I checked the thermalpads on the vram of my zotac 3070ti an those are 2mm not 1.5mm actually I tried using 1.5 mm and still had temp issues 2mm was a lot better I didn't replace the ones on the backplate
Same here, maybe its new revision of the card. When i replaced vram to 1.5mm thermal pads i got higher temps than on a stock ones... The correct size is 2mm for Vram and 2.5mm for Vcore.
@@sebastiankoataj2988 I have a zotac trinity 3070ti ..... I replaced the pads with gelid ultimate 2mm. At the front I could not find 2.5mm pads. The temperature of the ram has decreased by 15 degrees ..... the temperature of the gpu / hotspot is 5 degrees higher than before. It's normal?
Great job. You make it look really easy. I will give mine a try soon as I feel ready. Also, after applying the new pads, how much did you tighten? A lot or at slight resistance? Thxs in advance.
Your biggest problem is not the pads but the heatsink , bad soldering on fins maybe, but mostly because is Ni coated on contact plate of heatsink with GPU and other hot components, which is very poor thermal conductor. Same as my EVGA 1080 Ti FTW, Nickel coating and black paints is just hindering more than you want the efficiency of any radiator. You can add a 92 mm just on the backplate of the card sitting in angle that will take at least 5-6 C off your GPU temp @ only 1000 RPM. As for GPU RAM add copper shims and less tick thermal pads and that will take another 5-8 C- make sure pressure on GPU is right after placing shims - basically if original pads are 2 mm thick put a copper shim of 1 mm + 1 thermal pad of 1 mm which will bring it to 2 mm. Don't keep your card at 70-80 C cause other parts in computer will suffer as well. GPU or RAM won't burn but remember at 80 C other parts on on your video card will might not work as expected or jus die.
First time replacing thermal pads, was looking to improve my thermals to make my rig more silent so that I can lower the fan curve. => bought 3 sizes (Thermalright Odyssey Extreme 2.0, 2.5, 3.0). Opened my gpu up, removed and verified that the pads zotac used matched with my 3.0 and 2.5 (foreshadowing, uh oh) by lining them up and comparing => applied 3.0 back, 2.5 front. After applying the new pads VRAM was insanely cool, but my core temps were absurdly high, fans spinning etc. thought my gpu was gonna explode juding by my fans lol Reopened and the GPU-die thermal paste was barely spread on the cooler contact plate. I squished the front 2.5 down a bit with to make them even with the gpu-die, should be a good fit now, core temps are fine but the VRAM is sitting at 70 degrees in idle. After about 6 hours of either Core or VRAM being too hat & constant reapplying and testing this is kind of a bummer. Well, I guess my pads are way too hard, Zotac used very soft ones. My backplate is also a bit bent from the thick 3.0 on the back. What I will take from this is watch out what hardness you buy for your thermal pads, unless you exactly know the specifications of your needs, softer ones will make things way easier as they allow for more margin of error. Will try again with softer pads
Thank you for this, I also have an issue with this card with the fans making a strange noise, it's almost like a grasshopper, and seems to be from the fans rubbing on the shroud. Do you have that too?
I had something similar and I tried to reseat the shroud and fans but still, then I went with a vertical mounting - that eliminated the issue. Check out also that the one of the fans isn’t touching any cable. Another thing you can try is to spin at 100% for a minute the fans and check out the noise. What I’ve noticed to my GPU is that having the speed of the fans below 45% the noise is strange, I managed a custom fan curve in MSI Afterbruner like so: 0-50C fan speed 45%; 60C/50%; 70c/60%; 80c/70% (really noisy and aggressive) give it a try. I think the shroud+fans in this card aren’t really well done assembled.
@@JULVEXGAMING Thanks for the reply! Mine also behaves differently depending on the speeds. I also found pulling on the edge of the shroud seems to alleviate it so I wedged a small piece of plastic under it and it's mostly quiet now. A vertical mount hadn't occurred to me I'll try that. I agree about the build quality too, my vram temps are pretty horrible, fans sound like a jet engine over 70%
Thanks for guiding us. Do you know what is the purpose of the square black plastic around the GPU core (front side) for? I followed your guide and managed to get my temps down by 8C - 10C. But as I was busy cleaning up, i just realized i took the black plastic out to clean the it separately and FORGOT to put it back on.
