This video has grown so much. It’s difficult to keep up with comments after the first one. Please message me and our discord link in the description! Otherwise, I might not be able to get your comment to help you.
You simply did such an amazing job with this test and deserve every positive comment here :) Also, I would like to add a small positive side of keeping memory cooler and core hotter : Since metals (like copper/tin, etc.) get physically bigger the more heated they are, keeping PCB temperature even/uniform across all areas should help in preventing traces/joints being damaged from material fatigue over long period of time (because difference in temperature creates different stretching/compressing mechanical forces inside PCB layers). I don't know how relevant this is to longevity of the card, hence I classified it only as small positive :) It may not even be a problem for 5-10 years, since it depends on max. temp difference and number of hot/cold cycles each card had after all.
@@TekShinji I really enjoyed this... (concerned what this says about me) lol. But the Thermalright pads are the ones i used on mine after seeing loads of videos and recommendations. Mine dropped from 104C to 88C although i did not just replace the pads on the memory modules i did the VRM pads too which i suppose would allow the heatsink to draw closer to the modules and thus create the proper contact which given the thickness deviation from what was advertised is probably quite important. Could you or did you test the thermalright pads but with more than the memory modules replaced? I and others have had excellent temp drops with them and sadly it seems your results are the anomaly which is distressing as yours seems to be the most concise testing i have seen.
Dude this is awesome - thank you for not just stopping at which one performed best - seeing the core temps rise due to the increased sinking heat from the memory was something I had never considered
This kind of information is worth gold. Thank you for doing this (and it makes me feel better about getting the Gelid's in the first place). Comprehensive, easy to understand and to the point. Also worth to note, using thicker pads means the contact pressure between the cold plate and GPU core is less, which usually translates into a few C increase on the GPU core. It's a tradeoff, cooler RAM means hotter core.
Hey there, thanks for the comment on the feedback. Regarding the thicker pads. This also happened with the thinner pads and every pad we used was relatively close to the tolerances listed. Because of that is why we determined there was more thermal saturation in the cooler.
Great job. Some of my favorite videos are ones by guys like you and Seb who put in a lot of hours like this to produce a single video. It does not go unnoticed, my friend. Adding all the graphs and numbers gave this one a real Gamer's Nexus feel and I know you must take any comparison to them as a high compliment. You deserve it. As for others who may read this, I've used a generic Amazon pad similar to the one TekShinji does in this test. The kind you can get in the big 200mm x 200mm sheets which is HUGE (you can change many card's pads with one of those). They got the job done for me at a low price. They won't win you any bragging rights if that's what you're in to, but my Gigabyte 3080's been happily hashing away for 5 months with no loss in performance or degradation (knock on wood). My VRAM temps stay at 90C with 80% fan speed with pads replaced on both sides of the PCB. Nothing additional added to the metal backplate. No special fans directing additional air for supplemental cooling. Just one card inside a gaming PC case with the side panel off.
P.S. It's also worth noting, that in my situation, my 3080 required 3 different thicknesses to match what came on my card. As TekShinji mentions in this video, it's extremely important that you get the fit as close to perfect and not substitute the wrong pad sizes because you don't have what you need available. Because I was looking at 3 times the expense of someone doing a card that uses only 1 pad thickness, it led me to try the generic type as well.
Hey there buddy, thank you so much for those kind words. It really does mean a lot to us, this video took nearly two weeks to make and a lot of help from our wonderful resident scientist. Andrew. He really did almost all the work on this video. I just let him go ham! We hope to make more videos like this in the future. It's just it's extreme time commitment and a lot of work. Regarding the thickness of pads, that is true. Every card takes a different type of pad. That's why we are comparing the brands and their claims rather than making a how-to or which pad to use on which card since it varies so much.
Unbelievably comprehensive, informative video. I do need to repad my graphics card, but I'm a lot less worried about squeezing every degree of cooling out of it. You raised a good point about increasing core temps, I've never heard anyone else point that out before. Kudos
Glad you enjoyed the video! Make sure to get subscribe and hit that like button for the algorithm! Yes, that’s why we concluded that the best pad is depending upon your application. Not necessarily the best thermal transference.
I can't thank you enough for making this video I was really struggling with deciding on a thermal pad brand. Before watching your video I was honestly going to just dish out for the Fujipoly but thanks to your video i ended up with GELID. 1070ti is now sitting at Now: 30.5MH/S @ 60c, 105W, Fan 60%, 290~300kH/W Before: 30.5MH/s @70~75c but with maybe 200~250kH/W Overall (thank to your video) I've noticed a SIGNIFICANT increase in CONSISTENCY. The 1070ti (Founders Edition) required .5mm, 1mm, 1.5mm, and 2.0mm thermal pads. I measured each and every pad to ensure they remained the same thickness. I do have to warn though that the GELID thermal pads are very tough to reuse as they are like a clay substance that sorta crumbles after taking off / putting back on.
I think a contributing factor to the increased GPU core temp with certain pads, is that they are reducing the pressure applied to the core, leaving a thicker paste layer. I expect some temp increase with better VRAM cooling, but not 20c+ Would be interesting to see a high end thermal putty in comparison, since it allows for optimal core contact / pressure.
I actually just got a 6700 XT that was a factory refurb with fresh paste and pads. The consistency / density of the pads appeared to be on par with Gelid Extreme, so pretty good quality to begin with. I tested with Furmark (1080p 8X) before and after switching to U6 Pro thermal putty and PTM7950 for the core. It dropped -2c core, -10c hotspot, -4c VRAM and -5c VRM. Went from a hotspot delta of 18c, to 10c. Ambient temp was identical between tests and I used a static fan speed. Let the cards heatsoak and reach steady state before recoding any data.
It´s good to add, that you should always measure your old thermal pads and add new of the same height. Otherwise, you might encounter issues, like higher GPU core and/or VRM thermals, when using thicker pads. Or higher memory/VRM temperatures, when using thinner thermal pads. Another necessary point is, each card can have multiple sizes of pads on various places, so you need to have a supply of multiple sizes of pads.
@@TekShinji i was replacing the pads on my GPU last weekend and these were my findings. I had some trouble with measuring stock pads, so i used thicker pads for VRAM, as a result, i had insufficient contact at core and VRM. There is similar issue with backplate, especially when there aren’t enough screws near the padded area. It’s also necessary to understand, which area should be covered by the pads, since the manufacturers tends to use pads as simple & cheap as possible - they´d rather use 1 longer pad, than 2 or more individual pads for each VRAM chip.
Thermalright pads are NOT measured on the OO scale. (Sc) means they were measured using a shore C durometer which gives a very very different hardness rating. Although that rating can give you an idea of its hardness, it can not be compared directly with the others. The hardness of the Thermalright pads is noticeable higher than the others which is part of the reason they appear to give the most inconsistent results between users. They can not compress anywhere near as much as the other pads so they struggle to fill the gaps between the cold plate and memory modules. The main reason pads are used on memory modules is because their heights from the pcb are inconsistent due to the production line's soldering process.
It depends on your application and the exact contact pressure required for your card. However, in our testing they do not appear to be made correctly so they automatically fail
This video was excellent. You managed to answer every question I've ever had about thermal pads in less than 20 minutes, thank you. Fun fact, I just picked up a MSI Ventus 3x OC 3090 from Best Buy and it has the grizzly thermal pads on the rear from the factory.
Thank you so much! I'm glad you enjoyed the video. We put a lot of work into this and it's nice when we see stuff like this. You sure it has the actual grizzly brand? they're like a salmon / pink color doesn't necessarily mean it's from them.
Funny. I was coincidently recommending Gelid Extreme pads to a friend with many 3090's exactly 3 days ago lol. I personally use Fugipoly but I don't think it's always worth it to invest so much in 17w/mk pads for the so little difference it can "potentially" give. From personal experience, Gelid always been the ones that gave me the best overall results. Nice video by the way!
Exactly, from our testing. There doesn't seem to be much noticeable difference if any. Between the two thing that it's significantly cheaper, it doesn't make much sense.
The Gelid tested here is 'Ultimate' which is considerably more expensive than 'Extreme' (as in twice as expensive) and has much higher conductivity. The Gelid Extreme wouldn't match Fujipoly Ultra Extreme as Gelid Ultimate does.
@@EliteRock hmmm obviously?😂 what's your point exactly? Actually, Gelid extreme is 12w/mk while Ultimate is 15w/mk. Fugipoly equivalents are 13w/mk and 17w/mk. Honestly, the difference between both extreme vs Gelid Ultimate doesn't justify the price gap between them. Only Fugi Ultimate is really better but still, 75$ for a single pad is overpriced for the little improve over the others it gives. Beside, pricing will differ from one location to another. And like Shinji have said, it all depends on the use you have for them since resistance will vary between most of them. I know Fugipoly extreme and Ultimate have the same resistance but I don't remember for the others..
@@danielchicoine79 Gelid Extreme is 8.5w/mK, not 12w/mK. I paid UK£20 for two 120x20x1.5mm Gelid Ultimates six months ago (for VRMs on a MB that are notoriously under-cooled) but they've dropped to c. £16 since then. Looking at prices here in the UK again, Ultimate now seems to be about 30% more than Extreme (for some reason sheets of Ultimate are bigger than Extreme at 90x50 vs 80x40) not twice as expensive as I stated.
@@EliteRock , odd - just checked the boxes for my Gelid GP-Extreme and they all clearly state 12w/mk. Am I missing something with them being 8.5w/mK somehow?
Hi, very nice and complete testing, however I can bring my results which defer a lot from yours .. both me and a friend got a 3080Ti FE. Both suffering from very high stock VRAM temps. (throttle 108-110° during heavy mining, fan immediately to 100%). We both did a repad at almost a similar time (end of August). I went for the Minus Pad 8 , he went for the Thermal Odyssey, both 1.5mm front and back (i tried 2mm first, that did not work, not enough Core contact). We had roughly the exact same temps during mining afterwards during gaming or mining (tried ETH, RVN, ERG), within the same 2°, both excellent, easily down to the 80° with about 60% fan. BUT, more recently I noticed that my temps were getting really bad again, to the point of being back to the 104-108°, at the same or even higher fan speeds (I was mining constantly since then). I took it apart again and ... it's like the Minus 8 pads "leaked", and had water under them now sitting between the ram module and the pad. (I have some pics). Cleaned up everything, and went for the same Thermalright Odyssey this time around. Back to the amazing 70°ish for gaming, 80°ish mining. So yeah, durability and ability to withstand during time is actually super important as well. Short sum up from my side : Minus 8 Pad and Thermalright Odyssey perf is similar (approx -24° VRAM same conditions), but Minus 8 Pad degraded within 2 full months of constant mining. Let's see about the Odyssey
Jaques dewever hello can I ask you a question? today I ordered the thermal right thermal pads precisely the thermal right Valor Odin 15w, I ordered them at 1mm for the memories, because ek waterblock declares that on the Vector ftw3 only the 1mm ones can go, I did well to order them right at 1mm and not more than? I hope they don't press too hard to reduce their thickness. thank you
*WARNING* I use water cooling on 3080 and went from 68 C to 96 C in 8 months on Gelid which is tested in the roundup. All silicone based thermal pads are leaking silicone and sooner or later will become junk. ~100 % of pads in consumer market are silicone based but Google it - non-silicone based thermal pads do exist. You will not see claims like 13-17 W/mK but if you do not need to replace them every 6-8 months it is fine I guess.
@@BlackPhilosopher nice, I was wondering about durability as well. Despite the good results on the three most right, there was no way for them to show durability. Thanks for sharing!
I just take a university class or what. Was so compact and informative yet still easy to digest. Takes me a while to find peeps I can comfortably listen to but I think I just found one. Thx for all the knowledge!!
Try Gelid Extreme instead. They're much softer compared to the Gelid Ultimate. One possible flaw with this test is the harder pads do not compress as much, therefore leading to insufficient core mounting pressure. Another idea is to try thermal putty (TG-PP10). They don't have any hardness, so they can potentially allow the heatsink to be closer to the modules/core allowing for higher clamping and thermal transfer.
