a pad
- the cost might be part of the good answer
Honestly, I believe I speak for more than just myself when I say that if you tried to de-lid the 8700K (because of it intense heat levels) and tried to cool it, and also make it a "How to", people would be very interested!
Like I myself want to build my system with an 8700K, and I'm using a custom cooler from EK, and different Thermal paste, but, let's be honest, it's an area heater.
Or just grab a Ryzen 7 and you won't have problems like this.
@@nonci6 oh no problem with heat huh? lol
@@taylonjudd923 AMD re-did their manufacturing process and their CPUs don't heat uo as much as before like their FX series CPUs.
Jayz, Please make a water cooling system where one tube is connected to a water tap and then the outflowing tube from block is flowing into the sink so there is no loop and water is never heating up! I've never seen it done and haven't found anyone doing it, would be cool to see cold tap water cool a cpu!
There might be a reason its not done.
And the reason might be that theres so many that dont have acces to clean water, so it would be quiet hypocritical to do.
But thats just my opinion.
I'm mainly talking about the big youtubers such as Linus, Bitwit and such. Sorry if I sounded bad in anyway, didn't mean it!
dont apologize ....its a cool idea! he should try it
I really get the idea, but honestly it is terrible. For starters it will get costly quick and is ridiculously wastefull. Tapwater isnt free and running a faucet whenever your pc is on will get expensive.
PLS ADD BIG PACET CPU COLER SO I CAN SEND A EMAIL
Seems like goldleaf would be ideal for this sort of application.
Berserkasaurus Rex my exact thoughts! They really need to do this
Gold isn't all that good at thermal conductivity and is also electrically conductive. The only advantage gold would have is that it wouldn't corrode. Gold leaf would also be very difficult to apply without getting it everywhere.
@@LordMoebius Gold leaf sticks in place pretty well, in my experience. It's certainly less messy than paste, which seems to be the main issue with thermal paste. It's also easy to apply. Just put a little pressure on your sizing, and the gold leaf will adhere just fine.
Not sure why you think Gold isn't a good thermal conductor, it's one of the best metals for thermal conductivity. In fact, aside from some odd alloys, only Silver and Copper are better thermal conductors, but while both can also be used in a leaf application, they are thicker and will corrode, unlike Gold. The only thing better than those three is Diamond, which obviously wouldn't be applicable here without some serious technological development.
@@berserkasaurusrex4233 gold leaf has a tendency to flake during the application process especially when applying it on an intricate or small surface and you also need to use adhesive to hold it in place. So pretty sure it wouldn't be the best for thermal conductivity. And yea silver and copper are better conductors at a cheaper price point compared to gold. The only advantage you'd get is that gold wouldn't corrode but it's sorta irrelevant in this context so you're just paying for a material that performs worse than cheaper materials.
Plus gold leaf would have the same problem, it will not be able to flow into those microscopic valleys.
Try Graphite thermal pads
That is the wrong one, you mean graphene.
And yes it's a lot better then copper.
@@N0N0111 jayz made mistake, thermal pad works well with heatsink. Watercooling make thermal doen't melt suddenly frozen
@TheBlasterGameGuy doesn't not you understand words mouth coming from increase logic.
@@claritoresdiano1021 lmao wtf
@@probablynotabigtoe9407 english please
Thermal pads do make sense in GPU blocks if you consider them as spacers. It would be very costly I imagine to design a cooling block that is correctly spaced for all the chips on the card and it is much easier to use an X thickness pad to achieve contact. Another problem is that some pastes can short the chip's legs if it runs down (most are so small that it is difficult to push the correct tiny amount needed from the paste tube and even spreafing can make a mess of things). While some memory chips on GPUs tend to run rather hot, those temps are nowhere near enough the GPU's temp without a heatsink (they're probably in the region of 50-60C uncooled which tbh is not a high temp as far as chips go.
Next thermal paste on vrms? :P
nah just dip the whole thing in paste just to be sure
nah he'll put the thermal paste on the chokes
Αλέξανδρο ς very very bad idea
put IC Graphite on VRM'S (mobo and gpu ) and ( like jay sad: gpu ) to see if it's better than thermal pads.