I assume you're referring to the black square shape of the core in that case you have to put it back on it can prevent the core touch other surfaces that should not be in touch
Thanks for the Video this was very helpful. I have Zotac 3070 ti trinity. Just one quick note - I think the dimension for the last in memory module shud be 30×15?. Also could you please let me know the screw type or rating which u used to open the GPU. I will order it online.
@@JULVEXGAMING Hey hi. Sorry to trouble you. I planning to do the same and have few more queries on this: 1) Was GP extreme 80×40 mm enough for backplate?. The dimensions u had given for the backplate in description are not entirely fitting for 80×40. I tried drawing on paper to check. 2) At 6:37 I see a large thermal pad attached to right of heatsink. Did u replace that as well?. Would it be possible to give the thickness and dimension for that? 3) Due to availability issues in my country I am planning to go with GP extreme 2mm 80×40 for memory and GP Ultimate 3mm 90×50 for back plates. This should work rite?. Your thoughts please :-)
@Annmol Mahtani 1. yeah the dimmensions are approx - consider that the moment you put back everything (in this case the backplate) the thermal pads will "squish" about 1-2mm; sp if you decide to cut it some mm less in order to fit all the drawn you can do that 2. yes because I had one Gelid spare thermal pads, but I think that migh not be that much important in heat transfer anyway the dimensions are approx 80x9mm the thickness is ok 1.5mm (if the thermalpad is "squishable" enough) 3. I've into that as well (seems that the thermal pads have more shortgae than gpu's :) ) as I can see that would work but consider one very important thing: squish it a bit before installing all back otherwise the heatsink surface might not be touching the core surface (don't forget to apply properly a good thermalpaste); and the result of this will be maybe low temps on memory but high temps on the core.
Hello, thanks for your video, I can’t get the heatsink off. It is stuck with the pads..any ide on how to remove it without breaking everything? Thanks!
Hey thanks for the great content, it really helps. I got a question, I am about to buy this 3070ti and was wondering if I must change the thermal pads as soon as I have the card or if I must wait, thanks in advanced
Thanks. I suggest you to don't change immediately, what I would do is installing as it is and check the temps while gaming for some days, after that you can decide if you're ok with those temps or change the pads and also, important repaste the core.
Those who are doing this, PLEASE buy 2mm for the vram ESPECIALLY IF YOU ARE USING THERMALRIGHT PADS. 1.5mm is complete bs. I tried and there is minimal contact with the heatsink. Before: 98-100 After: 100-110+ EDIT: so i went full “dicko mode”. I used the remaining 1.5mm pads and layered them on top of the existing pads. Squished them all down gently with my finger, re-applied thermal paste, reassembled, and temps are now 1 layer: 100-110+ 2 layers (please remember to squish it): 88-94 Do let me know if it works for u too! 🙏🏻
Normally the card shouldn't have any big heating issue, the thing is that depends on how demanding the game(texture/resolution, etc) is, and also it depends on what condition you're buying it. Keep in mind that you might lose the warranty -depending on where you live, better ask your regional Zotac support just to be sure
@@JULVEXGAMING nice but that's not half the power drow. Can you you please test your GPU in 150w or on it's half power and see how much performance you're losing and temps. If you're still using 3070 ti.
@@vzero8612I don't have that GPU anymore, but usually, I would recommend undervolt instead of using the power slider just to avoid strange behavior with the power spikes.
Hello, just some questions if you have time as I see you know a lot of stuff, I boguht 1 month ago a 3070 ti zotac trinty, to play and when i am not playing mine with it and i get 94 - 100° in memory and 43° in the core, with my air on i get betwen 90 to 94 in memory, you consider those are good temps? i fear it could make some damage as are too high, I want to change the pads but you know, dont want to lose the warranty, but on the other hand after 1 month it works fine and have no issues and i dont think it will fail, so i think it may not worth to stay with the warranty and not change the pads, do you know if the warranty works if the gpu get damage for stay in 94°+- 24/7?
