Possible 1660 Super Gaming X problem?

[Fergutor]

I bought this MSi GTX 1660 Super Gaming X 2 months ago, which means, here in the south that it was cold in winter.
With low room temperatures, bit less than 20ºC my card got to around the expected 67ºC with furmark and around 1500RPM
More recently, still with low temperatures, but just a bit higher (less than 5ºC more) I noticed 73ºC and 1700RPM or more.
Today with around 27ºC room temperature: 77ºC and more than 2200RPM (don't remember, probably up to 2500RPM).

So while room temperatures went higher, the card didn't increase temperatures proportionally and accordingly, but actually it looks like it's getting terrible with time.

Also I noticed that pressing the backplate from the backplate and shroud the tempreature decrease significantly, today from 76ºC with open case to 67ºC. And sicerely the card feels filmsy and all parts move when touching them. Like a cheap delicate toy.

Also I noticed when first used it the card was boosting beyond 1900Mhz for all the gaming sesion, even getting to 2000Mhz at the beggining, then, still cold days, 1875Mhz is the faster I saw, and 1830-1850 are the normal numbers...Now it still do the same but with way higher RPMs.

Is there anyone with the same card model that could verify my observations,? Or is my card dying? Possibly wrongly assembled?

Here are some reviews:

MSI GeForce GTX 1660 Super Gaming X Review

The MSI GeForce GTX 1660 Super Gaming X comes with a large dual-fan cooler that makes it the quietest GTX 1660 Super we've reviewed today. It also comes with the highly popular idle-fan stop feature. Temperatures are good, too, and a backplate is included as well.

www.techpowerup.com

MSI GeForce GTX 1660 SUPER GAMING X review

MSI released their GTX 1660 SUPER GAMING X, a product with a more than capable cooler as well being factory tweaked. The SUPER edition card comes with faster GDDR6 memory but remains pretty much to be...

www.guru3d.com

MSI GTX 1660 Super Gaming X Review: Pricey Performer

While its price is high compared to the new 5600 XT, this 1660 Super performs well.

www.tomshardware.com

In the Tomshardware review they test with furmark, and you have both temperatures and fan RPMs...

 

[Zach_01]

First I would suggest to stop using furmark every day. This is a burn/stress test to use once in while for stability purposes.
Its very abusing to any card, and if you continue to use it frequently, you will kill the card eventually.

Your GPU temp findings according to abmient(room) temp increase seem right to me. It cant be a proportional increase. Your room is heated, so case is heated and PC components are heated even more.
You can improve case air flow for starters.

Also you can remove GPU card backplate check if the GPU cooler is tight on the PCB and maybe add some thermal pads behind GPU and memory chips to transfer heat to backplate. If backplate is metal. If its plastic wont help much.
But check if cooler is tight or if it moves around.

Uninstall the card first and have it on your hands. Try not to knock off any small components from PCB, or you could end-up with more and difficult issues.
Those little capacitors are somehow easy to be removed.

 

[Fergutor]

First I would suggest to stop using furmark every day. This is a burn/stress test to use once in while for stability purposes.
Its very abusing to any card, and if you continue to use it frequently, you will kill the card eventually.

Your GPU temp findings according to abmient(room) temp increase seem right to me. It cant be a proportional increase. Your room is heated, so case is heated and PC components are heated even more.
You can improve case air flow for starters.

Also you can remove GPU card backplate check if the GPU cooler is tight on the PCB and maybe add some thermal pads behind GPU and memory chips to transfer heat to backplate. If backplate is metal. If its plastic wont help much.
But check if cooler is tight or if it moves around.

Uninstall the card first and have it on your hands. Try not to knock off any small components from PCB, or you could end-up with more and difficult issues.
Those little capacitors are somehow easy to be removed.

No, I used Furmark only those times and a couple more and use only for few minutes, and I did it because I noticed strange temperatures, also at buying the card of course to check it.
Please take in account that I resumed a lot not to make my original post too long. Don't think I'm a total newbie. I'm asking for data from people who have this same card or people really experienced with GPU problems (RMA, repair enthusiast people). My many other GPUs never acted this strange. One of the things I didn't put explicitly but you can deduce is that I also tested with open case, but temperatures weren't very different than with closed case.
I can't dissasemble the card given that I, as I said, bought it 2 months ago and that will void my warranty...