Evolution[NL] Graphite is electrically conductive, so that's a bad idea
Late to the party! In engineering we choose thermal pads when you need to close gaps where you do not have control of component tolerances or a wide gap in component height that needs cooling. It’s a lot better than nothing and vary greatly in “thermal resistance”
"never gonna see this kind of load, especially playing games" - Someone forgot to play Cities Skylines at 200k+ population ;P
I rebuilt my daughter's EVGA 1080 that had the thermal problems using the EVGA thermal pad kit, but I used Thermal Grizzly Kryonaut instead of their TIM. Dropped temps by 8 degrees C. Thermal pads only work on lower wattage heat sources as the pads have a lower amount of heat transfer rate. This is what accounts for the variance in various brands of TIMs as well; they all have different heat transfer rates. Kryonaut has one of the highest transfer rates out of the other TIMs. You have to go to a liquid metal for the better heat transfer rates. When looking at any cooling solution look at the amount of heat it will transfer given the application. Also, if you lap the surfaces of the CPU lid and the cooler plate to make them smoother (and make sure they are absolutely flat (I've seen some cooler plates that aren't)) , this will also increase heat transfer rates as there is more metal to metal contact.
they definitely make sense on ram and vrms because they produce low heat.The pad will not be a bottleneck when it has to transfer like 5 watts of heat
Also because they're electrically insulating. You get some thermal transfer without risking shorts when in close proximity to 100+ Amps. Thermal pads are rated in terms of watts/area and they're a fraction as efficient as thermal paste. If you had a 3ft square heatspreader, then a big thermal pad would be able to transfer all that heat. A pad that small is probably only rated to transfer a few watts of heat, not the 300 watts or so from an overclocked 6 core.
Allison Pell 100Amps? Are u sure what u saying? Isnt that "100 A"? Maybe im missing something but... U got like i limit of 12 A in the wall, or it will cut out the power...
Gaspar Silva
you don't understand electricity at all my friend
amps times volts is watts. if you are pulling 100 amps at 1 volt, then you're drawing 100 watts from your power supply. your power supply manages how to pull the volts and amps so that you're not drawing too much power.
if you have two electrodes and there's a hundred volt potential between them, you aren't drawing power because there aren't any amps (no electricity is flowing). if you link the two electrodes, current will pass, meaning amps will start existing. the voltage will drop, depending on the resistance of your wire. if your wire has a high resistance, you won't draw as much as if it didn't resist at all (a short circuit). in a short circuit, you have no electrical potential (there are no longer two potentials because there is no two distinct electrodes) but you have a high current, meaning a shit ton of electrons are passing through your wire. this heats it up (pretty much because of friction) and probably melts it.
I'm dubious that there are 100 amps going through a PCB because if you want to have more amps going through a wire, you need a fatter wire. welding wires use twelve amps and they can easily be a centimeter in diameter (welding is short-circuiting metal to heat it up to the point where it melts).
VRMs may pull out about 60W of heat, but the heat is spread across a way bigger area than on a CPU/GPU.
oh, and VRMs aint flat, there are solderings etc arount it which also get contacted.
and since most thermal pads are non-conductive, you can't short out a FET and get 12V into the CPU.
under LN2 you see way more heat output, but since the CPU/GPU is sub-0, most parts of the Board around it are frozen however so heatsinks usually just heat up the Board, cuz they are contacting something which is not at - 50°C or smth.
Hi Thermal pad work excellent where the source like Vram, GPU, Ram or chip heat in these chips heat gradually increase so the transfer pads work well.
But like CPU where heat generate suddenly Pads fails. So it's better we use Thermal paste that evenly spread heat and removes spikes of heat that can damage CPU cores. Its better to use 6W/mk to 13W/mk for CPU. But check does the heatsink or water block sits properly on CPU surface. Note if not fitted properly on CPU NO Thermal can pass heat to the heatsink. Don't apply too much that will also ineffective. Apply the proper layer of CPU . It's only to fill microscopic holes of CPU surface and heatsink..
4:46 I took apart my two PS2 Slims and found plastic on the thermal pads. Taking them apart is easy and I recommend any other PS2 owners check their system!
@JayzTwoCents That type of thermal pad isn't for what you think. its meant to go on things like ram to spread heat through the chips in the case of vrm and normal ram. This is not for cooling but when the silicon heats and cools when in use so a hot spot could cause that part to expand while the rest is still stiff causing the sodder and silicon to warp damaging connections in sodder. This is the same for the case with heat spreaders on normal ram sticks.