Really impressive, professional video. Heck, it’s better than professional. Very thorough breakdown of the stats in an easy to understand way for the lay person. I love the narrator’s calm, relaxing voice too!
We only have the best narrators! Yes, Andrew did a lot of work on this video. There's nearly two weeks of work in this video. We're glad that it was easy to digest and easy to understand!
you are absolutely correct in your analysis: evacuate the heat from the heatsink as well to avoid increased core temps. I would like to see a full copper sheet used vs the more local "raspberry pi" heatsinks.
@@TekShinji check out ebay if you guys are interested. The chinese have the concept for sell at different thicknesses and tape applied. I went with scrap 5/16 copper tubing, but for testing & videos the plates are ideal.
@@TekShinji im referring to copper plate in this comment. They are sold in different thicknesses on ebay. Send me a friend request on discord and ill send you a link. Again this isn’t necessary for what we do with regards to mining but it would extend the life of the component by removing heatsoak.
I've repadded a lot of 3090 FE , I've had no problems with the thermalright extreme odyssey , they measure as 1.5mm as advertised using a micrometer . My benchmarking involves running the card at 105C max on the hotspot .On a stock 3090 this usually involves underclocking the ram significantly and usually gets something like 80Mhash . With the thermaltakes and a repaste I'll get 120Mhash with a very high overclock and temps will usually stay at 98 or below with fan sitting at 60% ( which means the card can be pretty much silent when mining). I've used Gelids as a comparison and found zero difference in performance. I use gelid and thermaltake on Zotac and Msi and not noticed much difference between the two . Only difference being going for 0.5mm thicker on the gelids vs stock pads due to softness . I generally prefer to use the thermalright and take care to precisely measure the pads with a micrometer ( its all 1.5mm on the 3090 FE) on most 3080 and 3090 other than FE you only need to replace the memory pads front and back , the other components are usually fine with the stock pads whihc is just as well as they often use odd thicknesses like 1.3 and 1.7mm ( its totally fine to sandwich pads together to get the thickness required by the way)
I refurbish gaming consoles and have been using Gelid for a while now. They are affordable and result in a huge noise reduction, especially on PS4 Pro. It's good to see some numbers backing up my experience with them. Thx for the in-depth video.
Hey, I'm a little new to thermal pads, but I was wondering: can you use these Gelid thermal pads inside a laptop? There's this separate chipset inside my old laptop that the heat pipe also goes to, and there's a little thermal pad on the chipset. I've measured the dimensions of the thermal pad and the new pad should be a 15mm square that is 1mm thick. I needed thermal paste for the new CPU and figured I'd replace the little thermal pad on the other chipset too while I had it apart. TL;DR: Can you use these Gelid thermal pads on chipsets inside an older laptop?
Good vid...though a little misleading. I run 2 3090 FE's in one of my mining rigs and use Thermalright. They sit at 74 and 78 degrees (virtual memory temp) at my house. You have to get the thickness perfect. I think there are more than just 1.5mm pads that get used in this GPU. I wasted the money and bought all sizes .5mm - 3.0mm and then just matched it to the compressed thickness of the stock pad that was used. I had to remount a couple times to figure it out, but now it's perfect. It seems obvious with that poor of a result on a quality product that there's an issue you may want to retest. Just some things to consider! You're definitely right about there being a lack of content, especially for the rare 3090fe. Thanks for making it.
Same here. For my 3080 I used thermalright 2mm and at first my mem temp went from 106C to 88C but my gpu die increased from 48C to 68C. I then took the card out and really pressed very hard on the locations where the pads were and also on the center of the cooler where the gpu die is. Now Mining 48C gpu die and 88C memory at 102Mhs.
Thermalright 1.5mm for gpu memory on my watercooled Aorus Xtreme 3090 with Bykski Waterblock. They're definitely firmer than other pads, but 1.5mm is perfect. 76C max vram temps on a mining test with a higher vram memory overclock.
Thank the lord you made this video. I've spent hundreds and hundreds buying thermal pads. Then I finally thought, "Do I need to keep buying these expensive pads from Amazon?" I then go to AliExpress and see literally the same pads you found. I then prayed I could find a video of someone testing this before I dropped money and time doing it myself haha! THANK YOU!
Wow, I love how comprehensive this is. I really appreciate all the work you put into this video. It's hard to wade through all the threads on Thermal Grizzly Minus, Gelid and Fujipoly, having a video to reference cost vs effectiveness is awesome! I think I'd go with the Gelid pads, seems like the perfect middle ground. Do you have any experience with the TG-PP-10 putty? The EVGA 3080/90/Ti stock cooler uses custom sized pads and their own putty, trying to find a good alternative to bring down VRAM/Junction temps.
I want to check that stuff out too. Haven't tried it yet myself though! Thank you for the kind words. A lot of work went into this so we appreciate it:-)
I love the methodologically sound and accurate approach to presenting data. Keeping as many variables stable/unchanged was a moment I started nerding out, so good to see reviewers take a scientific approach fully and seriously!
That's weird, I would have thought that reducing memory temperatures and so reducing temperatures on the PCB and so neighbouring components would lead to decrease of the core temperatures, but it's opposite, weird... Note to self: Best choice pads are Gelid GP Ultimate 15W or same under diffrent branding NAB Cooling NB Supermax 15W
note to you 😁:I can confirm after 4m of trial, gelid are very solid pads(i dare to say the best thermal pads. just make sure to get the right thickness). they are very good if you go with whatever the manufacturer had in there too if you add thicker they won't compress and you will get inconsistent/poor results
I know it's gonna sound crazy, but an ultra thin layer of thermal compound on each pad before making contact gave me INCREDIBLE results versus just thermal pads direct to the surface.I mean the smallest amt of paste you can use, super spread out thin with no extra to squeeze out, just enough to make it tacky.
I've done 10ish thermal pad swaps on 3080/90 gpus. I haven't seen core temp increase unless the density of the pad was too much or the size of the thermal pad was to thick. On some gpus I even got a core temp reduction because the stock thermal paste was trash Edit. I'd love to see you test the gelid gp 12w extreme, it's what I've had the best results with personally.
Hi I have a 3080 TI FE and only pad I can get are Minus pad 8 and Gelid Ultimate, From reddit I found that people are getting more success with 1.5mm pads on core side, 2mm on VRM and 2MM on back side now they all are using Gelid Extreme which sadly I cannot find in my country. Do you recommend Gelid Ultimate?
Oh man, what a good and great tested guide. The speaker has a calm voice and the Infos are very good. Gelid and nb are really the kings of the game. Great testing and overview. Thanks for the work. Just awesome 👌🏼👍🏼
Great video thanks for taking the time to test I have just re padded my 3090fe was hitting 104c on memory just gaming re padded using 1.5mm odyssey pads with great results max v ram temp now 80c I was so so nervous doing this tho but it's worked core temp has stayed the same max 69c and hotspot max 82c
I must say, well done! This is a great video with only the relevant information. Interesting about the raised core temps on the ones that lower the mem temps the most. Keep up the great content.
Yeah, it's crazy. This is not something we expected. Interesting how increasing chord temps come from the efficiency of heat transference from the memory module. Chips. Makes sense though, the cooler didn't get any bigger but it started taking more heat
you solved the Mistry that was bugging me. i was confused of the gpu core temp raising after replacing the thermal pads. and your take on that makes perfect sense. thank you.
I almost always use thermal grizzly, but this time I used the one that you recommended and it ended up being worse thermals then the original thermal pads. I personally didn’t have a good experience with those and will stick with the ones I know.
Thermal grizzly is a very good. They tend to leak more for some reason but your situation will vary. That's the thing too, everyone's mileage will vary for sure. Based on all the variables we can only go by consistent testing on her end.
Thanks for the video! Can I also suggest that maybe the core temps were higher, due to the better thermalpads being harder than the stock ones? Because hard pads would compress less, leading to less contact pressure between heatsink and gpu core, leading to higher gpu core temps.
Hey there, this would prove to be true. However, it seems to be happening across the board with other pads that were software as well that had a higher thermal conductivity. In other words, the lower the memory temperature, the higher the core temp in most situations. It's very interesting though! We thought so too. However, we don't think that was the case now.
This is a MUST SEEN video. Thanks very much folks for doing this. This is awesome. Clear, straight to the point and conclusion is perfect. Very pragmatic!
I haven't hanged any, but I have changed some :D I used some cheap ones and it worked fine but it was for gddr6 not gddr6x very excited to watch this. Might update my comment after I finish watching.... :D Edit: 2:15 I have some 5mm pads. 7:24 those the pads I used. Buying some NAB pads now. LOL
fun fact. i did this without recording and posting on UA-cam and my choice was gelid by far. granted i have my cards water-cooled but compared with cheap ones gelid keeps my 3080's and 3080ti's founders at 62-68 junction temp(while mining eth with +1400 mem) while the cheap one after 2 3 months they started loosing thermals. at first the cheap ones(well actually ek ones on one of the cards and different random brands on the other ones) were keeping the cards at 76 degree but than in a span of 4 months the temp started raising/averaging all the way up to 96 degree on some cards
gelid extreme 2.0mm @ 12.8 w/k on all 6x memory. 110pl on raven equates to 88c with roughly 80F ambient. copper plate and added fan on the card is ideal to remove heatsoak, but not required for hashing purposes.
Kudos! After viewing your video, I purchased quantity three of the mod/smart Fujipoly Ultra Extreme XR-m. I will mod my ASUS RTX3070-8G-EK, my RTX 3090 FE liquid cooled by both the EK-Quantum Vector FE RTX 3090 D-RGB BLK SE and the EK-Quantum Vector FE RTX 3090 D-RGB Active Backplate BLK SE, and my RTX 3070 Ti FE. Armed with your research, I discovered that the thermal pads supplied by EKWB are rated for thermal transferability of 3.5W/mK. Keep up the good energy!
Why dont i use shimis? why not use this pad? best place to let us know and FINDout WHAT ELSE we did in our discord! Tek Shinji Community Discord discord.gg/mjVHe2YMkR
You should try K5 Pro Viscous thermal compound it’s amazing thermal pad replacement it replaces pads up to 3mm thick I found it through LTT and his more FPS for $10 video
@@kdl0 umm yea but if you are using after market thermal pads then you probably shouldn't be using a regular cooler. They are most useful with water. The cooler will always be saturated when your just using a fan to blow air against a piece of metal. Air as a fluid is not very efficient at dissipating heat unlike water.
The best video I’ve seen in a looong, long time on comparing thermal pads/paste/OCing etc. FANTASTIC. Please give us more content like this! Subscribed!!!
Dude is super Charismatic, love the opening, helped me pick the perfect thermal pad. Thank you. Information: 10/10 Looks:10/10 Charisma: 10/10 Great content thanks brother
HOW did I miss this video for a year?! Great work! I have used the NAB cooling pads and the generic blue 6 W/M K pads. Glad to know they actually WORK. * If you ever decide to revisit this and review some more pads: Owl Tree makes Blue 6 W/M K and Grey 12.8 W/M K pads. They come in a 4-pack of 0.5, 1.0, 1.5, and 2.0mm thickness, 100x100mm in size. Arctic makes Pink and Blue pads, TP-1, TP-2, TP-3 and TP-4, in various thicknesses and sizes, up to 200x100mm. XPC and Kritikal make a Black thermal pad (carbon?) claiming 20 W/M K but it was hard to find in 1.5mm thickness.
Did your core temps increased while swapping thermal pads like in this video ? I find it hard to believe what they say in the video because when I swapped mine between different pads/brands on Gainward RTX3080ti I have never experienced core temperature difference of more than 1C initially and also it is just because of this video that I started swapping pads with very low w/mk -5 while being scared of increasing temps and when I did changed to 5W/mk I got VRAM temps increase to 20 C over stock thermal pads.So I'm still looking for as effective thermal pads as the stocks were ,ordered GP ultimate but have not tried them yet thinking they might be a bit too hard.