 

[sneekypeet]

Aside for possible lower clocks from temps reducing boost, what is the issue?

If nothing, don't worry about it.

I have had a few MSI cards aver the years, and I can say, unless it was a top tier offering, they do feel cheap and flimsy compared to others I have owned.

 

[Fergutor]

Aside for possible lower clocks from temps reducing boost, what is the issue?

If nothing, don't worry about it.

I have had a few MSI cards aver the years, and I can say, unless it was a top tier offering, they do feel cheap and flimsy compared to others I have owned.

The issue is that, apparently, is getting bad with time, and at this rate in 2 more months it could be unusable, especially deep in the summer. 77ºC is fine...but not at this room temperature with this kind of cooling and much less 76ºC with open case at this room temperature. I never had a card that did this. I expect this behaviour after ~7 years of use with maintenance, after the paste go bad, the fins too dirty, the fans struggling and the screws get loose, not two months. I could say I have the warranty but sincerely, by how I bought it, I'm not sure I could use that warranty...but will see.

About feeling cheap and filmsy, yeah, totally. Before this one I had a GTX 970 Strix OC and that was a tank! And all other cards were way better at that than this one. Really dissapointed with it.

 

[rtwjunkie]

I bought this MSi GTX 1660 Super Gaming X 2 months ago, which means, here in the south that it was cold in winter.
With low room temperatures, bit less than 20ºC my card got to around the expected 67ºC with furmark and around 1500RPM
More recently, still with low temperatures, but just a bit higher (less than 5ºC more) I noticed 73ºC and 1700RPM or more.
Today with around 27ºC room temperature: 77ºC and more than 2200RPM (don't remember, probably up to 2500RPM).

So while room temperatures went higher, the card didn't increase temperatures proportionally and accordingly, but actually it looks like it's getting terrible with time.

Also I noticed that pressing the backplate from the backplate and shroud the tempreature decrease significantly, today from 76ºC with open case to 67ºC. And sicerely the card feels filmsy and all parts move when touching them. Like a cheap delicate toy.

Is there anyone with the same card model that could verify my observations,? Or is my card dying? Possibly wrongly assembled?

What are gaming temps, or if not gaming, other GPU work temps? Furmark running hotter in the summer and abusing your system is like saying air is for breathing. Real world examples would be much more helpful in diagnosing a problem.

yes, unless it is a high end card by MSI, like a 1080 or 1080 Ti, you will find that the heatsink is attached to pcb by only 4 screws, and so there will be some flex and sagging.

So, there will be some difference in cooling when you squeeze a little, as this minimizes heatsing pulling away from the core and RAM. On the higher ends the whole unit is much more solid.
You may find that a support brace installed will keep this sagging from happening.

 

[Fergutor]

What are gaming temps, or if not gaming, other GPU work temps? Furmark running hotter in the summer and abusing your system is like saying air is for breathing. Real world examples would be much more helpful in diagnosing a problem.

yes, unless it is a high end card by MSI, like a 1080 or 1080 Ti, you will find that the heatsink is attached to pcb by only 4 screws, and so there will be some flex and sagging.

So, there will be some difference in cooling when you squeeze a little, as this minimizes heatsing pulling away from the core and RAM. On the higher ends the whole unit is much more solid.
You may find that a support brace installed will keep this sagging from happening.