I've used ceramic thermal paste ( some generic stuff for audio components , and Arctic white) to adhere some old zalman ram sinks onto a 6850 with a zalman vf1000 led heatsink "back in the day" ran it like that for a few years and they never came off/ fell off untill I changed the paste....just throwing it out there, I think it actually did better then pads
Designed for those cold winter days. Your CPU deserves a blanket sometimes.
"I'm gonna apply this with my finger."
Oh god its The Verge all over again
Take the leaf blower idea, but add in an intercooler from a junkyard turbocharged engine and get a second leaf blower to blow air across it and chill the charged air.
Hey Jay, I’ve got a dumb idea/suggestion for a video. It more of a curious experiment.
Here it is:
Delidding a CPU to replace the TIM between the IHS and the die with a graphite thermal pad. 🤷🏽♂️
Needed to renew my VPN service and switched to Nord from PIA with Jay's code. Been using it for a few weeks and so far I am loving it more than PIA.
The pads that everyone has been using are made out of graphite and are way better than the generic ram pads you were using. Kinda no comparison. Nice vid thought.
Btw your crazy rad setup with 4 480s has me looking into a 1280 rad external setup now. Maybe you can compare one of those to your setup.
That would be really cool. With the CPU thermal pad.
You going for a box fan our just a passive setup?
Passive really but Ill probably use a box fan if I have 2 as I don't want to buy 9 fans.
Yeah nine good fans can get expensive quick.
Themo pads are used where there are large gaps between heat spreaders and the module. VRMs have typically large gaps and there is higher heat transfer through a thermo pad than through the air. In general, conduction transfers more energy efficiently than convection. (solid vs air)
Thermo pads are also used where you cannot feasibly put paste. It would be impractical to cover VRMs in paste enough where it actually contacts the heat spreader. Also, the paste would ooze out of the card eventually due to gravity.
Amazing content!
I wished he had added CPU thermal paste to both sides of the thermal pad and tested to see what that does, would have been VERY interesting...
I wouldn’t have known about the pads if you didn’t make the video. And can’t wait to see the cpu pads in action. It’s doesn’t matter who made the video first without you making the video then where’s my excitement in my life. Thanks jay.
2:54
"I beat Linus"
"Fortnite Hi-Res"
bruh
10:02 better view of all the apps
The thermal pad predominantly helps to resolve alignment issues. Like if the plane of the cpu isn’t parallel to the cooler, only parts of the cooler will touch the cpu and not a lot of heat can transfer. How the pad does this, is that it is squishy and metal isn’t with the pressures involved.
To clear up, I wasn’t talking about a specific material of thermal pad, but rather just the general idea
2:57 Those are some interesting icon names you chose Jay. “I beat Linus” “Fortnite Hi-Res”
If not stated previously the surfaces are micron smooth -must use tissue paper not paper towel to clean surfaces and alcohol to clean out micro-pores first. when spreading paste it should be thin, thin, thin and worked back and forward several times so it "gets in" to those pores.
Thermal pads are cleaner and allow for a greater space to be crossed, if you look at some pads on cards they are thicker or thinner for optimal contact with the gpu/heatsink, less metal machining and cleaner. They are also used in testing applications, as in hey lets see if this cpu boots as fast as possible then onto the next. OH AND they work well in situations where something is warmer than expected due to bad silicon or design, like the recent fiasco with evga? sending thermal pads to 1080 buyers.
also helps with unevenness of components
Makes sense, and you kind of answered your own question. The thermal paste is used to cover micro cracks and displace the air, air is a poor heat transfer medium. The thermal pad just creates pockets of air underneath and in the cracks plus introduces an additional thick layer to the heat sink that creates additional heat transfer resistance.
That crunch in the toothpaste isn't the "Whiting Crystals".
Toothpaste is an abrasive. It's microparticles in tasty soap. Basically silicon and soap. It cleans your teeth and whitens your teeth by literally scrubbing and scraping them.
The whitening crystals (lol) are usually just things like peroxides and light bleaches in crystalline form.
I have a Turris Omnia router, which has a small aluminum cube as a cooler on top of Armada CPU, I changed the thermal pad on the CPU to a thinner one and added another on top of the cube, so it started transfering heat to the chassis (which is metal). Long story short - average temperature dropped by 30 degree.
So it works in certain situations.
3:03 Really appreciate the zoom
TIM is meant to fill gaps, it has worse conductivity than metal, but better than air, so you want it to be as thin as possible. VRM components have much higher variability in height than a heat spreader would, so much that without a thermal pad, they probably don't even contact the heatsink.