Hey there, the tjunction of this memories GDDR6X is around 105C so you won't damage anything (the GPU itself won't let you 😀 ), on other hand having those temps on idle isn't that good honestly - in short terms (usually 3/5yrs) nothing would happen but in long terms, you might have some issues. That difference between core temps and memory temps isn't that normal tbh. I would try to increase a little bit the airflow or fan speed in your build. If, where you live, changing the pads would invalidate the warranty and you don't want to do it then don't go for that. Also, in idle 94C 24/7 isn't something I would recommend 😉
My 3070ti OC Trinity already gets 78c running that benchmark on the stock thermal pads so not sure this would do anything. Where it hits 100-106c is running NiceHash mining. Wonder if I should bother trying. Seems like Zotac may have updated their thermal pads.
I guess yes, but the stock thermal pads were really terrible. I tested also with NiceHash and reached before changing the pads, I assume that the first 3070ti GPUs that came out kinda have a really bad quality thermal pad mine were like oiled and melted after almost 2month, but if your temps are ok and you’re ok with that really not need to change them
@@JULVEXGAMING thanks for the response. Can you also comment on the thickness of the pads? Most of the recommendations say to use 2mm pads, can you comment on how you came up with the 1.5mm thickness? Not saying it's incorrect but curious if you measured this somehow?
@@thepiecesfit5049 that's true, but it depends a lot on the type of thermal pads you're using (more specifically on the shore/hardness), on how thermal paste also you use on the core - because that will increase about 0.2/0.5mm all the total thickness, I tried to measure the old one with a caliber but then I consider all these factors, and also because doesn't make much sense to measure a thermal pad after that has been applied; so went for the gelid ultimate 1.5 which has a shore of 60/70 which isn't really "squishable" (I know isn't the most technical but just for giving the idea ;) ). In my opinion, it depends on these 3 factors: shore of the thermal pad, amount/thickness of thermal paste on the core, and how much all the screws are tightened; with that said I think the most precise thermal pad, in the range of shore 30-50 might is 1.7mm, which I assume doesn't exist so this is why I went for the GP Ultimate 1.5mm which have a shore of 60/70 and in my case were doing ok. In the past, I went for the gelid extreme 2.0mm and wasn't really satisfied with the temps, and how much they were "squished". Do keep in mind that all these results can change from GPU to GPU based on how each one of us goes through the process. My final recommendation is: if you have thermal pads that are in the range of 30/50 shore you might go with the 2mm, if you have thermal pads with a higher shore then a 1.5mm thickness might be the right choice.
i have some doubts, when hot spot is 90c board power draw is 281w chip power is 194w pcie slow power 43w 8 pin 1 power draw 121 8 pin 1 power draw 116 tdp 91% when hot spot was 95 board power was 290w , is this normal for zotac 3070 ti oc with 5900x 1:51:54 Time: 260.615 11:51:54 Frames: 46163 11:51:54 FPS: 177.131 11:51:54 Min FPS: 60.5659 11:51:54 Max FPS: 374.183 11:51:54 Score: 4461.94 Settings: Render: Direct3D11 Mode: 1920x1080 8xAA Fullscreen Preset: Custom Quality: Ultra Tessellation: Extreme No OC, No Undervolt , bechned on heaven
So you used 1.5 mm on the die side and 3mm on the backplate? I used 2mm die side and 3mm backplate and the core temp is very high now. I think I need to go back to 1.5mm die side.
Yes I used 1.5mm pay attention to really te thermal paste correctly on the die and before close definitely the heatsink give it an eye that the is touching correctly the die otherwise won’t transfer the heat
@@JULVEXGAMING unfortunately you did not show how to apply thermal paste correctly, I'm worried that I did something wrong, although the core temperature has become lower, but the memory temperature is higher. maybe I applied too thick a layer of thermal paste and the heatsink is not touching the pads?
Hello there, you can try, with a lot of patience, to replace them without opening all the GPU; if opening the GPU isn't going to void the warranty where you live, well I suggest you opening completely would more easier
I guess you missed the exactly place by couple of inchs of the Back Plate Thermal Pad on the Right... U can see that when u blink the image from the older to the new. Nice video BTW! Ty!
Hello thank you for your video! I tried the process but I stopped since I couldn't remove the heatsink, it seems that the heatsink is really stuck on the pads and doesn't move..any idea how to get it unstuck? Thanks in advance!!!