Gaming temps are another issue because gaming gives a variable load so it goes fro fine to bad, but of course still lower than furmark. Now, what's the point of Furmark? That's how/why I'm using it. And only few minutes at a time (from 5 to 20). I still never tested it in hot summer, simply because it gets too hot here and didn't wanted to suffer more heat by testing this (not the card, but me). I did test when the temperature went down. Still, the videocard is supposed to work at higher temperatures. Especially this one that is known to be very good at staying "cool" (which takes me here asking) and especially given I had videocards that consumed more with worse cooling and no problem. But, will see when I decide to use the card in one of those hot days. Listen, my Asus 970 Strix OC never went beyond 75ºC, and probably was lower before I changed the paste (from original to MX2)
The Strix was also attached by 4 screws...as most other modern cards I had...and no sagging, no flexing, no moving...
I did said pressing but actually only by touching gently from below is enough to lower the temperature, which is incredible to me. Also, the heatsing doesn't touch the rams...
Now I tried to squeez the 3 remaining screws, you know that 1 is covered with the warranty sticker, and I know this is not recommended due to asymmetrical pressure. But I had to see if they were properly adjusted...and they were! I could barely move 2 of them.
So personally I guess that if there's actually a problem is either a faulty heatsink (some manufacture failure), lack of thermal paste or improperly applied, or faulty heatsink's retention.

I added some reviews and what I forgot to mention about the boost clocks in the original post.

 

[Fergutor]

Yep definitely somethings not right. I'm baffled how nobody can see the problem here.
First bench 11 days after I bought I bought it and the second, today. Today here the temperature is very low, I'm sure a lot lower than the days of the first bench, and the minimum temperature in todays bench is 11ºC lower than in the first bench. And still the maximum temperature is 4ºC higher and the score lower.

Nobody here has this card and could share some data?

 

[xman2007]

Yep definitely somethings not right. I'm baffled how nobody can see the problem here.
First bench 11 days after I bought I bought it and the second, today. Today here the temperature is very low, I'm sure a lot lower than the days of the first bench, and the minimum temperature in todays bench is 11ºC lower than in the first bench. And still the maximum temperature is 4ºC higher and the score lower.

View attachment 172732View attachment 172733

Nobody here has this card and could share some data?

That's around 1% and well within the error of margin, I see no issue.

 

[ThirdhandTaters]

What is your fan curve setup to be? 1700 RPM with 67ºC seems like your fan curve isn't setup for what you are doing. I have the exact same card and mine doesn't go above 61º C while gaming, and it's OC'd to 2100MHz core 7950MHz memory.
.

This is the fan curve that I use and it's universal for whatever program you might use, this is Argus Monitor. IF your card were to still get to 67ºC+ with this curve then the fans would be going at about 85% of their max speed, close to 2700 RPM I think, and therefore cool the card to where the fans could slow down.

One other thing that I did, and if you choose to do the same you do so at your own risk, is I changed the thermal paste on the card. I'm using Noctua's NT-H1. I did notice some coil whine after doing so but I have no way of knowing whether or not the whine started because of the change or not. I did notice when I took off the cooler that there was A LOT of thermal paste on the chip itself, I can't remember right now what temps were like before the change so sorry on that question. if all else fails and you haven't tried changing the thermal paste, and don't want to, see if MSI can switch cards with you, I don't see why they wouldn't because they are keeping you as a customer and they can get the possibly defective card back to fix the problem and still be able to sell it for something.

Another thing that I have seen from people like Linus Tech Tips, JayzTwoCents and a couple other tech youtubers is that the first test will always show a higher score because the card is just "starting" to do some work, much in the same way you might do work, you are good and strong at the beginning but after some time you slow down some, however minute, because now you are getting tired and worn out.

Try out that fan curve and see if it makes a difference, it may look aggressive but the reason why below 30ºC the fans are at 0% is because if I'm not gaming the card is barely being used so the fans can actually turn off and save their lifespan.

 

[Fergutor]

That's around 1% and well within the error of margin, I see no issue.

What, the score? That's not the important part, the temperatue is. Is higher (and fan RPM higher too but that is not showing) than it was just 2 months ago, with lower room temperature too.

What is your fan curve setup to be? 1700 RPM with 67ºC seems like your fan curve isn't setup for what you are doing. I have the exact same card and mine doesn't go above 61º C while gaming, and it's OC'd to 2100MHz core 7950MHz memory.
.View attachment 172735

This is the fan curve that I use and it's universal for whatever program you might use, this is Argus Monitor. IF your card were to still get to 67ºC+ with this curve then the fans would be going at about 85% of their max speed, close to 2700 RPM I think, and therefore cool the card to where the fans could slow down.