As for temporary solutions, I'd be very curious to see how a drop of water(distilled) performs. Obviously you'd need a flat surface to mount the heatsink, but once mounted, surface tension should keep it in place long enough to do some testing.
Guess you can use non-conductive thermal paste on VRM and memory?
To answer your question it has to do with heat capacity vs thermal conductivity. The pads have a low capacity but ok conductivity. Dumping too many joules into the pad, more than it was designed to handle, will saturate it and create an insulating effect. It's like the gauge of copper wire, kind of. Its conductivity isn't changed by the thickness of wire, but can transfer more heat the thicker it is. The pads and paste have different physical characteristics and properties, but the concepts are the same.
I bought one of those ic diamonds thermal pad and my temps increased by 10C at idle.
Damn my cpu went opposite lol. Haha my experience was better than yours!
Random, crazy idea to try: without any fans attached, submerge the radiator part of a water cooling loop inside a bucket of ice water (variations could include dry ice and other crazy stuff) and see how well it cools (or goes gloriously wrong). You could also try potentially lowering the temperature of the water using salt (like ice cream salt for those homemade ice cream makers). Give it the ol' ice bucket challenge this Summer!
The thing about thermal pads is: with a cpu they get saturated with heat and they can't transfer anymore heat, but on something like vrm or ram they have a lot more area to transfer the very little heat (in comparison to a cpu) and don't get saturated. It's like spilling water and cleaning up with a sponge and then seeing a flooded house and trying to clean it up with the same sponge
Damn, who would have thought? But I'd definitely agree, the installation theory really makes the most sense as far as the thermal pads.
Folk: I use regular thermal paste to help cool down my CPU, what do you use?
Me: It's a little complicated...
Hey, you should mention that thermal pads are not meant for high psi CPU mounts. They're bulk gap fillers.
Under a high psi CPU mount, the copper IHS and the copper cold plate is being elastically deformed by pressure to reduce the need for gap fills.
With thermal paste, the remaining patches of poor contact is filled to prevent an air gap.
If you use a thermal pad, it prevents all metal-metal contact and increases thermal resistance over just bare copper.
use bacon as a thermal pad
The reason and the application where those pads are used, is for when you have a single heat spreader or cooling block for multiple chips because no matter how exact the mounting of the chips is, there will still be differences in height between them. If you put a flat cooling block on top, it will make contact with some, but have significant gap (well, significant in terms of heat transfer. Like 0.01mm) to others. So while you would get better cooling on some chips, you would get virtually no cooling on other chips. By using these pads that are a bit thicker, you can even that out, and since they are a tiny bit squishy, the ones on the thinner chips will just have a little bit less pressure on it which doesn't affect cooling. The reason it becomes worse than nothing for the CPU, is because compared to direct touch, it actually is worse since no paste or pad, while you will have air gaps, you will also have direct touch to SOME degree and these pads, while superior to air, are far FAR worse than directly touching, even if that direct touch is over a much smaller area in total. And it also works as an electrical isolator (be warned, the CPU version you talked about at the end, is different and IS CONDUCTIVE) but that isolation is more of an added bonus, not the reason they're used.
How about extreme polishing on both the water block and the heat spreader on the CPU? Smoother, super flat surface should help transfer the heat more efficiently without the need for paste. DO IT! Please :)
do you have millions to do it ?
No ?
That's why we can't.
When you polish something, you're just making thinner scratch/dents in the material.
For a PERFECTLY flat surface, you need an ultra-precise machine, with an ultra-expensive cutting tool. It can easily cost hundreds of thousands of bucks.
Thermal Pads are actually transferring a bit of heat away from VRMs/RAM/Coil and Stuff.. The special thing is, they can bridge the gap inbetween the Cooling plate and the Parts on the PCBs. We use Thermal Pads at work also. The Parts on PCBs are not always the same height after soldering, and to catch all the tolerances, thermal pads are used. When PCBs for Spacecrafts are built, the cooling plate is actually fabricated after the build of each individual PCB. The PCB is measured in 3D to micrometers, yep, really! and then the Cooling Plates are milled to these measured Values. This makes the gaps between the Parts on the PCBs and the cooling Plate really small, so we can use very thin Thermal pads. The Thinner the Pad, the better the heat is transferred. There are also HUGE differences between some Thermal pads, the are som cheap really crappy ones, and there are some good ones. The ones we use for Spacecrafts are awesome but expensive as HELL.