Amazing job ! Thank you for explaining that you locked the fan speed of the card. Very few people explain what they do with fan curve, so their temps results are pretty much useless
Very thorough and well done job. Thank you for taking a look at the systemic effects of changing the thermal pads on the memory and considering the system as a whole.
HEY, just to give you a quick update to my EVGA RTX 3070 XC3 BLACK thermal pad replacement situation. I contacted you before, the part where you show 17:00 mark, the temps actually increasing for 15w/mk + above pads. WELL, thats VERY TRUE. For my 3070, ordered GELID 15w, made my GPU run hotter. ESPECIALLY, for mining. Maybe it will be better for 3080+ above cards. But for my 3070, the 15w/mk made things worse. Eventually contacted EVGA, just got thermal pads for their deparment recently. Everything is working fine now.
@@TekShinji Like cutting the pads for GELID and installing it? It was easy, but not sure if it was making my heatsink contact better with my GPU DIE, even though it was 2mm, but i am sure it was the 15w/mk that made things worse, because i did order Thermalright 12w/mk and that made things a BIT better... But, right now for mining with EVGA PADS i am back to normal. So in the long run, wasn't worth the stress and hassle and wait to replacing the pads. The only thing i would've done was replaced the thermal paste. Since that was rock hard.
Re. Gelid GP-Ultimate - it's effectively pre-formed strips of thermal putty rather than pads as the latter are usually understood (and you can actually re-use it as putty if you ball it up!). People who've used thermal putty have reported the best overall temperatures by some margin over pads. That said, good putty is as or more expensive in terms of coverage than GP-Ultimate but _much_ fiddlier and more difficult to install (and the finished appearance is off-putting for those with OCD!). I now almost excusively use GP-Ultimate for VRM and memory. Mind-bogglingly thorough test though, bravo. ETA >> oops, 2 months and just noticed I forgot to hit 'like'!
Which one brand should i consider and how big the pad is sufficient to cover galax rtx 2060s and thinkness Reply will be appreciated ❤ I have watched entire without any skip the you have done so tiring So great of you 🐈
Honestly, any replacement pad will do. The biggest factor is your goal of GP or memory temperatures and the thickness you can find in your country/area. It doesn’t matter if you have the best pad in the world if it’s the wrong size
I'd be interested in seeing if you went up a size on the Thermalright Pads to make better contact due to the higher compression rate on the pads, how they would preform in comparison.
It is a GREAT vídeo, bringing light in an issue people talk too little about, and It's about time you guys do a follow-up one! I've been researching the subject in those past few weeks because I'm about to do the first clean-up and repasting of my new PC, so far everything is still stock. It is running great and I'm not having any problems, but I like to do the first repaste after one year and then just doing it again when necessary. Some suggestions: 1 can we replace thermal pad with a very thick layer of thermal paste? I think we can't because we would trade conduction for convection, and that's bad, but thermal paste, after it has curated, may end up behaving more like a solid, so, test it I say. 2 if we put thermal paste WITH thermal pads, what happens? I suggest doing it in 3 configurations, paste in both sides of the pad, paste in the dissipator side of the pad only and paste in the die side of the pad only, with 3 specs, matching the thermal conductivity of both the paste and the pad, with the paste being weaker, and with the pad being weaker. 3 is there a way to make it work? if we improve the heat dissipation to a degree heat gets blasted away before returning to the die?
Hey there, thanks for the comment. It’s kind of hard to follow up on all of that in one comment/UA-cam comment if you want to join our discord and join the conversation I can definitely talk to you more about there!
I really wish this reviews can come earlier, I got the thermalalright pad 2 weeks ago, they just barely improved my 3080FE mem temp for about 2 degree C over the stock pad. Thank you for reviewing them, I know what not to wast my money on next time.
We're glad that this video was helpful! We want to make our videos high quality and easy to understand as much as possible and that takes time. Sorry we couldn't make it sooner!
Well, I know people want a black and white answer but it's difficult to give. It really depends on what your goal is, do you want the lowest memory temperature? Or do you want the lowest GPU core temperature?
Great video. However one thing most keep forgetting is that lower temperatures also mean less power use = less power throttling. Resistance of most materials reduces with heat therefore requiring more amps at the same voltage.
I know this was made a year ago but this is quite an invaluable video for thermal pads in general- Amazing work! I suspected the core temp to increase if the pads are actually effective, but do you tlhink it will be that much of a difference on say a 3080 or 3080 ti which have much less vram chips on them? Thanks!
Hey there buddy, they technically don't have less. Because the 3090 non-TI are the ones that have the chips on the back. So there's overall higher thermal envelope. However another cards only have memory chips on one side.
Thank you so much for this video, truly appreciated not many reviewers will take the time and effort for this knowledge and accuracy 👏 you are the best I subscribed and like a millions times .
Really good comparison video, helps a lot with my upcoming repad project. NAB pads for sure. Good thing to note though, the conclusion with the gpu thermals being substantially increased due to the pad swap is likely specific to memory heavy applications, as we can see the gpu chip power draw is also about 60 W. In a core heavy workload, gaming for example, it seems likely that the benefit for the memory would lessen but the core temp would be less affected because power draw would be substantially higher on the chip. Interesting results though. I'm going to try a copper shim mod for my memory so I wonder what the results would be for that, compared to these.
This is the only video i was able to find on this topic. Thank you! I found Pullsar Ice Dragon Thermal Pad 17 W/m*k pads but cant find any comperison videos.
First time viewer, this is a pretty damn great video - I think you've got yourself another subscriber! Personally I've used Thermalright pads on a couple of 3090 FE cards after fitting EK blocks. In both cases, the temperatures dropped around 10c under memory load compared to the pads EK supply which is about where they should be given the stated conductivity... Since I'm working with blocks, the results aren't directly comparable with yours obviously, and I use liquid metal on the die and the (free and softer) EK pads where heat transfer isn't needed, like on top of the chokes, which would likely mitigate any compression issues the pinned commenter brought up... or maybe the 1mm pads are just made to tighter tolerances or weren't stamped on at the amazon warehouse... (they did measure at 1.0mm).
hey there Thanks for the feed back! water blocks thermal capacity is MUCH higher than standard air cooled so it wont be as much (or even at all a issue)
Mjb10 hello can I ask you a question? today I ordered the thermal right thermal pads precisely the thermal right Valor Odin 15w, I ordered them at 1mm for the memories, because ek waterblock declares that on the Vector ftw3 only the 1mm ones can go, I did well to order them right at 1mm and not more than? I hope they don't press too hard to reduce their thickness. thank you
Great video and info.... although I think something is going on with the thermalright pads. The first card I re-padded was the 3080 gig aorus master(w/LCD) The thing is like 3 bricks so I figured it would have great cooling but it would thermal about 30 seconds after power up... replaced ALL pads on card using 3 different sizes of Thermalright... temps dropped by 20c and card has not thermaled since.
Hey man great content, I've considered some of the thermal pads described in the vid and they would be for normal copper heatsinks for like a raspberry pi (in my case for a ps4 ram chips) my problem is I don't want the heatsink to fall off so I would need one quite sticky. I dont want to buy a thermal epoxy due to the low thermal conductivity and being permanent. Any suggestions on which would stick more effectively in a position like sideways or even upside-down? Lastly good gob on the vid, not much out there on the specific differences on the very different pads. Keep up with the good content😁👍
Very thorough and learned something at the end that I wouldn't have expected, in regards to the impact of transferring more heat from the VRAM to the core. I wonder if active Water-cooling would make a significant impact here, looking forward to your follow-up video on this.
yes, water cooling will be much better, i have Asus Strix 1070ti, on air i can get only +320mhz on mem with stock pads, after adding (aliexpress) pads to backplate, i could go to + 450mhz, with water cooling (EKWB and EKWB pads), no backplate, +600mhz, and these are full stable no restart in 2months of mining clocks.
Hi, absolutely fantastic video and flawless testing methodology with precise yet concise conclusions, thank you 😊. It would be awesome if you could redo the test with a few more brands included - Kritical Thermal Pads, MX-5, Alphacool Eisschicht Ultra Soft, Thermal Grizzly Carbonaut, Thermal Putty (TG-PP10).
@@TekShinji Sure, will do, and thanks for the reply 😊 I am especially curious to know how Carbonaut would perform against all other thermal Pads including your best contenders so far.
After so many annoying videos of "I tried new pads, and it worked better!" It is nice to see someone do a real side-by-side of different pads, and seeing what actually makes a difference and how much. Main issue with my 3090 hasn't been the main cooler, but the back plate. Even with a heat pipe under the back, you can still easily fry an egg on the back of the card! Took a block from an old Pentium II and just a little thermal paste (more to help it stick than as a permanent solution), and my reported VRAM temps dropped 10*c. That tells me that the VRAM under the main cooler is getting plenty of love from the fans, but the VRAM on the back is essentially the thermal bottleneck of the system. Whats more, after a month of mining at ~305W and 110-115MH/s my temps have risen significantly from 94-96*c back up to around 106*c which appears to be the thermal limit of my GPU where it will start slowing itself down a bit. From what I have read, this is expected for my card (Dell 3090), so now I am ordering some aftermarket pads... Hopefully it will work as expected and get me back up where my card belongs. In the mean time, lowering the power level to ~250W and getting closer to 90-100MH gets me back down to the more cozy 94-96*c again. The idea of being able to get closer to 90*c at full tilt... that would make me feel a heck of a lot better. I know on paper 120*c is where damage can occur, so the built-in 106*c is giving plenty of margin for error... but the lower I can get it below 100*c the better. Obviously while mining I am underclocking the GPU itself, so my GPU temps are a frosty 50*c, so no chance of danger there. It is only the RAM temps that causes me some worry.
Perfect! After looking at what my Dell 3090 needs I was able to find the NAB 15W pads for $8.50 for .5mm and $11 for 1mm. Ordered a 90x50cm sheet of each. The Gelid pads were $1 more expensive, but I'd still check as I would imagine a sale or price fluctuation may make one or the other cheaper. For those with similar Dell 3080 and 3090 cards: On the front of the card you will need 1mm pads between the RAM and the black thermal plate that covers the card. Between the heatsink and the black plate you can use thermal paste, but it needs to be thicker paste that wont melt/run/move over time. The other option here is to use .5mm pads instead which will compress down and will be more consistent over time. On the back, if on a 3080 just keep with the stock pads, they aren't doing much, and there is little to no point to replacing it (especially if it is sucking more heat out the front of the card more effectively). For a 3090, especially if mining, the RAM modules on back are likely your bottleneck. The built in pads look like they are pretty good quality... but just don't cover the whole chip, so not all of the heat is being dealt with. So you can either keep the stock back pads and add a .5mm pad to cover the area a little better (they are stackable and will squish enough to not be a problem), or get a 1.5mm pad to take care of the rest.
Hi there, thanks for the comment. You're asking if it's possible test them on a CPU or GPU since the thermal load is higher. Not exactly sure what your question is since the video is literally about testing it on GPUs? Technically, a CPU's thermal load on most computers is much lower than the GPU. Could you please restate your question?
As many people on Reddit say Gelid Extreme is slightly better than Ultimate (by ~2°C) since it's softer and provides a better contact with baseplates. Great video btw, finally a good comparison! :)
@@TekShinji Moreover I applied them (pads) on literally everything I could: mobo's VRM, RAM ICs, VRAM, GPU VRM and the GPU raditor itself ultra chill lol
Hey mate, nice video, i like the gp ultimate. What you think about the kriticalpads or the gpuriser pads with 20W/mk ? Is there a possibility for future testing ? Greetz
Physically speaking most of the pads of 90% to 95% of the most they will ever be unless technology is change or materials drastically change it’s not gonna go up by much. A lot of the readings and readings from people are by the companies enough by third-party so it’s very difficult to take those numbers at face value. Come join a discord to continue the conversation!
Well sure glad I only bought 2 packs of Thermalright pads instead of the whole range like I was going to. Thanks for the great info and saving our chips!
