HELP - RX470 Crashes at 61deg .. never 59 or 60 .. only at 61 deg

Forums › Desktop computing › HELP - RX470 Crashes at 61deg .. never 59 or 60 .. only at 61 deg

GoranZ

415 posts

Ultimate Geek
+1 received by user: 57

ID Verified

Trusted

#271769 26-May-2020 14:10

So I got an RX470 4GB Gigabyte G1 Gaming for my sons DELL (it just fit) ..
https://www.gigabyte.com/nz/Graphics-Card/GV-RX470G1-GAMING-4GD#kf

It was ok for a few days, but the fans were really loud and when I spun by hand they felt horrible.
I got new fans from AliExpress and they work well, no noise and solid RPM etc but since I put them on it crashes at 61deg. I put the old fans back but it still crashes at 61deg.

I reflashed vBIOS from Gigabyte Website but didnt help.

If I use AMD settings or MSI afterburner to push fans to about 60% it stays under 61deg and will not crash, can run FurMark for over 2 hours, but drop fans back to 40% or leave them on auto (start at 33% and then got to 50%) and temp hits 61deg and it crashes.

I think the silicon must be OK, RAM passed all the tests I could find .. I'm assuming somewhere internally when it hits 61deg it thinks its hit 95 and shuts down for safety.

Anyone seen this before ?
Could swapping the fans I have done some damage ?

Got any ideas on a solution as the noise from running fans this fast is getting a bit annoying now.

Apsattv

2452 posts

Uber Geek
+1 received by user: 724

#2491878 26-May-2020 14:36

What do you mean when you say crashes?

What are your other system temps?

Is the pc an oem Dell ex office pc? If its shutting down its perhaps too much load on the power supply?

tehgerbil

1110 posts

Uber Geek
+1 received by user: 884

ID Verified

Subscriber

#2491879 26-May-2020 14:39

Did you reapply the heatpaste?

Andib

1395 posts

Uber Geek
+1 received by user: 974

ID Verified

Trusted

#2491883 26-May-2020 14:45

Sounds like the PSU is hitting its power limit and shutting down. What size PSU does the machine have? I know the OEM machines typical have quite low wattage.

<#
.DISCLAIMER
Anything I post is my own and not the views of my past/present/future employer.
#>

GoranZ

415 posts

Ultimate Geek
+1 received by user: 57

ID Verified

Trusted

#2491884 26-May-2020 14:46

Cheers for taking an intereset .. Answers ..

Yep I redid thermal paste .. done this about 25 times with no issue in my life and use non conductive stuff, but maybe I did something

When it "crashes" fans ramp up to 100% and then PC reboots .. when back in windows its all OK until 61deg again.

Other temps are in check .. CPU is never more than ~55deg (I5-4460) .. put a new 550W PSU in the DELL too ..

If I put it in my Ryzen 2700X system it does the same thing at the same temp. My own GTX1070 runs fine in the DELL so confident its not the other components.

Some googling and there was suggestion the VRM temps might be getting out of hand or a VRM temp sensor causing the shutdown but not sure how to test for that.

SpartanVXL

1498 posts

Uber Geek
+1 received by user: 666

#2491886 26-May-2020 14:48

I know the recent AMD cards have two temp readings, core and tjunc. It mostly relates to I think Vega cards but is it possible your taking the heatsink off to replace fans isn’t making proper contact? (Related to the washer mod).

Edit: on that note you could try underclocking the card slightly. If it is a memory module as some review have mentioned (too close to the core) then lowering memory clock a bit might help. Undervolting or lowering power limit too might help with thermals.

GoranZ

415 posts

Ultimate Geek
+1 received by user: 57

ID Verified

Trusted

#2491892 26-May-2020 15:08

SpartanVXL: I know the recent AMD cards have two temp readings, core and tjunc. It mostly relates to I think Vega cards but is it possible your taking the heatsink off to replace fans isn’t making proper contact? (Related to the washer mod).

Edit: on that note you could try underclocking the card slightly. If it is a memory module as some review have mentioned (too close to the core) then lowering memory clock a bit might help. Undervolting or lowering power limit too might help with thermals.

Will have a play around .. the fact I can just blast it with air means its got to be a temp issue ... might remove backplate and see if that helps as the GDDR6 is flip chip so hottest part is on the back.

Geekzone

Want to support Geekzone and browse the site without the ads? Subscribe to Geekzone now (monthly, annual and lifetime options).

Apsattv

2452 posts

Uber Geek
+1 received by user: 724

#2491906 26-May-2020 15:39

Did you buy this card new? if so send it straight back if still under warrenty.

If you bought it used there is a chance it was used flatout for coin mining.. 61C is nothing in heat terms for that card and you have run it in another pc with the same issue occuring. Which probably rules out a windows issue. What about the driver version? are you running Adrenalin 2020 Edition 20.4.2 ? AMD has had a lot of issues with their drivers

Some users of 470's have flashed the bios from the 570/580' on them.. Hopefully it hasn't been tampered with in that way if you got it used?

https://www.guru3d.com/news-story/wanna-flash-your-rx-470-towards-a-570-it%E2%80%99s-possible.html

GoranZ

415 posts

Ultimate Geek
+1 received by user: 57

ID Verified

Trusted

#2491910 26-May-2020 15:50

Yeah 2nd hand and 100% must have been mined but I have flashed it with stock vBIOS .. using whatever the latest drivers were at amd.com

gzt

18679 posts

Uber Geek
+1 received by user: 7809

Lifetime subscriber

#2491950 26-May-2020 15:56

No idea. I'd want to know exactly where the sensor is and check in that area first.

gzt

18679 posts

Uber Geek
+1 received by user: 7809

Lifetime subscriber

#2491952 26-May-2020 15:58

Maybe some utilities to confirm clocks are running in spec.

Apsattv

2452 posts

Uber Geek
+1 received by user: 724

#2491954 26-May-2020 16:00

GoranZ:

Yeah 2nd hand and 100% must have been mined but I have flashed it with stock vBIOS .. using whatever the latest drivers were at amd.com

Which stock bios did you flash it with? there is more than one available

https://www.gigabyte.com/nz/Graphics-Card/GV-RX470G1-GAMING-4GD/support#support-dl-bios

also i note this in the features

3D Active Fan with LED Indicators

The semi-passive fans will remain off when the GPU is under a set loading or temperature for low power gaming. The LED fan indicator on the top of the graphics card provides an instant display of the fan status

New World

Shop on-line at New World now for your groceries (affiliate link).

GoranZ

415 posts

Ultimate Geek
+1 received by user: 57

ID Verified

Trusted

#2491971 26-May-2020 16:58

Thanks everyone for your input .. I might have fixed it.

Just to answer the questions, the BIOS's I used were the F3 from techpowerup and the F6 from Gigabyte.
Both worked fine and GPU-Z said it was in spec plus the driver is 2020 edition

But to the solutioon .. maybee ..

When I repasted and swapped the fans I did not remove the backplate as you didnt need to.
I have removed it this afternoon based on some googling and advice from here and there was what appeared to be a large water stain right in the VRM area. I have cleaned it off with CRC contact cleaner, and left the backplate off.

Just popped the card back in and running a stress tes, after 6 min it hit 64deg and no crash, the fans then ramped up a bit more, so it droppped back into the 50's and now the card has been running for 22 min with no crash.

Temps ranging from 57 to 69 .. fans going from 33% to 50% based on load .. as they should and core closks and FPS are as expected

So if anyone finds this thread, or has asimilar issue this might help point them in the right direction

I'll keep running tests and let the son play some fortnite and report back if we have any crashes .. but its already a lot better.