I built a dope gameing / content creating pc and accidently forgot the "WARNING!!! PEAL OFF BEFORE USE!!!" sticker on on the cpu block of the masterliquid 240. I used it for a half year with an overclocked 7700K and complaind a lot because the temps i got with a prety good AIO. Than i decided to delid the cpu and when the liquid metal arived and i went to take apart the whole pc i recognised the f*cking sticker i just replaced the thermalpaste and the idle temps moved from the high 40s to the low 20s.......
Thermal pad makes sense in a situation where you have lot of different heat sources, in that situation efficiency doesnt matter as much as being able to transfer heat to any surface that can be cooled
Thermal pad = Thermal blanket.
He is using the wrong ones, because he could not get the graphene pads.
Don't get fooled with the word "Thermal pad"
If you dont want to use Thermal Paste on your CPU, I recommend Graphite Pads. Innovation Cooling (The maker of IC Diamond) sells them at Amazon.
They are quite good and they are reuseable.
2:54 awesome shortcut names..
I like the one above Fortnite Hi-Res
dear Jay. if you would have consulted the technical properties of these thermal pads, you would have noticed the very little thermal conductivity of the compound.
the reason it is used is because it is the best solution for a heat conductive material that is elastic. because the heat sink isn't as firmly connected to the ram/vram and it needs an elastic connection.
the only possible better solution to thermal paste or liquid metal is to weld the cooler onto the surface of the processor. that is the only way i know without a thermal barrier between two materials.
thank you very much for all your testing and even dumb ideas. they are entertaining and sometimes shone light onto a matter from a very different angle. keep it up.
Another idea/experiment: use pencil graphite instead of thermal paste... 🤷🏽♂️
Power MOSFETs generally only dissipate a couple of watts. They get hot because they are often not actively cooled, and so generally have a high thermal resistance to the air (on the order of 50°C/W, meaning each Watt of power output raises the steady-state temperature by 50°C).
Thermal interface materials have thermal conductivities expressed in Watts per meter-Kelvin. The weird unit comes from the fact that thermal conductance (the inverse of thermal resistance) is proportional to the area and inversely proportional to the thickness.
A cheap 1W/m.K 0.5mm x 10mm x 10mm thermal pad will have a thermal resistance of about 5°C/W, meaning if a single MOSFET dissipated 5W, the difference in temperature between the MOSFET and the heatsink would reach about 25°C.
Meanwhile, a CPU is capable of dissipating 100W or more of heat. A 0.5mm x 25.4mm x 25.4mm 1W/m.K thermal pad would result in a steady-state temperature difference of about 80°C between CPU and heatsink at 100W.
The best silicone thermal pads have thermal conductivities of about 5W/m.K, resulting in a steady-state temperature difference of about 16°C between CPU and heatsink (less if compressed) at 100W.
The IC Graphite thermal pad has a thermal conductivity of about 35W/m.K, which assuming the pad is about 0.1mm thick 30mm x 30mm pad would give a thermal resistance of about 0.003°C/W, resulting in a temperature difference between CPU and heatsink of about 0.3°C at 100W.
I've got a strange idea for you, Jay: Install something like Jack Daniels, or Everclear 190 proof in a watercooling setup, lol ;)
heat energy transfer relies on the temperature gradient and thermal conductivity. If you can accept a high temperature gradient, you can have the same heat energy transfer with a bad thermal conductor. But if you a re limited to 100°C on the chip, you need a better conductor for better performance. Pads are in the end just cheaper and are able to hold small cooler in place without screws and such. You can use them if you want to oc you raspberry pi or smt like that.
Can you recommend me a non electrical conductive thermal paste for a laptop cpu and gpu?
KPx High Performance Thermal Compound 10G
To answer your question jay i believe thermal pads typically make sense in areas where your heart source and sink are farther away since pads can be made thicker than paste and transfer better than air
when will Intel replace their thermalpaste?