Thanks! I suggest you to check if you have unscrewed all the screws, if this is the case then I guess it might be that the heatsink is stucked into the core instead of the thermal pads (similarly when you remove the cpu cooler after a long time and the thermal paste is really dry); pay attention also you have disconnected also all the fan/rgb cables too.
You can loose the warranty. I usually don't mine but I tested anyway for about 1hr with nicehash and the memory temp didn't went more than 92C - bear in mind that these depends a lot on the airflow of your case, the speed of the fans of your gpu, room temperature and many other factors (overclocking included). I used all default and my room temperature is around 20C more or less.
The core temps, with stock settings, basically changes 2/3 degrees less because of the thermal paste I’ve used (Noctua NT-H2); but if you want to for less core temps try out undervolting which is the way to go with the RTX 30 series in my opinion - if you’re interested check out my videos on MSI Afterburner tutorials
Yeah. Undervolting is a must. I have the same card, and it was unacceptably loud and hot at stock, hitting the 83degree limit with fans running @ 85+%. I am now running at 850mV for the same core clocks, dropping core to 73degres and 74% fan speed. I have ordered 1.5mm gelids for the front and will replace paste. Hoping that the memory decrease will drop the fan speeds more. Not to sure if I am going to bother with swooping rear pads.
@@kabal911 rear pads can help dissipate better the memory (this GDDR6X are tremendously hot) and all the temps in general; try, while gaming, touch the backplate…it’s a oven 😅
You are probably right. Might as well do it while I am disassembling it anyway. Thanks for this video btw, otherwise it would have been very difficult to know what size pads to orde
just to follow up here, IMO, 1.5mm pads for the front are not correct. the stock are 2mm. my cpu core went down (likely due to new paste), but memory went up from 90C to 100C when switching over. my fault, I shouldnt have continued with the pad swop even though I measured the stock ones before removing them all. have ordered 2mm and will replace
Hello gentlemen, could you please help me choosing the right card, I write down my choices: These are the 3070 Tis with price tags (after tax) in my region: Asus Strix: $1030 Asus Tuf: $900 Zotac AmpHolo: $900 Gigabyte Aorus: $885 Zotac Trinity: $825 Zotac Trinity OC: $835 I'm leaning toward the zotac because of their low prices, do you think the zotac trinity, trinity OC, or ampholo edition good cards? is there any difference between these three cards in terms of hashrate, cooling performance, or heatsink.
Hi, If I had to choose probably the Strix would be my first choice beacuse of the heatsink and the VRMs, then the TUF and Gigabyte Aorus. But, this depends on your budget and what are you planning to do with the GPU. Whatever you choose in this list is going to serve you well.
Hello there, if the high temps are on the core try to repaste it, it happens to me that i have to repaste due to the shore of the pads, otherwise if the high temps are on the memory try to squish it or check if you tighten properly (don't overtight or you can damage the pcb!)
It might be, the only thing to consider is the shore of the pads to me some (not all) pads with a soft shore were ok others not. From what I’ve seen the exact measure (measured with caliber) is 1.7mm which afaik aren’t available. With most of the 2mm you should be ok
If you guys are running temps in the 80’s you have serious problems with your case and/or ambient temp of your room. I have a Zotac 3080 Trinity OC 12gb that were released last month and I’ve yet to see my card go above 62C during benchmarking/stress testing as well as playing Cyberpunk on max settings.
I have this exact card and when I’m gaming at 2K my core temp gets up to 82C and my hot spot gets to 92C. Should I be worried about this? I don’t feel that comfortable at this temp but I don’t want to void my warranty. Really what I’m asking is if I continue to use this card like this for gaming (3hrs a day average) will I experience any permanent damage?
Hi there, the core temp is a little to high in my opinion considering that you are gaming in 2K (if RTX is on the temps can go high, that's true), hotspot usually it is similar to the memory temps and 92C, while gaming, is ok. The thing is, if you continue to use it like this in short term (let's say 3/5yrs) you'll be more than fine for me and you shouldn't face any permanent damage . Some additional tips: try to improve your the airflow of your case - if you haven't done it yet, or most importantly you can try to undervolt the GPU with MSI Afterburner (it will be a guide on that in the channel), that will dramatically reduce the core temp and you won't loose basically nothing in performance. Certainly, re-apply the thermal paste can help a lot, but if you don't feel comfortable don't do that especially if that would void the warranty.