One other thing that I did, and if you choose to do the same you do so at your own risk, is I changed the thermal paste on the card. I'm using Noctua's NT-H1. I did notice some coil whine after doing so but I have no way of knowing whether or not the whine started because of the change or not. I did notice when I took off the cooler that there was A LOT of thermal paste on the chip itself, I can't remember right now what temps were like before the change so sorry on that question. if all else fails and you haven't tried changing the thermal paste, and don't want to, see if MSI can switch cards with you, I don't see why they wouldn't because they are keeping you as a customer and they can get the possibly defective card back to fix the problem and still be able to sell it for something.

Another thing that I have seen from people like Linus Tech Tips, JayzTwoCents and a couple other tech youtubers is that the first test will always show a higher score because the card is just "starting" to do some work, much in the same way you might do work, you are good and strong at the beginning but after some time you slow down some, however minute, because now you are getting tired and worn out.

Try out that fan curve and see if it makes a difference, it may look aggressive but the reason why below 30ºC the fans are at 0% is because if I'm not gaming the card is barely being used so the fans can actually turn off and save their lifespan.

Finally!! One with the same card!
Thanks for all the tips and info.
That's a very, very aggressive fan curve. I think the default one is fine: MSi prioritize acoutics over temperature, which I like (when I bought it I couldn't hear the fans, but now, you bet I can). But yeah, all what I'm doing here is with the card at stock. At stock is fine, my problem is the card quickly getting worse with time, getting way hotter and also slower with only 2 months of use and not much gaming (and yeah I cleaned and checked all my hard).
I will not change the paste for the moment (I have MX-2 2019 edition that I used on my GTX970 with not very good results) because if I do I would be voiding he warranty (screw with sticker, at least) and I have to see if my card is actually dying. I tell you I am so tempted to open it and see if there's something wrong, change the paste, thighten the screws, put some washers if the screwws aren't enough, mod or fix the heatsink if it's something wrong with it...but no! I have to wait and see if I have to send it to warranty.
But thanks again. And hey, even if you changed the paste could you make a quick test with Furmark or Superposition in default settings (clocks and fan)? If you do also tell me the room temperature. Thanks!

 

[MaurizioC]

I have a MSI 1660 Super Gaming X Card.

My card temperature is always between 67-69° under load.

I have been using it for folding@home almost night and day full time. Temperature is in that range and the fan is at about 1500rpm.

What changed in your hardware and software configuration to cause this temperature delta? Mine even ifoverclocked stays below 70°

Did you check if both fan are spinning under load? Did you turn off some other fan in the case? Did you install any software that could temper with fan curve or gpu settings?

The score variation is not relevant, but the temperature variation is strange. If rpm are going up this way maybe one of the fan is broken?

 

[ThirdhandTaters]

Stock


Overclocked


Maybe it's the difference in CPU?

 

[Fergutor]

I have a MSI 1660 Super Gaming X Card.

My card temperature is always between 67-69° under load.

I have been using it for folding@home almost night and day full time. Temperature is in that range and the fan is at about 1500rpm.

What changed in your hardware and software configuration to cause this temperature delta? Mine even ifoverclocked stays below 70°

Did you check if both fan are spinning under load? Did you turn off some other fan in the case? Did you install any software that could temper with fan curve or gpu settings?

The score variation is not relevant, but the temperature variation is strange. If rpm are going up this way maybe one of the fan is broken?

I knew it!
Yeah, there's something wrong here...
No idea what could it be...
I'm betting faulty hardware. Something went wrong with the heatsink or it's mount.
No change in hardware and much less change in software. I did change three fans but not in a way that could make things worse, on the contrary, and besides I'm getting the same results with open case.
(But for those who could hang on this I changed: an old Cooler Master fan that was malfunctioning and I tried to fix but didn't worked. For an Aerocool with more blades (all I could get). That one only throws air through the HDDs, old case. Then I removed a fan in the side panel that I was using to remove air next to the GPU. I removed it because it was actually worse. In fact I tested two times, yesterday was the second, that in extract and intake both ways made it worse so it was better not putting one. The thirs was the extract fan at the rear, replaced the other old Cooler Master with the exact same problem than the other for an old re-oiled Spire with double the RPM which I will remove again because, just as with the CM fans the cleaning and re-oiling just didn't worked. And I used propper machine oil. So as you can see the changes couldn't make it worse, just better and as I said I tested with open case too anyway).