Cavey Manta i vote for you being a CEO at intel
Covio dont tell anyway, but its called sarcasm shhhh
When AMD gets at least 35% market share
Adam Smith from sales point of view amd is roughly tied with intel now
Well with Sandy bridge they used solder
Thanks for the video. Regarding your question about the idea of thermal pad and how it works, I should say that when heat (or any other things like sound, light, motion, electromagnetic waves, etc) conducted through a medium everything depends on the conductance capacity of that medium and as long as the medium is homogeneous (no abrupt change in the material) everything is fine. However, when heat reach to an abrupt change like passing from one medium to another one the challenge happens. Exactly in the junction you have an abrupt homogeneity which means that the conductance capacity changes abruptly because of two reasons: first is that the junction surface is not microscopically perfectly flat and there will be thin layer of air there (air is a heat isolator) the second reason is the difference between heat conductance capacity of the two medium (the more difference the worst the case). To envision it assume that you are on the roof of a building and want to go to the nearby building if the height of two buildings are the same (conductance capacity) then you can easily go on the roof of the nearby building but if the heights are different you cannot go and you need a ladder (conductance matching) which helps you to smoothly pass the heterogeneous junction. You can model the microscopic air-gap with the space between two buildings. This is what thermal paste or pad plays in the junction, providing a ladder for the heat to easily pass the junction for both removing the microscopic air-gap and conductance difference. The specification that the thermal conductance should have is that it should have a conductance capacity that matches to both mediums.
Interesting. Does this mean we should use thermal paste instead of pads on the gpu rams and vrms?
No, though you could if you want for personal preferences. Thermal solutions carry eg. 1w per square cm surface, at 0.1mm solution thickness, at a temp differential of 10C, or 2W at 20C, linearly. Depending on the surface available for transferring X watts of heat at a predetermined differential, you can use different compounds or methods. A CPU puts out around 10-15 watts of heat per square cm, so you want paste that can transfer eg. 10W/cm2/10C (150W CPU / (10W x 10cm2 surface) x 10C = 15C differential); VRAM and FETs only put out 0.5 - 2W per cm2, so in this case you can use potentially cheaper, easier to handle, less messy and reusable pads that can transfer eg. 1W/cm2/10C (2W FET / 1W x 10C = 20C differential). If conductivity requirements aren't particularly demanding, whether you want the potential benefits from pads or not, the choice is yours.
Obviously, this means you should use toothpaste on your VRMs
I've had great results with graphite thermal pads less than .02 inches thick. Thickness and material composition counts. Silicone pads are 1-2 W/mK rated.
2:57 The shortcuts are hilarious, 3dMark says beat linus, PUBG saying fortnite hi-res.
Use a thick layer of thermal paste on memory chips and VRM instead of a thermal pad, got it!
Btw, you have to try homemade peach marmalade, it's the best household item to use as a thermal paste substitute.
VRMs output much less heat and diffuse it in a very different way. You don't even need VRM heatsinks even with an overclocked 7980XE as long as you have good airflow over the MOSFETs. You obviously can't cool a CPU die with airflow alone. Thermal pads work, just much less effective than thermal paste, and they cannot conduct heat very quickly, hence the naked setup outperforming the thermal pad. There is a reusable thermal pad for CPUs now that is actually a revolutionary product as it's effectively solid thermal paste. I would have liked to see that tested in this video vs. the stuff that you shouldn't use, but I'm weird in that I like useful content.
Is it OK to double up on a silicon pad if it isn't thick enough? Is it OK to use a pad by itself with no heat sink?
Have you considered pumping chilled air into a case rather than using liquid cooling? It's a thought that some may say runs the risk of getting condensation on your hardware, but unless you pump chilled air in after you power down the system or you make the hardware colder than room temp (something that would be crazy to do anyway) you should have no trouble with condensation. At least the science of it all says that the hardware has to be colder than the air it's in to cause such a problem.
I've considering using a solid state cooler to chill the air that is being drawn into my laptop as a means to help it run cooler, and that's where my question comes from.
Lap CPU, then try pad again?
Probably LAZY to do a lap CPU.
I would say lapping and thetgraphene pad could be interesting.
Nisco Racing Lazy how?
Well it takes elbow grease and many hours.
Nisco Racing So you're saying Jay is too lazy? lol
i mean why not i guess but lapping it when you're using a Thermal Pad is pointless, haha
it's already 'hella thick' either way :D
the pad is there to somewhat cool the VRAM but has thermal resistance to prevent the heat coming from the GPU to back draft.
6:33 OMG USE BACON!