This is a great video to reference now after my 3090 FE needs maintenance after 3 1/4 years. Although there is something odd about your results with the Thermalright Oddysey pads, as those were the pads I used, and I got the same sort of temperatures you got with the Gelid Ultimate pads. My 3090 had the memory junction temperature at 112 degrees when new, and I ended up replacing the pads just two months after I bought it, which was very annoying and stressful. I used the 1.5mm thickness everywhere and got the memory junction temperature down to only get as high as 86 degrees under load when 3D rendering, which is also heavy on memory usage. I don't mine, so can't say how well it did for that, but it seems to be anything that is heavy on memory usage makes for good testing. At the moment, my 3090 is running a lot warmer than it used to on the core, so the thermal paste needs redoing, but the memory juction temperature is still the same as when I did it back then. I found this video as I'm concerned that even though the pads work alright still, they could end up crumbling if I try to remove them, so I think it's going to be safer for me to have some new ones ready to put in, and if the Thermalright ones can be inconsistent, I'll go with the better Gelid ones this time around. Thank you for such a good video.
Awesome! I have 3 cards (one which is a 3090 FE) I want to replace pads on and this is invaluable! I was going to go with Gelid, but may consider NAB based on this video. Thanks for doing this!
Although the GPU temp increases were very eye opening as well. It looks like the fans will remain at 100%. Fans are a cheap replacement vs. replacing the entire GPU. I'd really appreciate understanding if putting those little heatsinks on the backplate help anything.
Thank you very much for this very comprehensive test. For use on PC laptops for non-gaming everyday work and personal uses, which thermal pads would you suggest for the CPU, GPU, RAM and SSD, and what thickness?
What a great video, thank for posting this. Will you say the same rules apply to a 1080ti FE or Titan Xp? I want to change the pads and searching on the subject is how I landed here. I know the blower style is the worse as it is but will the core temps get higher with pads like 12W/mK? Maybe start with 6W/mK? any suggestions are welcome. Thanks and keep up the great videos!!!
Hey there, thanks for the comment and I'm glad you liked it. It would be dependent on what your goal is. Lower memory temps? Lower core temps? And really depends. Based on our data. You can pick the pad that works best for you!
Interesting result for the Thermalright pads, I have just redone all of my pads on a FE RTX 3090 withn those same 1.5mm pads and contact is perfect, my memory temps dropped from 100c max down to 80c max.
If I got it right, changing the pads only on gddr memory modules, you get lower temp on gddr, but higher temp on gpu core? So it's a mixed feeling situation? Or I got it wrong? Nice test and congratulations for the video
Essentially, according to our testing. The cooler was designed to dissipate X amount of heat energy per time. When the efficiency of heat and energy transferred per time increases, it also affects the performance of the cooler.
I've never personally worked on one of those myself. However, it appears that may be the case. Best idea to Google search yourself. This video is a comparison between the claims and the actual performance of the pads. Not a how to do video on a founder's edition. Since all pads tend to be slightly different depending on the card.
Hi!, do you think the NB supermax is going to last the same time in good condition compared to the gelid gp?, I don't know if the price difference means the nb supermax is worse.
Geat content, thanks for this! I have the 3090FE, I’m only gaming and cannot decide if I should go for a re-pad or not given the increased heat transfer to the core. The reason I am considering doing this is fan noise - clearly the memtemps are what’s makes the fans spool up on auto while gaming. What’s to expect here? Would a repad, let’s say using the gelid’s, make the core hotter while gaming and in the end not make any difference in fan noise/make it worse or should I expect a quieter card with better ”average combined” temps?
It would be dependent on your situation. However, lower memory temps can never hurt. If you're looking for the lowest noise, then your memory temps are going to be high. If you're looking for the lowest temperatures, then your noise is going to be higher. There's no way to get a free lunch. A lot of people like to go for the average or middle of the road. However, I did find more efficient heat transfer. Always turns out better in the long run
This is all very great info. I'm not changing pads, but it seems to be a reoccurring topic as of late. very interesting to see the results on core temp. i would love to see a follow up regarding different heat sinks.
17:22 you sayin the Founders heat sink around the entire board isn't good enough? Maybe the Ko will do better as a heatsink coming from Asus and all. Right?... Wait that's a good question. Is it better to use decent thermal pads on Asus cards?
It's about it's designed purpose. For memory intensive applications. It's not a good design. However, for 90% of gaming applications, it's fine since the memory doesn't get taxed as much. I don't understand the question though. Are you asking if it's good? I don't think it's bad.
Some Thermalright odyssey 1.5mm pads have issues with mounting pressure on the back side of the card I changed my 3090 FE pads to 1.5mm thermalrights, I never saw the reductions other people were reporting, Few months later got another pack this time 2mm, I changed the pads on the back memory to 2mm and then I saw the large temp drop, in my case 1.5mm was too thin for the back memory modules, though 1.5mm was correct for the front modules Great video, Wished it had been made a year earlier
One more comment.... I would love to see that video about adding thermal sinks to the backs of cards. People report some pretty significant improvements in memory temp with that but I've only been able to knock off 2 - 3 DegC personally. I have actually gone to removing the GPU backplates entirely so the backs of the cards can radiate the heat more effectively and that did better than attaching heatsinks to the backplate. If you do the backplate heatsink test you should give this a shot -- you may start the TekShinji backplateless wave with that idea. :)
We actually have a video similar to that already on the channel. It was earlier on in our history. However, it approved to be not as effective because the thermal pads were already bad to begin with. Now we know it's better to replace thermal pads other than putting heat sinks on only.
@@TekShinji You are exactly right. I have fans on both my rigs to keep air moving across the cards. My current thinking is that the ambient air temperature is way more important that I initially thought. I was brand new to most of this at the time so I watched and read a lot but didn't know what info was reliable/trustable so I wound up doing my own tests for the most part. Misfit Mining has been a great find for me personally. Now I have too much good information to process! :) Have a GREAAAAAAAAAAAAAAAAT day!
@@TekShinji Thanks for the quick reply! I should've been more clear. I mean pads of different thickness from the same manufacturer and the same lineup. As you mentioned, there's no data on the original pads and limited hardness or compression data for the aftermarket pads. It seems likely to me that the more conductive pads have less compressible filler material and more rigid conductive material, causing the GPU die to have less mounting pressure than it needs for better cooler. I think the impact of heat transferring from the VRAM into the GPU die is much less than mounting pressure and die coverage. It's impossible to conclude either with multiple variables changing simultaneously. Hopefully you could get a video out of it!
Thanks for this. Need thermal pads for some boards that are worth less than some of these pads, consoles, and laptops. Nice to see reviews that break the one soze fits all mold.
This video just proves that the Morpheus 2 is vastly much more better than what thought of initially. It's not even a competition anymore, the cooler king truthfully!
This video has grown so much. It’s difficult to keep up with comments after the first one. Please message me and our discord link in the description! Otherwise, I might not be able to get your comment to help you.
You simply did such an amazing job with this test and deserve every positive comment here :)
Also, I would like to add a small positive side of keeping memory cooler and core hotter :
Since metals (like copper/tin, etc.) get physically bigger the more heated they are, keeping PCB temperature even/uniform across all areas should help in preventing traces/joints being damaged from material fatigue over long period of time (because difference in temperature creates different stretching/compressing mechanical forces inside PCB layers).
I don't know how relevant this is to longevity of the card, hence I classified it only as small positive :)
It may not even be a problem for 5-10 years, since it depends on max. temp difference and number of hot/cold cycles each card had after all.
Thanks, I’m glad someone can appreciate our hard work! It took a long time and several hours to put this together, so we really appreciate it!
i commented not looking at the fact the video was 2 years old just take it with a grain of salt hehe.
Then the best is nab cooling?
Very, VERY few reviewers take the time to do such thorough testing. Bravo 👏 and thank you.
Thanks a lot, make sure to tell all your friends and get subscribed 🙂 a lot of people didn't like the videos so I appreciate it when you do!
@@TekShinji I really enjoyed this... (concerned what this says about me) lol. But the Thermalright pads are the ones i used on mine after seeing loads of videos and recommendations. Mine dropped from 104C to 88C although i did not just replace the pads on the memory modules i did the VRM pads too which i suppose would allow the heatsink to draw closer to the modules and thus create the proper contact which given the thickness deviation from what was advertised is probably quite important. Could you or did you test the thermalright pads but with more than the memory modules replaced? I and others have had excellent temp drops with them and sadly it seems your results are the anomaly which is distressing as yours seems to be the most concise testing i have seen.
And they who take that time seem to grow big.
You know videos like this are made for yt creators to grab money while sponsoring all pads XD
Next content,, battle of thermal paste pleassseeeee
For a non native english , your video is a must. You speack clearly with calm and I UNDERSTAND every word. Tank YOU. Great video and great speacker
Thank you, and you’re welcome! We worked really hard on this video to make sure it was easy to digest as possible.
Yeah, agreed. I would lose or change the music in the background though.
Thanks for the feed back!
@@TekShinji Thanks for the videos. Cheers
Dude this is awesome - thank you for not just stopping at which one performed best - seeing the core temps rise due to the increased sinking heat from the memory was something I had never considered
Thanks buddy im glad you enjoyed it! join our discord linked inthe discrption for more!
This kind of information is worth gold. Thank you for doing this (and it makes me feel better about getting the Gelid's in the first place). Comprehensive, easy to understand and to the point.
Also worth to note, using thicker pads means the contact pressure between the cold plate and GPU core is less, which usually translates into a few C increase on the GPU core. It's a tradeoff, cooler RAM means hotter core.
Hey there, thanks for the comment on the feedback. Regarding the thicker pads. This also happened with the thinner pads and every pad we used was relatively close to the tolerances listed. Because of that is why we determined there was more thermal saturation in the cooler.
Great job. Some of my favorite videos are ones by guys like you and Seb who put in a lot of hours like this to produce a single video. It does not go unnoticed, my friend. Adding all the graphs and numbers gave this one a real Gamer's Nexus feel and I know you must take any comparison to them as a high compliment. You deserve it.
As for others who may read this, I've used a generic Amazon pad similar to the one TekShinji does in this test. The kind you can get in the big 200mm x 200mm sheets which is HUGE (you can change many card's pads with one of those). They got the job done for me at a low price. They won't win you any bragging rights if that's what you're in to, but my Gigabyte 3080's been happily hashing away for 5 months with no loss in performance or degradation (knock on wood). My VRAM temps stay at 90C with 80% fan speed with pads replaced on both sides of the PCB. Nothing additional added to the metal backplate. No special fans directing additional air for supplemental cooling. Just one card inside a gaming PC case with the side panel off.
P.S. It's also worth noting, that in my situation, my 3080 required 3 different thicknesses to match what came on my card. As TekShinji mentions in this video, it's extremely important that you get the fit as close to perfect and not substitute the wrong pad sizes because you don't have what you need available. Because I was looking at 3 times the expense of someone doing a card that uses only 1 pad thickness, it led me to try the generic type as well.
Hey there buddy, thank you so much for those kind words. It really does mean a lot to us, this video took nearly two weeks to make and a lot of help from our wonderful resident scientist. Andrew. He really did almost all the work on this video. I just let him go ham!
We hope to make more videos like this in the future. It's just it's extreme time commitment and a lot of work.
Regarding the thickness of pads, that is true. Every card takes a different type of pad. That's why we are comparing the brands and their claims rather than making a how-to or which pad to use on which card since it varies so much.
Unbelievably comprehensive, informative video. I do need to repad my graphics card, but I'm a lot less worried about squeezing every degree of cooling out of it. You raised a good point about increasing core temps, I've never heard anyone else point that out before. Kudos
Glad you enjoyed the video! Make sure to get subscribe and hit that like button for the algorithm!
Yes, that’s why we concluded that the best pad is depending upon your application. Not necessarily the best thermal transference.
I can't thank you enough for making this video I was really struggling with deciding on a thermal pad brand.
Before watching your video I was honestly going to just dish out for the Fujipoly but thanks to your video i ended up with GELID.