Then yes, both fans are spinning, I checked. Both are in perfect condition. No other software is acting upon the GPU. It's all in stock and I can corroborate that the fan curve is what it should be, I said in the other posts too: the temperature just went up and the RPMs too. The cooling system is simply not working as it should, as it was when I bought the GPU.

By the way the reason I get (almost) the same results with open case as with closed case is that the side panel has a breathing grill (?).

Stock
View attachment 172758

Overclocked
View attachment 172761

Maybe it's the difference in CPU?

Look at those temps!
Yep, something's not right here...
But hey! I get higher score in the stock setting
Room temps?

No, the diff in CPU can't be, at least in the sense that at 3.7Ghz mine isn't putting much heat and neither Furmark or Superposition use the CPU a lot and I have a tower cooler with the airflow to the rear directly to the exhaust fan...hmmm could it be my PSU?!?! Since I bought this GPU I'm getting a funny smell but I tested everything, and even removed the PSU and couldn't pinpoint the origin of that smell...

 

[MaurizioC]

Here are the benchmark screenshot.

I knew it!
Yeah, there's something wrong here...
No idea what could it be...
I'm betting faulty hardware. Something went wrong with the heatsink or it's mount.
No change in hardware and much less change in software. I did change three fans but not in a way that could make things worse, on the contrary, and besides I'm getting the same results with open case.
(But for those who could hang on this I changed: an old Cooler Master fan that was malfunctioning and I tried to fix but didn't worked. For an Aerocool with more blades (all I could get). That one only throws air through the HDDs, old case. Then I removed a fan in the side panel that I was using to remove air next to the GPU. I removed it because it was actually worse. In fact I tested two times, yesterday was the second, that in extract and intake both ways made it worse so it was better not putting one. The thirs was the extract fan at the rear, replaced the other old Cooler Master with the exact same problem than the other for an old re-oiled Spire with double the RPM which I will remove again because, just as with the CM fans the cleaning and re-oiling just didn't worked. And I used propper machine oil. So as you can see the changes couldn't make it worse, just better and as I said I tested with open case too anyway).

Then yes, both fans are spinning, I checked. Both are in perfect condition. No other software is acting upon the GPU. It's all in stock and I can corroborate that the fan curve is what it should be, I said in the other posts too: the temperature just went up and the RPMs too. The cooling system is simply not working as it should, as it was when I bought the GPU.

By the way the reason I get (almost) the same results with open case as with closed case is that the side panel has some breathing grill (?).


Look at those temps!
Yep, something's not right here...
But hey! I get higher score in the stock setting
Room temps?

if the heatsink were mounted in a wrong way you would have had bad temperature from the beginning.

it make no sense that without any reason temperature get bad suddendly.

can you describe how many fan do you have for intake and exaust?
what are your gpu clock?
can you post gpuz screenshot at idle and sensor after 5 minutes of full load?

 

[Fergutor]

Here are the benchmark screenshot.


View attachment 172778View attachment 172779



if the heatsink were mounted in a wrong way you would have had bad temperature from the beginning.

it make no sense that without any reason temperature get bad suddendly.

can you describe how many fan do you have for intake and exaust?
what are your gpu clock?
can you post gpuz screenshot at idle and sensor after 5 minutes of full load?

View attachment 172781
View attachment 172782

Well it can happen if the mounting or the heatsink is faulty, given it's hanging, right? It's upside down and gravity all the time pulling it. If there was something bad it could ultimately fail in some way. What I mean is that there is a force that is acting on the heatsink trying to separate it from the core. So a weakness could be evident here.

My fans are what I explained in the previous post in small text: one intake, one exhaust, theres an opening in the case which is an old Cooler Master Centurion 534 Plus...but all this doesn't matter given that all the tests I did so far give almost the same result with closed and open case, at least after the GPU started to, in my opinion (although now it's evident), go bad.