Thermal pads work when the surface area is large when compared to the amount of heat being generated. On a GPU and RAM the surface area is large and as a result the insulating property has less effect. But on a CPU the heat is concentrated in a small surface and therefore the insulating properties prevents a larger portion of heat from transferring. The pad essentially creates a more even transfer of heat on the microscopic level but due to its insulating effects it also creates a heat transfer bottleneck. The bottleneck can be lessened by a greater surface area.
So basically, thermal pads work fine when the demand for heat transfer compared to surface area of the transfer surfaces is low.
"old ass game" is the name of the FSX shortcut, love it
One comment suggested Gold foil, made me look at some charts and it seems gold and aluminum performs roughly the same.
Both are easily available , So please JayzTwoCents , test it. (aluminum foil & gold leaf)
Another thing you could test is the heat transfer goop, you can buy for engines. Its some kind of copper stuff, and I remember seeing it some time ago, wondering if that would work. ( off to the hardware store we go! )
the only thing I can think of for why the pads are used for VRM and such is because the surface area is smaller on those, so the pads help add more material for the heat to be absorbed into rather than help dissipate the heat into something else.
or at least the thicker pads. The new thinner ones are designed to actually transfer the heat
what i know of thermal pads on lesser components is that they are made to handle a higher temp. so they need less thermal dissipation but without it they still get to hot. the thermal conductivity is better at higher temperatures differences. those components can work fine at 80-90 degrees. the VRMs dont put out as much energy they are just designed to handle more heat.
Reasons for Thermal Pads:
1. If there is a moving connection between parts, which thermal pads can buffer.
2. If you have too get thermal connection without electrical conductivity.
3. If there is no airflow on very uneven surfaces.
Greetings from Germany ;-)
Can i repalce the thermal pads on the Vram with thermal paste ???
Please help me
Pads are intended for use where you need to transfer heat with less pressure applied, and an area where paste wouldn't be thick enough . As in, the gap from the component to the sink is too big. I remember back in the day when a lot of OEM's used pads on CPU's, urk.
You have to mix toothpaste and petroleum jelly half half. It works very well but only for a couple days then you'd have to re-apply.
Thermal pads are actually useful when you have a gap to bridge. it can be because the component surface is away from the cooler and you can not get the 2 surfaces to touch for some reason. Although insulating, they insulate far less than the air that would be there without pads.
Another interesting application is when you get a seriously uneven surface where the thermal paste would not fill the gaps reliably enough.
To be noted though is that thermal paste is actually the same... it insulates but just far less than the air that is there without paste. We see less of an impact as, you said it right, the layer is way thinner than with a pad.
How about a "radiator free" water loop with like a 25 gallon Rubbermaid tub to see how Long it will take for the water to exceed it's cooling capacity. Also use temperature fittings to monitor the water temp. It's something I've been wanting to try, but I currently lack the resources at the moment.
You should try gold leaf. It has excellent thermal properties and it is pretty soft, especially 24ct gold. It would take a little experimenting to get the right thickness, and to apply it with zero creases, but it should work well. Ironically, even 24ct gold is cheaper than many of the TIMs out there (gives you a little perspective on how expensive some of these compounds are).
I got the graphite pad yesterday. Ordered after watching linus's review.
It works well best used with a Water cooler.
The graphite pad tend to keep the temp of each core steadier then thermal paste.
I'd like to see metal hard lines, or a scrap/ car themed PC build. I think it'd be cool to see you doing odd stuff with pieces that don't belong in a computer; gluing sheet metal blades on to fan hubs, custom ducting, using a giant car radiator and e-fan mounted on the side of a case, driving a whole car ac system to use as a chiller. I know that kind of stuff can get really complicated, but I, and I think many others on this channel would love to watch it
Fuck this is a 6 month old video, I'm gonna post this on his next upload
Necrobumping here ... It's probably already been said ... but the pads are there to prevent an airgap. The pads suck vs paste or direct contact, but they're there to fill in when you'd have an air gap.
There is normally a small dip of height in the center of the intel cpu aluminium lid. So when you put a thermal pad as is, you will have the thickness of the thermal pad PLUS the air gap that you have when no thermal interface is used. so you will have high temps. try cutting a smaller area of a pad to cover the air gap only. you can use some paste to fill the gap and a straight exacto blade to pass over and gather the excess to see the gap.
@JayzTwoCents They should probably design a thermal material that comes in like a foil form. You know, it melts when in contact with heat and turn into gel form but returns into soild form when it cools down.
That would be nice.