[SOLVED] RX480 (580 Bios) Strange problem.

[GamerMove69]

Hello everyone, first time posting in this forum, sorry for common mistakes in advance.

First ill start with my PC spec's.

CPU: Ryzen 5 2600x
GPU: RX 480 8Gb (580 bios installed) Overclock was 1450mhz core voltage 1.175v (now i'm running stock voltage because of the problem) 2100mhz memory clock.
RAM: Kingston HyperX 2x8Gb DDR4 2x1600mhz.
PSU: Corsair TX750M
MB: Asus prime x470 - pro (latest bios)
SSD: Adata 450gb
HDD: Western digital 500gb.

Ok, ill start with history of RX480 8Gb (580bios), I had this card for 2-3 years now, I bought it used for 200eur I think.
It already had 580 bios flashed when I bought it, I never tried to change it.
Basically I use this card for gaming, I started overclocking it when I bought it, I remember the temps for 1450mhz clock were around 65-75c full load, it wasn't throttling or anything like that, though it had some hard crashes, black screen, artifacts etc, when overclock was unstable (was still trying to nail the sweet spot for OC).
After a year or two it started acting up, the clock speed was dropping though the temperature's seemed acceptable. So I changed the thermal paste for it, and the thermal throttling was gone I guess.

Now a year passed it started doing it again out of nowhere, and my temperature's was showing 56-64c max, I thought that was very strange, so this time I changed thermal pads and I changed thermal paste again. But the problem still persisted, I thought maybe I applied too little thermal paste or thermal pads weren't the right size/thickness.

Anyway I wanna cut the actual problem, basically I push down the GPU (give more sag to it) while it's running benchmark, the temperature shoot's from 56-60c to 70-90c but when the card is pushed down I don't get the thermal throttling which of course lowers my clock speed because I think the fans actually kick in at the right speed which it supposed to be, it even rises the temperature when I push the card on idle by 2-3c. So I cleaned off GPU'S contact's which go to motherboard (PCIE I think). Still no luck. I tried different GPU slot in motherboard, same problem.

Basically what I figured out is that the PC/GPU is giving wrong temperature readings if it's not pushed down a bit (sag's more). There for the GPU thinks it's running at good temperature and fans spin at average speed, but the thermal throttling still happens maybe because there is built-in GPU thermal protection in GPU itself.

I wanna know how does sagging/pushing the Gpu down can affect these readings, is it because of bad contacts and is it fixable?

Sorry for this mess that I wrote, I'm kinda in a hurry, ill try to edit a bit when I have more time and maybe add a video if I can.

EDIT1: Just tried putting it in another GPU slot, same problem.

EDIT2: I have repasted the card for the fourth time this month, but this time I putted more amount of thermal paste like never before (like really a lot in my opinion), ran the benchmark, OC clock speed holding up for couple of passes, temperature is 58-60c but only for half an hour of full load, fan speed was around 60-70%. I hope it fixed the problem, ill try to game a bit with this OC and see if the throttling is going to come back....

Youtube link, I hope it works.

View: https://www.youtube.com/watch?v=KsnXZwmB6jM



Thank you for responses and help in advance!

 

[ohio_buckeye]

If it's sagging, depending how your case is set up, maybe use something like this to try to limp it along until you can get a better card. You basically set it up so that metal finger pushes up on the end or side of the card to try to support it a bit.

https://www.ebay.com/itm/383920181526?epid=10039373745&hash=item59636d1116:g:vv8AAOSwQdhgD~jM

The resaon your card probably overheats when sagging is because the heatsink and thermal pads are probably not making solid contact like they would if the card weren't sagging. Also, if you set the PC on it's side with the card facing upwards, you might see you don't as much sagging on your gpu due to gravity.

 

[GamerMove69]

If it's sagging, depending how your case is set up, maybe use something like this to try to limp it along until you can get a better card. You basically set it up so that metal finger pushes up on the end or side of the card to try to support it a bit.

https://www.ebay.com/itm/383920181526?epid=10039373745&hash=item59636d1116:g:vv8AAOSwQdhgD~jM

The resaon your card probably overheats when sagging is because the heatsink and thermal pads are probably not making solid contact like they would if the card weren't sagging. Also, if you set the PC on it's side with the card facing upwards, you might see you don't as much sagging on your gpu due to gravity.

Thank you for the fast response!

Yeah you got it right I think, It's just the other way around I need more sag for the card not to overheat and throttle, when it's not sagging (not pushed down in one side) it overheats and throttles. I changed thermal pads, made sure they make good contact. And the force I use to push it down is very little and wouldn't affect the thermal pad connection, it's like something with GPU contacts that go into the motherboard is wrong. It's giving me bad temperature readings when it's not sagging, that's why it throttles on low temperature's. I think....

Thanks again for the help!

 

[spentshells]

Thank you for the fast response!

Yeah you got it right I think, It's just the other way around I need more sag for the card not to overheat and throttle, when it's not sagging (not pushed down in one side) it overheats and throttles. I changed thermal pads, made sure they make good contact. And the force I use to push it down is very little and wouldn't affect the thermal pad connection, it's like something with GPU contacts that go into the motherboard is wrong. It's giving me bad temperature readings when it's not sagging, that's why it throttles on low temperature's. I think....

Thanks again for the help!

Well if you need to push down on the card to cause sag and get normal temps... simply put you are tightening the heat sink down to far. Snug but not tight. tighten all your screws as usual, then loosen them a full turn. also if you feel you are using a lot of thermal paste, then you are and for gpu's this can be problematic.

Overclocking after a vbios flash, that's sorta gutsy. My RX 570 nitro+ is sorta a bum when it comes to oc.... considering all the other very substantial oc's I've seen with this model.

So yeah in closing you are tightening the heat sink to tight.

 

[GamerMove69]

Okay. so throttling started again after applying more thermal paste, I was using Cooler Master High Performance thermal paste which was kind of not thick enough I think it was white colored too, so I bought Arctic MX 4 it seemed better and thicker and was gray, I applied that to the GPU and clock speeds started holding good after 1-2 hours of intensive gameplay, temperature's didn't exceed 61c. So maybe it was the quality of thermal paste I don't know for sure. I will try to game more a bit after work to see if the throttling is gone for sure. But yeah... maybe the thermal paste thickness wasn't enough. Thank you all for responses and advices!

 

[ohio_buckeye]

Also, make sure you have enough airflow around the card.

Here's a cheap way to do that.

https://www.ebay.com/itm/224519417747

It's a bracket that you can slot 2 120mm case fans into. I actually ordered this and couple of 120mm fans. So I'm going to put the fans into that and slot it in below my new card just to keep air flowing well. For the few dollars spent it's a good investment I think.