Ok I restarted and did the test you asked for, this time with closed case:
IDLE:

LOAD (FURMARK 1.22.0.0):

Here, a screenshot of FURMARK right after I took the provided screenshots of GPU-Z:

 

[MaurizioC]

You are running the card at stock frequencies and it seems to me that the fan speed is correct for 72°.

Mine is 49% at 69° under load but here in Northen Italy is starting cold season.

So the fan rpm variation you reported at the beginning were indeed just the card reacting to temperature increase (I initially thought one fan to be defective and that would have led the other to spin higher).

Honestly now this seems normal to me if your environment temperature is warmer than before.

I do not believe a heatsink can be poorly mounted without being noticeble.

Did your CPU temperature rise too?

 

[Fergutor]

You are running the card at stock frequencies and it seems to me that the fan speed is correct for 72°.

Mine is 49% at 69° under load but here in Northen Italy is starting cold season.

So the fan rpm variation you reported at the beginning were indeed just the card reacting to temperature increase (I initially thought one fan to be defective and that would have led the other to spin higher).

Honestly now this seems normal to me if your environment temperature is warmer than before.

I do not believe a heatsink can be poorly mounted without being noticeble.

Did your CPU temperature rise too?

It isn't warmer. It was until 2 days ago (37ºC or more), but today it got to 14ºC max. The room temperature was about 19 or less when I did that test minutes ago.
Now, sure, the card is reacting to the greater GPU temp but the room temperature didn't increase! That's why I say the card is getting worse: same room temps but gpu variation from then to now is about 8ºC! And faster fan spinning, lower boost clocks too...
No, my CPU stays the same, motherboard CPU VRMs too.
Now I just fired Furmark again and it was 52% at 72ºC. So, with Afterburner, changed the fan to 60% and it got to 71ºC, then to 70%, 70ºC...
About the heatsink, well it kind of moves when I thouch it...not the die baseplate section (if I remember correctly), but the heatpipes and the fins do move...

 

[xman2007]

I bought this MSi GTX 1660 Super Gaming X 2 months ago, which means, here in the south that it was cold in winter.
With low room temperatures, bit less than 20ºC my card got to around the expected 67ºC with furmark and around 1500RPM
More recently, still with low temperatures, but just a bit higher (less than 5ºC more) I noticed 73ºC and 1700RPM or more.
Today with around 27ºC room temperature: 77ºC and more than 2200RPM (don't remember, probably up to 2500RPM).

What, the score? That's not the important part, the temperatue is. Is higher (and fan RPM higher too but that is not showing) than it was just 2 months ago, with lower room temperature too.

You stated 2 months was winter and it was cold, you then stated that it's hotter now/today so it was cold when you got cooler temps and now it's warmer they are higher, unless I'm missing something?

But your performance hasn't dropped off, you are still getting the same score today (warmer) than 2 months ago (colder) albeit with higher fan RPMS but still well within safe max temps for a GPU, have I missed anything?

Anddddd...... you're still running that power virus furmark on your GPU, that's one sure fire way to kill it early

 

[Fergutor]

You stated 2 months was winter and it was cold, you then stated that it's hotter now/today so it was cold when you got cooler temps and now it's warmer they are higher, unless I'm missing something?

But your performance hasn't dropped off, you are still getting the same score today (warmer) than 2 months ago (colder) albeit with higher fan RPMS but still well within safe max temps for a GPU, have I missed anything?

Anddddd...... you're still running that power virus furmark on your GPU, that's one sure fire way to kill it early

Temperatures vary day to day. Today is very cold, yesterday was cold, two days ago was too hot, the day before was brutally hot, all last week was WAY hoter than the week before. This week will be WAY colder than the week before around 18ºC colder in average. We are in spring and summer is ahead... I stated the room temperatures (in estimates). Don't know what's the confusion.

I'm getting the "same" performance in a way, the clocks are lower, the temps are WAY higher, the fan RPM's are WAY higher too AT THE SAME ROOM TEMPERATURES, AND GETTING WORSE.