1070ti is now sitting at
Now: 30.5MH/S @ 60c, 105W, Fan 60%, 290~300kH/W
Before: 30.5MH/s @70~75c but with maybe 200~250kH/W
Overall (thank to your video) I've noticed a SIGNIFICANT increase in CONSISTENCY.
The 1070ti (Founders Edition) required .5mm, 1mm, 1.5mm, and 2.0mm thermal pads.
I measured each and every pad to ensure they remained the same thickness.
I do have to warn though that the GELID thermal pads are very tough to reuse as they are like a clay substance that sorta crumbles after taking off / putting back on.
here there thanks for watching the vid! i hope things go well and good gaming!~
I think a contributing factor to the increased GPU core temp with certain pads, is that they are reducing the pressure applied to the core, leaving a thicker paste layer. I expect some temp increase with better VRAM cooling, but not 20c+
Would be interesting to see a high end thermal putty in comparison, since it allows for optimal core contact / pressure.
That would be interesting as well. Let us know how it goes for you!
I actually just got a 6700 XT that was a factory refurb with fresh paste and pads. The consistency / density of the pads appeared to be on par with Gelid Extreme, so pretty good quality to begin with. I tested with Furmark (1080p 8X) before and after switching to U6 Pro thermal putty and PTM7950 for the core.
It dropped -2c core, -10c hotspot, -4c VRAM and -5c VRM. Went from a hotspot delta of 18c, to 10c.
Ambient temp was identical between tests and I used a static fan speed. Let the cards heatsoak and reach steady state before recoding any data.
Very nice regarding your temperatures. Definitely in a good range.
It´s good to add, that you should always measure your old thermal pads and add new of the same height. Otherwise, you might encounter issues, like higher GPU core and/or VRM thermals, when using thicker pads. Or higher memory/VRM temperatures, when using thinner thermal pads.
Another necessary point is, each card can have multiple sizes of pads on various places, so you need to have a supply of multiple sizes of pads.
Thanks for the feedback!
@@TekShinji i was replacing the pads on my GPU last weekend and these were my findings. I had some trouble with measuring stock pads, so i used thicker pads for VRAM, as a result, i had insufficient contact at core and VRM. There is similar issue with backplate, especially when there aren’t enough screws near the padded area. It’s also necessary to understand, which area should be covered by the pads, since the manufacturers tends to use pads as simple & cheap as possible - they´d rather use 1 longer pad, than 2 or more individual pads for each VRAM chip.
Thermalright pads are NOT measured on the OO scale. (Sc) means they were measured using a shore C durometer which gives a very very different hardness rating. Although that rating can give you an idea of its hardness, it can not be compared directly with the others. The hardness of the Thermalright pads is noticeable higher than the others which is part of the reason they appear to give the most inconsistent results between users. They can not compress anywhere near as much as the other pads so they struggle to fill the gaps between the cold plate and memory modules. The main reason pads are used on memory modules is because their heights from the pcb are inconsistent due to the production line's soldering process.
I appreciate you pointing this out! I overlooked that it was SC. That would therefore be about 87-94.5 on the shore 00 scale.
i bought the 2.0mm thermalright odyssey in advance (failed to score a gpu yet), would they work better?
It depends on your application and the exact contact pressure required for your card. However, in our testing they do not appear to be made correctly so they automatically fail
So, in the end if there was a separate heatsink for the chips or better heatsink then the core temps of the gpu wouldn't have gone up?
It's likely, however, most GPUs share the heatsink these days. There's a few exceptions, yes, but a lot of them are together
This video was excellent. You managed to answer every question I've ever had about thermal pads in less than 20 minutes, thank you. Fun fact, I just picked up a MSI Ventus 3x OC 3090 from Best Buy and it has the grizzly thermal pads on the rear from the factory.
Thank you so much! I'm glad you enjoyed the video. We put a lot of work into this and it's nice when we see stuff like this.
You sure it has the actual grizzly brand? they're like a salmon / pink color doesn't necessarily mean it's from them.
Funny. I was coincidently recommending Gelid Extreme pads to a friend with many 3090's exactly 3 days ago lol. I personally use Fugipoly but I don't think it's always worth it to invest so much in 17w/mk pads for the so little difference it can "potentially" give. From personal experience, Gelid always been the ones that gave me the best overall results. Nice video by the way!
Exactly, from our testing. There doesn't seem to be much noticeable difference if any. Between the two thing that it's significantly cheaper, it doesn't make much sense.
The Gelid tested here is 'Ultimate' which is considerably more expensive than 'Extreme' (as in twice as expensive) and has much higher conductivity. The Gelid Extreme wouldn't match Fujipoly Ultra Extreme as Gelid Ultimate does.
@@EliteRock hmmm obviously?😂 what's your point exactly?
Actually, Gelid extreme is 12w/mk while Ultimate is 15w/mk. Fugipoly equivalents are 13w/mk and 17w/mk. Honestly, the difference between both extreme vs Gelid Ultimate doesn't justify the price gap between them. Only Fugi Ultimate is really better but still, 75$ for a single pad is overpriced for the little improve over the others it gives. Beside, pricing will differ from one location to another. And like Shinji have said, it all depends on the use you have for them since resistance will vary between most of them. I know Fugipoly extreme and Ultimate have the same resistance but I don't remember for the others..
@@danielchicoine79 Gelid Extreme is 8.5w/mK, not 12w/mK. I paid UK£20 for two 120x20x1.5mm Gelid Ultimates six months ago (for VRMs on a MB that are notoriously under-cooled) but they've dropped to c. £16 since then. Looking at prices here in the UK again, Ultimate now seems to be about 30% more than Extreme (for some reason sheets of Ultimate are bigger than Extreme at 90x50 vs 80x40) not twice as expensive as I stated.
@@EliteRock , odd - just checked the boxes for my Gelid GP-Extreme and they all clearly state 12w/mk. Am I missing something with them being 8.5w/mK somehow?
Brilliant work guys. You also conducted a fantastic analysis of the end-results instead of just posting stock temp data. Love it.
Thanks!
Hi, very nice and complete testing, however I can bring my results which defer a lot from yours .. both me and a friend got a 3080Ti FE. Both suffering from very high stock VRAM temps. (throttle 108-110° during heavy mining, fan immediately to 100%). We both did a repad at almost a similar time (end of August). I went for the Minus Pad 8 , he went for the Thermal Odyssey, both 1.5mm front and back (i tried 2mm first, that did not work, not enough Core contact). We had roughly the exact same temps during mining afterwards during gaming or mining (tried ETH, RVN, ERG), within the same 2°, both excellent, easily down to the 80° with about 60% fan. BUT, more recently I noticed that my temps were getting really bad again, to the point of being back to the 104-108°, at the same or even higher fan speeds (I was mining constantly since then). I took it apart again and ... it's like the Minus 8 pads "leaked", and had water under them now sitting between the ram module and the pad. (I have some pics). Cleaned up everything, and went for the same Thermalright Odyssey this time around. Back to the amazing 70°ish for gaming, 80°ish mining. So yeah, durability and ability to withstand during time is actually super important as well.
Short sum up from my side : Minus 8 Pad and Thermalright Odyssey perf is similar (approx -24° VRAM same conditions), but Minus 8 Pad degraded within 2 full months of constant mining. Let's see about the Odyssey
Thanks for the info 🙂😊🙂!!
Jaques dewever hello can I ask you a question? today I ordered the thermal right thermal pads precisely the thermal right Valor Odin 15w, I ordered them at 1mm for the memories, because ek waterblock declares that on the Vector ftw3 only the 1mm ones can go, I did well to order them right at 1mm and not more than? I hope they don't press too hard to reduce their thickness. thank you
*WARNING* I use water cooling on 3080 and went from 68 C to 96 C in 8 months on Gelid which is tested in the roundup. All silicone based thermal pads are leaking silicone and sooner or later will become junk. ~100 % of pads in consumer market are silicone based but Google it - non-silicone based thermal pads do exist. You will not see claims like 13-17 W/mK but if you do not need to replace them every 6-8 months it is fine I guess.
How is it holding up now?
@@BlackPhilosopher nice, I was wondering about durability as well. Despite the good results on the three most right, there was no way for them to show durability. Thanks for sharing!
GeLid is the best!
Did you have a lot of good experience with them?
@@TekShinji yes very good temp difference after using them
Im glad to hear!
which gelid? the ultimate or extreme?
@@AriR6Rboth are good
I just take a university class or what. Was so compact and informative yet still easy to digest. Takes me a while to find peeps I can comfortably listen to but I think I just found one. Thx for all the knowledge!!
Thanks! I’m glad u got good info out of it!
Try Gelid Extreme instead. They're much softer compared to the Gelid Ultimate. One possible flaw with this test is the harder pads do not compress as much, therefore leading to insufficient core mounting pressure.
Another idea is to try thermal putty (TG-PP10). They don't have any hardness, so they can potentially allow the heatsink to be closer to the modules/core allowing for higher clamping and thermal transfer.
Thanks for the information
Hello, does extreme perform worse than ultimate or is it better?
@@dolanthefacc879 did you find out
Really impressive, professional video. Heck, it’s better than professional. Very thorough breakdown of the stats in an easy to understand way for the lay person. I love the narrator’s calm, relaxing voice too!
We only have the best narrators! Yes, Andrew did a lot of work on this video. There's nearly two weeks of work in this video. We're glad that it was easy to digest and easy to understand!
you are absolutely correct in your analysis: evacuate the heat from the heatsink as well to avoid increased core temps. I would like to see a full copper sheet used vs the more local "raspberry pi" heatsinks.
Yeah, it's so crazy how this happens.
@@TekShinji check out ebay if you guys are interested. The chinese have the concept for sell at different thicknesses and tape applied. I went with scrap 5/16 copper tubing, but for testing & videos the plates are ideal.
The thing is we also have to keep the pads within the scope class realm of availability in practicality. We can't test something that no one can get.
@@TekShinji im referring to copper plate in this comment. They are sold in different thicknesses on ebay. Send me a friend request on discord and ill send you a link. Again this isn’t necessary for what we do with regards to mining but it would extend the life of the component by removing heatsoak.
Thanks for the information!
I've repadded a lot of 3090 FE , I've had no problems with the thermalright extreme odyssey , they measure as 1.5mm as advertised using a micrometer . My benchmarking involves running the card at 105C max on the hotspot .On a stock 3090 this usually involves underclocking the ram significantly and usually gets something like 80Mhash . With the thermaltakes and a repaste I'll get 120Mhash with a very high overclock and temps will usually stay at 98 or below with fan sitting at 60% ( which means the card can be pretty much silent when mining).
I've used Gelids as a comparison and found zero difference in performance. I use gelid and thermaltake on Zotac and Msi and not noticed much difference between the two .
Only difference being going for 0.5mm thicker on the gelids vs stock pads due to softness . I generally prefer to use the thermalright and take care to precisely measure the pads with a micrometer ( its all 1.5mm on the 3090 FE) on most 3080 and 3090 other than FE you only need to replace the memory pads front and back , the other components are usually fine with the stock pads whihc is just as well as they often use odd thicknesses like 1.3 and 1.7mm ( its totally fine to sandwich pads together to get the thickness required by the way)
Awesome, thank you for the feedback!
I refurbish gaming consoles and have been using Gelid for a while now. They are affordable and result in a huge noise reduction, especially on PS4 Pro. It's good to see some numbers backing up my experience with them. Thx for the in-depth video.
Thanks buddy, I'm glad you like the video. Make sure to check out others on our channel and get subscribed :-)
Hey, I'm a little new to thermal pads, but I was wondering: can you use these Gelid thermal pads inside a laptop? There's this separate chipset inside my old laptop that the heat pipe also goes to, and there's a little thermal pad on the chipset.
I've measured the dimensions of the thermal pad and the new pad should be a 15mm square that is 1mm thick. I needed thermal paste for the new CPU and figured I'd replace the little thermal pad on the other chipset too while I had it apart.