@@JULVEXGAMING Hey man, I really appreciate the reply and the advice. I found a really easy fix (you cant use the graphics card at full power) but in zotac firestorm my target gpu temp was 83C and this was making my temps reach 82C, I changed the target temp to 80C (graphics card is only at 88% power) and my new average gaming temp became 76C and 86C hotspot. Obviously the card isn't running at 100% power but this really helped me out and I'm now comfortable with the temp and performance. You can also set a custom fan curve in firestorm. I'm just leaving this here in case anyone wants to change gpu temps and not void warranty. Thanks again for your advice, your a great youtuber and keep up videos like this because this is exactly what people need to know, goodluck and keep up the hard work!
@@lanetomlin5381 thanks! :) I've used the Firestorm sometimes and low the power % just to get quickly low temps; also because it is the only way to turn off or change the RGB of the GPU. Doing so you aren't going to loose basically nothing on performance, just sometimes turn on RTX (whenever possible) or max the details of the game - this would help to understand if the GPU require more watts to drain ;) and avoid random crashes, for the rest happy gaming :)
Hey there, I assume so but you should verify locally with your retailer or Zotac. You can lower the temps a little bit by undervolting the GPU while maintaining basically the same performances, I've done a tutorial on how to do that, check it out if you're interested, Also keep in mind that the airflow of your case is playing an important in order to have good temps.
Don’t use 1.5 for the front, I used 1.5 and it wasn’t cooling correctly (exact same pads he was using in the video) I then saw people saying to use 2.5 gelid extreme so I did and the memory did go down to below 50 but it was too thick for the radiator to touch the gpu core so it wouldn’t cool properly. I’m now ordering 2.0mm and will let y’all know how it works out because I’m guessing it’s 2.0
@@leandrorberto for me yea, just make sure you clean off the old stuff real good and you don’t over tighten when reassembling because it can crack the gpu die
I did everything as in the video, but the memory temperatures only got higher. even started dropping frequencies. what could i have done wrong? maybe not pressing enough?
It could be possible, did you monitor also the core temps? Because that might give an idea if you have to press more or not, in fact if the core temps are high that means if the heatsink isn't pressing properly (except the fact that the thermal past is applied correctly); with that said it depends also on what thermopads you're. I would suggest trying to press a little bit but DON'T push too much.
@@JULVEXGAMING thanks for the answer. I used all the pads like yours. only I did not understand why the second thermal pad 120x20 1.5mm. Is it needed instead of that massive thermal pad that connects the heatsink and the backplate directly under the 3rd fan? I didn’t replace it, but it’s 2mm, and perhaps because of it, the new thermal pads don’t touch the radiator? 🤔
@@JULVEXGAMING I use video card in mining. It seems to me that immediately after the purchase, the temperature of the core and memory differed by about 20-30 degrees. Today, 9 months later, Before changing the thermal pads, I had a core temperature of 50-53 and a memory of 102-110 at 75% fan and a room temperature of 25 degrees. after the replacement, the core temperature is 41-43, but the memory temperature is more than 110 and the core starts to lower the frequency. I will also say that in idle temperatures 37/62 at 50% of the fan.
DESASTRE! ☹ Las termicas de 1.5mm son poco para esa tarjeta. es necesario las de 2mm. Ademas 12W/MK es justo para esto. resultado una vez cambiadas. se pone en 104º en segundos al meterle caña, y se reinicia solo. Las he vuelto a cambiar por las de 2mm, un modelo de 20W/MK, ahora no pasan de 74º
Creo que depende de las almohadillas térmicas y de la GPU, me pasó lo mismo con otra 3070Ti y en ese caso usé las almohadillas de 2 mm. Normalmente el Gelid GP Ultimate es de 15W/mK con un shore de 60-70 OO. Tenga en cuenta que, por lo general, todo esto depende de GPU a GPU, lamentablemente no todos son iguales, a pesar de que pueden parecer tan (perdón por mi español :))