As I said previously, I run Furmark only for few minutes. You know we play games for half an hour, an entire hour, hours...heavy games that put our GPUS to 99-100% load. Have the video card rules changed I wonder...?

 

[ThirdhandTaters]

Look at those temps!
Yep, something's not right here...
But hey! I get higher score in the stock setting
Room temps?

No, the diff in CPU can't be, at least in the sense that at 3.7Ghz mine isn't putting much heat and neither Furmark or Superposition use the CPU a lot and I have a tower cooler with the airflow to the rear directly to the exhaust fan...hmmm could it be my PSU?!?! Since I bought this GPU I'm getting a funny smell but I tested everything, and even removed the PSU and couldn't pinpoint the origin of that smell...

I can't say with 100% accuracy what the temp is but I am still using my apartment's built-in AC unit set to 70°F, ~21°C, even in the northeast US. I do have 6 case fans so maybe that is what the difference is? as for the PSU my understanding is that if there was something wrong with it it would most likely just stop providing power to the whole computer.

Googling your case it seems like you could use an upgrade. Are you using an ATX motherboard or smaller? Go on newegg and look at cases. I bought a Thermaltake S100 after buying a Rosewill Tyrfing, the Thermaltake had more fan mounts and a better size for my mATX mobo, and it wasn't that expensive either at $70 US not including shipping. I also got a 4 pk of some Cooler Master case fans for $35 US, so for a little over $100 US you could get something that will make sure that your card stays cool even if you decide to overclock some in the future.

Also I noticed that pressing the backplate from the backplate and shroud the temperature decrease significantly, today from 76ºC with open case to 67ºC. And sicerely the card feels filmsy and all parts move when touching them. Like a cheap delicate toy.

That, to me, sounds like the screws came loose while it was being transported from the manufacturing facility because if squeezing the backplate and shroud together, if I'm understanding that correct, reduces the card's temp by 9° that means that there isn't enough contact between the heatsink and the chip. If you're worried about voiding the warranty you can screw the 2 that are diagonal from each other and that could work for now. That could also maybe be a reason for the smell? Is the smell something familiar like burning plastic? I'm not sure what thermal paste smells like so I can't help you there. If it is plastic and you can't find any melted wires then it has to be something enclosed like the PSU or an SSD/HDD. Also, try to take as close of a look as you can at the motherboard, maybe something shorted out in some freak occurrence that wasn't enough to kill the entire computer but could possibly lead to something bad if not looked into.

 

[Fergutor]

I can't say with 100% accuracy what the temp is but I am still using my apartment's built-in AC unit set to 70°F, ~21°C, even in the northeast US. I do have 6 case fans so maybe that is what the difference is? as for the PSU my understanding is that if there was something wrong with it it would most likely just stop providing power to the whole computer.

Googling your case it seems like you could use an upgrade. Are you using an ATX motherboard or smaller? Go on newegg and look at cases. I bought a Thermaltake S100 after buying a Rosewill Tyrfing, the Thermaltake had more fan mounts and a better size for my mATX mobo, and it wasn't that expensive either at $70 US not including shipping. I also got a 4 pk of some Cooler Master case fans for $35 US, so for a little over $100 US you could get something that will make sure that your card stays cool even if you decide to overclock some in the future.



That, to me, sounds like the screws came loose while it was being transported from the manufacturing facility because if squeezing the backplate and shroud together, if I'm understanding that correct, reduces the card's temp by 9° that means that there isn't enough contact between the heatsink and the chip. If you're worried about voiding the warranty you can screw the 2 that are diagonal from each other and that could work for now. That could also maybe be a reason for the smell? Is the smell something familiar like burning plastic? I'm not sure what thermal paste smells like so I can't help you there. If it is plastic and you can't find any melted wires then it has to be something enclosed like the PSU or an SSD/HDD. Also, try to take as close of a look as you can at the motherboard, maybe something shorted out in some freak occurrence that wasn't enough to kill the entire computer but could possibly lead to something bad if not looked into.