TL;DR: Can you use these Gelid thermal pads on chipsets inside an older laptop?
@@neverendinchaos4800 age of the system is irrelevant, only the size and thermal dissipation. yes, you should replace it.
I would agree with what the other responder said!
Good vid...though a little misleading.
I run 2 3090 FE's in one of my mining rigs and use Thermalright. They sit at 74 and 78 degrees (virtual memory temp) at my house.
You have to get the thickness perfect. I think there are more than just 1.5mm pads that get used in this GPU.
I wasted the money and bought all sizes .5mm - 3.0mm and then just matched it to the compressed thickness of the stock pad that was used.
I had to remount a couple times to figure it out, but now it's perfect.
It seems obvious with that poor of a result on a quality product that there's an issue you may want to retest.
Just some things to consider!
You're definitely right about there being a lack of content, especially for the rare 3090fe.
Thanks for making it.
Hey there buddy, thank you for the feedback! Will definitely consider this.
Same here. For my 3080 I used thermalright 2mm and at first my mem temp went from 106C to 88C but my gpu die increased from 48C to 68C. I then took the card out and really pressed very hard on the locations where the pads were and also on the center of the cooler where the gpu die is. Now Mining 48C gpu die and 88C memory at 102Mhs.
Thermalright 1.5mm for gpu memory on my watercooled Aorus Xtreme 3090 with Bykski Waterblock. They're definitely firmer than other pads, but 1.5mm is perfect. 76C max vram temps on a mining test with a higher vram memory overclock.
for me, in 3080 82 degrees
"rare" thanks also to people like you. No offense, business is business ...
Thank the lord you made this video. I've spent hundreds and hundreds buying thermal pads. Then I finally thought, "Do I need to keep buying these expensive pads from Amazon?" I then go to AliExpress and see literally the same pads you found. I then prayed I could find a video of someone testing this before I dropped money and time doing it myself haha! THANK YOU!
Thanks buddy, I'm glad you really like the video! Make sure to hit that like button get subscribed hit that Bell icon to help us out!
Wow, I love how comprehensive this is. I really appreciate all the work you put into this video. It's hard to wade through all the threads on Thermal Grizzly Minus, Gelid and Fujipoly, having a video to reference cost vs effectiveness is awesome! I think I'd go with the Gelid pads, seems like the perfect middle ground. Do you have any experience with the TG-PP-10 putty? The EVGA 3080/90/Ti stock cooler uses custom sized pads and their own putty, trying to find a good alternative to bring down VRAM/Junction temps.
I want to check that stuff out too. Haven't tried it yet myself though!
Thank you for the kind words. A lot of work went into this so we appreciate it:-)
@@TekShinji I can imagine! I haven't done anything this involved yet, but definitely appreciate the work.
Yea it was like 20+ hours of work. Lol. It's best to come join our discord if ya want to chat and such! Let me know if you have any questions?
I love the methodologically sound and accurate approach to presenting data. Keeping as many variables stable/unchanged was a moment I started nerding out, so good to see reviewers take a scientific approach fully and seriously!
Glad you enjoyed it!
That's weird, I would have thought that reducing memory temperatures and so reducing temperatures on the PCB and so neighbouring components would lead to decrease of the core temperatures, but it's opposite, weird...
Note to self: Best choice pads are Gelid GP Ultimate 15W or same under diffrent branding NAB Cooling NB Supermax 15W
YUP we has NO idea this would be the result it truly is amazing because NO one has ever talked about this!
Dang it just googled my 3090 trio uses THREE DIFFERENT THICKNESSES!
Oh man! That sux! What were they thinking!?!?!
note to you 😁:I can confirm after 4m of trial, gelid are very solid pads(i dare to say the best thermal pads. just make sure to get the right thickness). they are very good if you go with whatever the manufacturer had in there too if you add thicker they won't compress and you will get inconsistent/poor results
I've used both gelud 15w and 12w, they have different uses. If you need a firm pad use the 15w, if you need a soft pad is 12w.
I know it's gonna sound crazy, but an ultra thin layer of thermal compound on each pad before making contact gave me INCREDIBLE results versus just thermal pads direct to the surface.I mean the smallest amt of paste you can use, super spread out thin with no extra to squeeze out, just enough to make it tacky.
Thanks for the feedback! Is it possible to ask for the exact procedure in details you used?
@djlowtek did you thermal paste the pads on just one side (the memory side?), or both the memory side and the backplate side of the pads?
I've done 10ish thermal pad swaps on 3080/90 gpus. I haven't seen core temp increase unless the density of the pad was too much or the size of the thermal pad was to thick. On some gpus I even got a core temp reduction because the stock thermal paste was trash
Edit. I'd love to see you test the gelid gp 12w extreme, it's what I've had the best results with personally.
Thanks for the feedback. Appreciate it!
Hi I have a 3080 TI FE and only pad I can get are Minus pad 8 and Gelid Ultimate, From reddit I found that people are getting more success with 1.5mm pads on core side, 2mm on VRM and 2MM on back side now they all are using Gelid Extreme which sadly I cannot find in my country. Do you recommend Gelid Ultimate?
I BELIEVE YOU because I got similar results but with fewer tries and exploration so thank your for your input!
Oh man, what a good and great tested guide. The speaker has a calm voice and the Infos are very good. Gelid and nb are really the kings of the game. Great testing and overview. Thanks for the work. Just awesome 👌🏼👍🏼
Yea for the price you can't beat it!!
Great video thanks for taking the time to test I have just re padded my 3090fe was hitting 104c on memory just gaming re padded using 1.5mm odyssey pads with great results max v ram temp now 80c I was so so nervous doing this tho but it's worked core temp has stayed the same max 69c and hotspot max 82c
Thanks buddy appreciate it!
Thank you for reviewing all these. I went with a Gelid extremes for my 3090 FE, and future 3090 FE cards.
You're welcome! Let others know and get subscribed and see us in the discord!
I must say, well done! This is a great video with only the relevant information. Interesting about the raised core temps on the ones that lower the mem temps the most. Keep up the great content.
Yeah, it's crazy. This is not something we expected. Interesting how increasing chord temps come from the efficiency of heat transference from the memory module. Chips. Makes sense though, the cooler didn't get any bigger but it started taking more heat
Yeah, really interesting result. But most logical, as Spock would say.
Long live Spock!
you solved the Mistry that was bugging me. i was confused of the gpu core temp raising after replacing the thermal pads. and your take on that makes perfect sense. thank you.
Absolutely thank you so much for watching the video. We really appreciate it. It was a big surprise to us when we figure that out.
I almost always use thermal grizzly, but this time I used the one that you recommended and it ended up being worse thermals then the original thermal pads. I personally didn’t have a good experience with those and will stick with the ones I know.
Thermal grizzly is a very good. They tend to leak more for some reason but your situation will vary.
That's the thing too, everyone's mileage will vary for sure. Based on all the variables we can only go by consistent testing on her end.
To give credit where it is due.
I only just noticed the your way of cutting the pads is actualy quite good.
I will try this.
Thx
Thanks Andrew is AMAZING
Thanks for the video! Can I also suggest that maybe the core temps were higher, due to the better thermalpads being harder than the stock ones?
Because hard pads would compress less, leading to less contact pressure between heatsink and gpu core, leading to higher gpu core temps.
Hey there, this would prove to be true. However, it seems to be happening across the board with other pads that were software as well that had a higher thermal conductivity. In other words, the lower the memory temperature, the higher the core temp in most situations. It's very interesting though! We thought so too. However, we don't think that was the case now.
This is a MUST SEEN video.
Thanks very much folks for doing this. This is awesome.
Clear, straight to the point and conclusion is perfect.
Very pragmatic!
Thank you, we put a lot of hard work and effort into it so we really appreciate it :-)
Have you changed Pads? What pads do you LIKE?
I haven't hanged any, but I have changed some :D I used some cheap ones and it worked fine but it was for gddr6 not gddr6x very excited to watch this. Might update my comment after I finish watching.... :D Edit: 2:15 I have some 5mm pads. 7:24 those the pads I used. Buying some NAB pads now. LOL
fun fact. i did this without recording and posting on UA-cam and my choice was gelid by far. granted i have my cards water-cooled but compared with cheap ones gelid keeps my 3080's and 3080ti's founders at 62-68 junction temp(while mining eth with +1400 mem) while the cheap one after 2 3 months they started loosing thermals. at first the cheap ones(well actually ek ones on one of the cards and different random brands on the other ones) were keeping the cards at 76 degree but than in a span of 4 months the temp started raising/averaging all the way up to 96 degree on some cards
Exactly that is something that would be much harder to test but thanks for the info. We will definitely consider this!
I hope your temps be low!
I changed the pads on my Asus tuf 3090 to gelid ultimate and it made no difference at all my memory is still at 90c+
gelid extreme 2.0mm @ 12.8 w/k on all 6x memory. 110pl on raven equates to 88c with roughly 80F ambient. copper plate and added fan on the card is ideal to remove heatsoak, but not required for hashing purposes.
Kudos! After viewing your video, I purchased quantity three of the mod/smart Fujipoly Ultra Extreme XR-m. I will mod my ASUS RTX3070-8G-EK, my RTX 3090 FE liquid cooled by both the EK-Quantum Vector FE RTX 3090 D-RGB BLK SE and the EK-Quantum Vector FE RTX 3090 D-RGB Active Backplate BLK SE, and my RTX 3070 Ti FE. Armed with your research, I discovered that the thermal pads supplied by EKWB are rated for thermal transferability of 3.5W/mK. Keep up the good energy!
That's awesome!
Why dont i use shimis? why not use this pad? best place to let us know and FINDout WHAT ELSE we did in our discord!
Tek Shinji Community Discord
discord.gg/mjVHe2YMkR
You should try K5 Pro Viscous thermal compound it’s amazing thermal pad replacement it replaces pads up to 3mm thick I found it through LTT and his more FPS for $10 video
So you bought all these pads but didnt buy alphacool or use k5 pro? which are the best. kinda defeats point of this video
1
@@AcidBombYT If they are better, it would exacerbate the problem highlighted at the end of the video regarding a saturated cooler heating up the gpu
@@kdl0 umm yea but if you are using after market thermal pads then you probably shouldn't be using a regular cooler. They are most useful with water. The cooler will always be saturated when your just using a fan to blow air against a piece of metal. Air as a fluid is not very efficient at dissipating heat unlike water.
@@shorty808100 I wish $10 got you more fps.
The best video I’ve seen in a looong, long time on comparing thermal pads/paste/OCing etc. FANTASTIC. Please give us more content like this! Subscribed!!!
Things a lot buddy. I appreciate it! Let all your friends know And if you like join our discord we will continue to discussion!
Dude is super Charismatic, love the opening, helped me pick the perfect thermal pad. Thank you.
Information: 10/10
Looks:10/10
Charisma: 10/10
Great content thanks brother
You’re welcome. I’m glad that our Video helped you out!
Thanks a lot. really great video.
I'd like to see a test concerning 18:25 after being repadded
Hey there buddy, did you watch the whole video? There should be a whole bunch of data there. You can always continue the conversation in our discord!
HOW did I miss this video for a year?! Great work! I have used the NAB cooling pads and the generic blue 6 W/M K pads. Glad to know they actually WORK.
* If you ever decide to revisit this and review some more pads:
Owl Tree makes Blue 6 W/M K and Grey 12.8 W/M K pads. They come in a 4-pack of 0.5, 1.0, 1.5, and 2.0mm thickness, 100x100mm in size.
Arctic makes Pink and Blue pads, TP-1, TP-2, TP-3 and TP-4, in various thicknesses and sizes, up to 200x100mm.
XPC and Kritikal make a Black thermal pad (carbon?) claiming 20 W/M K but it was hard to find in 1.5mm thickness.
You didn’t miss it, the algorithm chose not to show you for some reason ha, ha. That works, though, let me know how it goes for you!