Again, same room temps then and today, same computer, totally different results.
No, I'm in Argentina, sadly, and buying here these days is crazy expensive. Not the time (unless I sell some of my old parts, then maybe). But again the rest of the hard isn't the problem as two months ago, everything seemed fine.
Pressing reduced 9ºC that time I said but I'm thinking that was a typo, maybe not. It does get lower but not that much (I think). 4 to 6 ºC depending on the room temp is more realistic. I just tested and got 4ºC less by pressing (from 73 to 69, see? open case same results. Worse even). While in idle, no reduction...
I did try to tighten the other screws ( I mentioned it above), although not in the proper way you mentioned, but they were tight already...which is bad news of course as I thought that was an idea...
Yeah is something like burning plastic or burning dirt maybe too...I did checked and re-checked...for days and frankly can't find anything nor other symptom. Sure it's worrying!
The only other new thing, since 3 weeks ago or so, that I never noticed before is a random faint noise, not a coil whine, it sounds more like a mechanical thing. It appears at random, sometimes few seconds, sometimes minutes. No correlation with what is happening in the computer...I stopped all the fans, stopped all my mechanincal HDDs, and still appears at random. No idea where it comes from.

 

[ThirdhandTaters]

I stopped all the fans, stopped all my mechanincal HDDs, and still appears at random

do you mean you are turning the computer off? If that is what you are talking about and still hearing the noise then it can't be a part of the computer. Take out the power cable and hold the power button for at least 10 seconds to make sure that all the power gets used and see if the sound is still there.

 

[Fergutor]

do you mean you are turning the computer off? If that is what you are talking about and still hearing the noise then it can't be a part of the computer. Take out the power cable and hold the power button for at least 10 seconds to make sure that all the power gets used and see if the sound is still there.

No haha! You can stop the fans safely by pressing them by the center or disconnecting them, the psu fan I simply use a soft thing, like a straw. For the HDDs, configure the energy plan to stop the HDDs in 1 minute, wait 1 minute while not using them and that's it, or you can turn off disconnect them and turn on the PC.

Look, I found something!
I just removed the card and started inspecting it again. So I decided to make a video for all you to see how the damn heatsink moves

In the back there are some rubber stoppers, as in the video it's too dark you can see them here https://www.guru3d.com/articles-pages/msi-geforce-gtx-1660-super-gaming-x-review,4.html
Now I'm not sure now, and I will not do this again ,but it appears even the screws moves! And you can't tighten them more! How is it possible?! The last video card I had that did something like this was an EVGA 6600GT 128MB in 2005!! But that one didn't even had 4 screws! it had 2 (if I remember correctly) and rested on rubber feets! What the heck!?

Then, I made sure it was in a proper position, checked the screws again, moved it a bit and lastly pressed it by the rubber stoppers side and in the middle, installed, run furmark (sorry xman2007) and:

A ouple of degrees less closed case! Open case 68º (with 2 and half minutes, then I closed it, so probably 69º)! Also look at the temp curve is a bit different, and the fans % is a bit less too.

So (I think) it's as I was saying!! The damn thing isn't properly assembled (or whatever)!
I guess that some time I cleaned the machine (here is terribly dusty, epecially with the record drought this year, so cleaning often is a must) I probably touched in a part where it casues the heatsink to move, affecting the thermal paste and how firm the heatsink rest after release it and installing it...

 

[ThirdhandTaters]

I wouldn't go so far as it wasn't properly assembled. Who knows where the card was before you bought it. there might have been bumpy rides in the delivery trucks or someone was careless and dropped it when they were putting it on the shelf waiting to be sold. Contact MSI and see if there is anything that can be done. I know it sucks but if you can get a new card without spending much or anything at all then I think that's a win. Trust me when I say that a company would rather please the customer if it's simple enough than get that customer talking about how bad their products are, and this seems like a simple fix. I had something similar with my last computer and the CPU cooler, it was stock but designed so poorly that trying to take it out to put new thermal paste felt like I could have broken the motherboard and been totally screwed. I had to take out the whole motherboard one time because I couldn't get the cooler back on and couldn't get it off to try and see what was going on. Best of luck Fergutor.