Did your core temps increased while swapping thermal pads like in this video ? I find it hard to believe what they say in the video because when I swapped mine between different pads/brands on Gainward RTX3080ti I have never experienced core temperature difference of more than 1C initially and also it is just because of this video that I started swapping pads with very low w/mk -5 while being scared of increasing temps and when I did changed to 5W/mk I got VRAM temps increase to 20 C over stock thermal pads.So I'm still looking for as effective thermal pads as the stocks were ,ordered GP ultimate but have not tried them yet thinking they might be a bit too hard.
Are you asking whether we did more than one test? I’m not exactly sure.
Amazing job ! Thank you for explaining that you locked the fan speed of the card. Very few people explain what they do with fan curve, so their temps results are pretty much useless
Thanks appreciate it, make sure to get subbed and tell everyone about it!
Very thorough and well done job. Thank you for taking a look at the systemic effects of changing the thermal pads on the memory and considering the system as a whole.
You're welcome buddy, thank you for the comment and make sure to tell all your friends about it!
HEY, just to give you a quick update to my EVGA RTX 3070 XC3 BLACK thermal pad replacement situation. I contacted you before, the part where you show 17:00 mark, the temps actually increasing for 15w/mk + above pads. WELL, thats VERY TRUE. For my 3070, ordered GELID 15w, made my GPU run hotter. ESPECIALLY, for mining. Maybe it will be better for 3080+ above cards. But for my 3070, the 15w/mk made things worse.
Eventually contacted EVGA, just got thermal pads for their deparment recently. Everything is working fine now.
Oh wow that's crazy! Did you have any issues with the install?
@@TekShinji Like cutting the pads for GELID and installing it? It was easy, but not sure if it was making my heatsink contact better with my GPU DIE, even though it was 2mm, but i am sure it was the 15w/mk that made things worse, because i did order Thermalright 12w/mk and that made things a BIT better... But, right now for mining with EVGA PADS i am back to normal. So in the long run, wasn't worth the stress and hassle and wait to replacing the pads. The only thing i would've done was replaced the thermal paste. Since that was rock hard.
Re. Gelid GP-Ultimate - it's effectively pre-formed strips of thermal putty rather than pads as the latter are usually understood (and you can actually re-use it as putty if you ball it up!). People who've used thermal putty have reported the best overall temperatures by some margin over pads. That said, good putty is as or more expensive in terms of coverage than GP-Ultimate but _much_ fiddlier and more difficult to install (and the finished appearance is off-putting for those with OCD!). I now almost excusively use GP-Ultimate for VRM and memory. Mind-bogglingly thorough test though, bravo. ETA >> oops, 2 months and just noticed I forgot to hit 'like'!
Thanks for the feed back!
Which one brand should i consider and how big the pad is sufficient to cover galax rtx 2060s and thinkness
Reply will be appreciated ❤
I have watched entire without any skip the you have done so tiring
So great of you 🐈
Honestly, any replacement pad will do. The biggest factor is your goal of GP or memory temperatures and the thickness you can find in your country/area.
It doesn’t matter if you have the best pad in the world if it’s the wrong size
I'd be interested in seeing if you went up a size on the Thermalright Pads to make better contact due to the higher compression rate on the pads, how they would preform in comparison.
Thanks for the feedback!
Good comprehensive video. It must have taken a lot of time to do all the tests and I thank you for your time.
Thanks buddy, if you want to help us out, let all your friends know about it. Sure the word. Let them see this video!
@@TekShinji shared.
@@BlenderRookie thanks!!! ♥️♥️♥️♥️♥️
It is a GREAT vídeo, bringing light in an issue people talk too little about, and It's about time you guys do a follow-up one!
I've been researching the subject in those past few weeks because I'm about to do the first clean-up and repasting of my new PC, so far everything is still stock. It is running great and I'm not having any problems, but I like to do the first repaste after one year and then just doing it again when necessary.
Some suggestions:
1 can we replace thermal pad with a very thick layer of thermal paste? I think we can't because we would trade conduction for convection, and that's bad, but thermal paste, after it has curated, may end up behaving more like a solid, so, test it I say.
2 if we put thermal paste WITH thermal pads, what happens? I suggest doing it in 3 configurations, paste in both sides of the pad, paste in the dissipator side of the pad only and paste in the die side of the pad only, with 3 specs, matching the thermal conductivity of both the paste and the pad, with the paste being weaker, and with the pad being weaker.
3 is there a way to make it work? if we improve the heat dissipation to a degree heat gets blasted away before returning to the die?
Hey there, thanks for the comment. It’s kind of hard to follow up on all of that in one comment/UA-cam comment if you want to join our discord and join the conversation I can definitely talk to you more about there!
I really wish this reviews can come earlier, I got the thermalalright pad 2 weeks ago, they just barely improved my 3080FE mem temp for about 2 degree C over the stock pad. Thank you for reviewing them, I know what not to wast my money on next time.
We're glad that this video was helpful! We want to make our videos high quality and easy to understand as much as possible and that takes time. Sorry we couldn't make it sooner!
so which one exactly r u getting
Well, I know people want a black and white answer but it's difficult to give. It really depends on what your goal is, do you want the lowest memory temperature? Or do you want the lowest GPU core temperature?
Great video. However one thing most keep forgetting is that lower temperatures also mean less power use = less power throttling. Resistance of most materials reduces with heat therefore requiring more amps at the same voltage.
that is absoluty correct thanks for the comment! make sure to get subbed for furtre vids!
Andrew crushed it on this video. Very well done and explained.
Andrew did work hard on this video! He deserves the credit!
I know this was made a year ago but this is quite an invaluable video for thermal pads in general- Amazing work!
I suspected the core temp to increase if the pads are actually effective, but do you tlhink it will be that much of a difference on say a 3080 or 3080 ti which have much less vram chips on them? Thanks!
Hey there buddy, they technically don't have less. Because the 3090 non-TI are the ones that have the chips on the back. So there's overall higher thermal envelope. However another cards only have memory chips on one side.
Thank you so much for this video, truly appreciated not many reviewers will take the time and effort for this knowledge and accuracy 👏 you are the best I subscribed and like a millions times .
You're very welcome!
Really good comparison video, helps a lot with my upcoming repad project. NAB pads for sure.
Good thing to note though, the conclusion with the gpu thermals being substantially increased due to the pad swap is likely specific to memory heavy applications, as we can see the gpu chip power draw is also about 60 W. In a core heavy workload, gaming for example, it seems likely that the benefit for the memory would lessen but the core temp would be less affected because power draw would be substantially higher on the chip. Interesting results though. I'm going to try a copper shim mod for my memory so I wonder what the results would be for that, compared to these.
Hey there thanks for watching the video. Must know what you find out in our discord!
So much job in this video, very nice!
Do you have any suggest for a 3060Ti Tuf? i'm hitting 105 in hotspot
Thanks buddy. It's hard to talk about things like this over comments. Try joining our discord for a continued conversation linked In the description!
I've never been here before but GOT DAYUM your hair is majestic bro
Lol thanks man, check my other vids out !
This is the only video i was able to find on this topic. Thank you! I found Pullsar Ice Dragon Thermal Pad 17 W/m*k pads but cant find any comperison videos.
Thank you! It was not easy to make this video. Make sure to check out our other videos and get subscribed 🙂
First time viewer, this is a pretty damn great video - I think you've got yourself another subscriber! Personally I've used Thermalright pads on a couple of 3090 FE cards after fitting EK blocks. In both cases, the temperatures dropped around 10c under memory load compared to the pads EK supply which is about where they should be given the stated conductivity... Since I'm working with blocks, the results aren't directly comparable with yours obviously, and I use liquid metal on the die and the (free and softer) EK pads where heat transfer isn't needed, like on top of the chokes, which would likely mitigate any compression issues the pinned commenter brought up... or maybe the 1mm pads are just made to tighter tolerances or weren't stamped on at the amazon warehouse... (they did measure at 1.0mm).
hey there Thanks for the feed back! water blocks thermal capacity is MUCH higher than standard air cooled so it wont be as much (or even at all a issue)
Mjb10 hello can I ask you a question? today I ordered the thermal right thermal pads precisely the thermal right Valor Odin 15w, I ordered them at 1mm for the memories, because ek waterblock declares that on the Vector ftw3 only the 1mm ones can go, I did well to order them right at 1mm and not more than? I hope they don't press too hard to reduce their thickness. thank you
Great video and info.... although I think something is going on with the thermalright pads. The first card I re-padded was the 3080 gig aorus master(w/LCD) The thing is like 3 bricks so I figured it would have great cooling but it would thermal about 30 seconds after power up... replaced ALL pads on card using 3 different sizes of Thermalright... temps dropped by 20c and card has not thermaled since.
Very nice man, good to hear. There's many factors when it comes to thermal and cards.
Hey man great content, I've considered some of the thermal pads described in the vid and they would be for normal copper heatsinks for like a raspberry pi (in my case for a ps4 ram chips) my problem is I don't want the heatsink to fall off so I would need one quite sticky. I dont want to buy a thermal epoxy due to the low thermal conductivity and being permanent. Any suggestions on which would stick more effectively in a position like sideways or even upside-down? Lastly good gob on the vid, not much out there on the specific differences on the very different pads. Keep up with the good content😁👍
Very thorough and learned something at the end that I wouldn't have expected, in regards to the impact of transferring more heat from the VRAM to the core. I wonder if active Water-cooling would make a significant impact here, looking forward to your follow-up video on this.
Yeah man, we had no idea going into it as well. It was a lot of information we learned:-)
yes, water cooling will be much better, i have Asus Strix 1070ti, on air i can get only +320mhz on mem with stock pads, after adding (aliexpress) pads to backplate, i could go to + 450mhz, with water cooling (EKWB and EKWB pads), no backplate, +600mhz, and these are full stable no restart in 2months of mining clocks.
Very nice man!
Hi, absolutely fantastic video and flawless testing methodology with precise yet concise conclusions, thank you 😊. It would be awesome if you could redo the test with a few more brands included - Kritical Thermal Pads, MX-5, Alphacool Eisschicht Ultra Soft, Thermal Grizzly Carbonaut, Thermal Putty (TG-PP10).
Thanks for the recommendations! Make sure to share it and let all your friends know :-)
@@TekShinji Sure, will do, and thanks for the reply 😊 I am especially curious to know how Carbonaut would perform against all other thermal Pads including your best contenders so far.
Thanks!! 😘
Dude thanks you so much now I know which to choose when my thermal pads eventually become bad.
Awesome thanks for checking out or video!
This was a really helpful video. Very well researched and good job on the video production as well. Keep up the good work.
Thanks!
After so many annoying videos of "I tried new pads, and it worked better!" It is nice to see someone do a real side-by-side of different pads, and seeing what actually makes a difference and how much.
Main issue with my 3090 hasn't been the main cooler, but the back plate. Even with a heat pipe under the back, you can still easily fry an egg on the back of the card! Took a block from an old Pentium II and just a little thermal paste (more to help it stick than as a permanent solution), and my reported VRAM temps dropped 10*c. That tells me that the VRAM under the main cooler is getting plenty of love from the fans, but the VRAM on the back is essentially the thermal bottleneck of the system.
Whats more, after a month of mining at ~305W and 110-115MH/s my temps have risen significantly from 94-96*c back up to around 106*c which appears to be the thermal limit of my GPU where it will start slowing itself down a bit. From what I have read, this is expected for my card (Dell 3090), so now I am ordering some aftermarket pads... Hopefully it will work as expected and get me back up where my card belongs. In the mean time, lowering the power level to ~250W and getting closer to 90-100MH gets me back down to the more cozy 94-96*c again. The idea of being able to get closer to 90*c at full tilt... that would make me feel a heck of a lot better. I know on paper 120*c is where damage can occur, so the built-in 106*c is giving plenty of margin for error... but the lower I can get it below 100*c the better.
Obviously while mining I am underclocking the GPU itself, so my GPU temps are a frosty 50*c, so no chance of danger there. It is only the RAM temps that causes me some worry.
Perfect! After looking at what my Dell 3090 needs I was able to find the NAB 15W pads for $8.50 for .5mm and $11 for 1mm. Ordered a 90x50cm sheet of each. The Gelid pads were $1 more expensive, but I'd still check as I would imagine a sale or price fluctuation may make one or the other cheaper.
For those with similar Dell 3080 and 3090 cards:
On the front of the card you will need 1mm pads between the RAM and the black thermal plate that covers the card.
Between the heatsink and the black plate you can use thermal paste, but it needs to be thicker paste that wont melt/run/move over time. The other option here is to use .5mm pads instead which will compress down and will be more consistent over time.
On the back, if on a 3080 just keep with the stock pads, they aren't doing much, and there is little to no point to replacing it (especially if it is sucking more heat out the front of the card more effectively). For a 3090, especially if mining, the RAM modules on back are likely your bottleneck. The built in pads look like they are pretty good quality... but just don't cover the whole chip, so not all of the heat is being dealt with. So you can either keep the stock back pads and add a .5mm pad to cover the area a little better (they are stackable and will squish enough to not be a problem), or get a 1.5mm pad to take care of the rest.
That's awesome, Thanks for the comment!
Very good review! Is it possible to test those (and maybe others) on a cpu or gpu, since the thermal load is higher?
Hi there, thanks for the comment. You're asking if it's possible test them on a CPU or GPU since the thermal load is higher.
Not exactly sure what your question is since the video is literally about testing it on GPUs? Technically, a CPU's thermal load on most computers is much lower than the GPU. Could you please restate your question?
As many people on Reddit say Gelid Extreme is slightly better than Ultimate (by ~2°C) since it's softer and provides a better contact with baseplates.
Great video btw, finally a good comparison! :)
Thanks! Is that wat u endure using
@@TekShinji Moreover I applied them (pads) on literally everything I could: mobo's VRM, RAM ICs, VRAM, GPU VRM and the GPU raditor itself
ultra chill lol
Very nice! I’m glad it’s working :-)
Hey mate, nice video, i like the gp ultimate. What you think about the kriticalpads or the gpuriser pads with 20W/mk ? Is there a possibility for future testing ?
Greetz
Physically speaking most of the pads of 90% to 95% of the most they will ever be unless technology is change or materials drastically change it’s not gonna go up by much. A lot of the readings and readings from people are by the companies enough by third-party so it’s very difficult to take those numbers at face value. Come join a discord to continue the conversation!
Well sure glad I only bought 2 packs of Thermalright pads instead of the whole range like I was going to. Thanks for the great info and saving our chips!
Thanks for watching! I hope you found information on what works best for you❤️❤️
How come you only have 6k subs, it should be at least 60k, like for the Algorithm.
It's because I don't make clickbait nonsense videos that have no science and a bunch of junk food lol
Been waiting for this video! Thank you TekShinji! Curious to see if you can compare the Gelid Extreme vs Gelid Ultimate pads.
Ah perhaps that may happen !
No doubt, i saw four times this video, is the most interesting a fully of information in all youtube. Thanks a lot for a great job!!!!
Thanks! 🥰🥰
This is a great video to reference now after my 3090 FE needs maintenance after 3 1/4 years. Although there is something odd about your results with the Thermalright Oddysey pads, as those were the pads I used, and I got the same sort of temperatures you got with the Gelid Ultimate pads. My 3090 had the memory junction temperature at 112 degrees when new, and I ended up replacing the pads just two months after I bought it, which was very annoying and stressful.
I used the 1.5mm thickness everywhere and got the memory junction temperature down to only get as high as 86 degrees under load when 3D rendering, which is also heavy on memory usage. I don't mine, so can't say how well it did for that, but it seems to be anything that is heavy on memory usage makes for good testing. At the moment, my 3090 is running a lot warmer than it used to on the core, so the thermal paste needs redoing, but the memory juction temperature is still the same as when I did it back then. I found this video as I'm concerned that even though the pads work alright still, they could end up crumbling if I try to remove them, so I think it's going to be safer for me to have some new ones ready to put in, and if the Thermalright ones can be inconsistent, I'll go with the better Gelid ones this time around.
Thank you for such a good video.
Awesome I’m glad that it worked out for you! Please let me know if you have any other questions!
Awesome! I have 3 cards (one which is a 3090 FE) I want to replace pads on and this is invaluable! I was going to go with Gelid, but may consider NAB based on this video. Thanks for doing this!
Awesome! I'm glad it helped. If you don't mind tell all your friends about it, if it helped you!
Good luck!
Although the GPU temp increases were very eye opening as well. It looks like the fans will remain at 100%. Fans are a cheap replacement vs. replacing the entire GPU. I'd really appreciate understanding if putting those little heatsinks on the backplate help anything.
Sure we are guaging if ppl want a follow up since this project reviewed some crazy stuff!
No idea GPU core temp would go up!! It's truly incredible!
Thank you very much for this very comprehensive test. For use on PC laptops for non-gaming everyday work and personal uses, which thermal pads would you suggest for the CPU, GPU, RAM and SSD, and what thickness?
It would depend on what you’re using for that area. The best way is to buy a set of pads and measure them before you replace it.
What a great video, thank for posting this. Will you say the same rules apply to a 1080ti FE or Titan Xp? I want to change the pads and searching on the subject is how I landed here. I know the blower style is the worse as it is but will the core temps get higher with pads like 12W/mK? Maybe start with 6W/mK? any suggestions are welcome. Thanks and keep up the great videos!!!
Hey there, thanks for the comment and I'm glad you liked it. It would be dependent on what your goal is. Lower memory temps? Lower core temps? And really depends. Based on our data. You can pick the pad that works best for you!
Been waiting for someone to do this for a while now. Good job here.
Thanks buddy, we put a lot of work into this. Nearly two weeks of work and a lot of man hours. Andrew did a great job as well!
Interesting result for the Thermalright pads, I have just redone all of my pads on a FE RTX 3090 withn those same 1.5mm pads and contact is perfect, my memory temps dropped from 100c max down to 80c max.
Hey there, did you watch the whole video on the reasons for that?
Hearing the narator speak,
it's like entering the quantum realm of some kind. The voice is so futuristic :D
Only the best quantum vortex!
I would love a follow up on this 👍
Thanks buddy! Make sure to tell your friends and get subscribed 🙂
If I got it right, changing the pads only on gddr memory modules, you get lower temp on gddr, but higher temp on gpu core? So it's a mixed feeling situation? Or I got it wrong?
Nice test and congratulations for the video
Essentially, according to our testing. The cooler was designed to dissipate X amount of heat energy per time. When the efficiency of heat and energy transferred per time increases, it also affects the performance of the cooler.
Thank you very much for all this testing!
So 1.5mm pads should be right for a 3070 Ti Fe aswell right?
I've never personally worked on one of those myself. However, it appears that may be the case. Best idea to Google search yourself. This video is a comparison between the claims and the actual performance of the pads. Not a how to do video on a founder's edition. Since all pads tend to be slightly different depending on the card.
I just started watching your killer videos. This one is next level from my laymen mind. But you know what? I like it!
Thanks!
Hi!, do you think the NB supermax is going to last the same time in good condition compared to the gelid gp?, I don't know if the price difference means the nb supermax is worse.
Honestly, I’m not sure. I haven’t tried that one yet, but it seems that the ones that are softer don’t last as long
This is awesome and it's like the only thermal pad around up video around
I'm glad you like it! Let all your friends know about the good content you found! We definitely plan on doing more!
Geat content, thanks for this! I have the 3090FE, I’m only gaming and cannot decide if I should go for a re-pad or not given the increased heat transfer to the core. The reason I am considering doing this is fan noise - clearly the memtemps are what’s makes the fans spool up on auto while gaming. What’s to expect here? Would a repad, let’s say using the gelid’s, make the core hotter while gaming and in the end not make any difference in fan noise/make it worse or should I expect a quieter card with better ”average combined” temps?
It would be dependent on your situation. However, lower memory temps can never hurt. If you're looking for the lowest noise, then your memory temps are going to be high. If you're looking for the lowest temperatures, then your noise is going to be higher. There's no way to get a free lunch.
A lot of people like to go for the average or middle of the road. However, I did find more efficient heat transfer. Always turns out better in the long run
This is all very great info. I'm not changing pads, but it seems to be a reoccurring topic as of late. very interesting to see the results on core temp. i would love to see a follow up regarding different heat sinks.
Awesome!
@@TekShinji Outstanding job on condensing some highly-technical Data & Information while presenting it in a fashion that we can understand! 🇺🇸 😎👍☕
You're welcome! Make sure you get subscribed and hit that like button for future videos!
17:22 you sayin the Founders heat sink around the entire board isn't good enough? Maybe the Ko will do better as a heatsink coming from Asus and all. Right?... Wait that's a good question. Is it better to use decent thermal pads on Asus cards?
It's about it's designed purpose. For memory intensive applications. It's not a good design. However, for 90% of gaming applications, it's fine since the memory doesn't get taxed as much.
I don't understand the question though. Are you asking if it's good? I don't think it's bad.
Some Thermalright odyssey 1.5mm pads have issues with mounting pressure on the back side of the card
I changed my 3090 FE pads to 1.5mm thermalrights, I never saw the reductions other people were reporting, Few months later got another pack this time 2mm, I changed the pads on the back memory to 2mm and then I saw the large temp drop, in my case 1.5mm was too thin for the back memory modules, though 1.5mm was correct for the front modules
Great video, Wished it had been made a year earlier
yes eveyones milage will varry its not possible to test evey situation unless we were doing it our selves but thanks for the feed back
One more comment.... I would love to see that video about adding thermal sinks to the backs of cards. People report some pretty significant improvements in memory temp with that but I've only been able to knock off 2 - 3 DegC personally. I have actually gone to removing the GPU backplates entirely so the backs of the cards can radiate the heat more effectively and that did better than attaching heatsinks to the backplate. If you do the backplate heatsink test you should give this a shot -- you may start the TekShinji backplateless wave with that idea. :)
We actually have a video similar to that already on the channel. It was earlier on in our history. However, it approved to be not as effective because the thermal pads were already bad to begin with. Now we know it's better to replace thermal pads other than putting heat sinks on only.
Another big portion is airflow. If you add heat sinks to the back it's great but you got to have air flow to remove the heat as well.
@@TekShinji You are exactly right. I have fans on both my rigs to keep air moving across the cards. My current thinking is that the ambient air temperature is way more important that I initially thought. I was brand new to most of this at the time so I watched and read a lot but didn't know what info was reliable/trustable so I wound up doing my own tests for the most part. Misfit Mining has been a great find for me personally. Now I have too much good information to process! :)
Have a GREAAAAAAAAAAAAAAAAT day!
You're welcome, I'm glad you got a lot out of it!
The increase core temps is likely due to compression. Would you do follow up tests with thicker and thinner variants of the same thermal pads?
Hey there, This result came from testing on all of the pads. Which had different levels of hardness and thickness
@@TekShinji Thanks for the quick reply!
I should've been more clear. I mean pads of different thickness from the same manufacturer and the same lineup.
As you mentioned, there's no data on the original pads and limited hardness or compression data for the aftermarket pads.
It seems likely to me that the more conductive pads have less compressible filler material and more rigid conductive material, causing the GPU die to have less mounting pressure than it needs for better cooler.
I think the impact of heat transferring from the VRAM into the GPU die is much less than mounting pressure and die coverage.
It's impossible to conclude either with multiple variables changing simultaneously. Hopefully you could get a video out of it!
Thank you! Now I don't know if I should buy the thermalright for my MacBook
You're welcome, it'll be based on what you're looking for and what your goals are.
Thanks for this. Need thermal pads for some boards that are worth less than some of these pads, consoles, and laptops.
Nice to see reviews that break the one soze fits all mold.
Correct it’s something we found that it’s not “the best” but it’s situanttional
This video just proves that the Morpheus 2 is vastly much more better than what thought of initially. It's not even a competition anymore, the cooler king truthfully!
Yes, they're great!
Good works, the redundancy is appreciated.
Thanks!