Jump to content



Photo

Was I sold a broken AMD Radeon R9 280X?

Answered Go to the full post

  • Please log in to reply
13 replies to this topic

#1 Yusuf M.

Yusuf M.

  • Tech Issues Solved: 1
  • Joined: 25-May 04
  • Location: Toronto, ON
  • OS: Windows 8.1 Pro
  • Phone: OnePlus One 64GB

Posted 17 June 2014 - 05:06

I recently bought two Gigabyte Radeon R9 280X 3GB WINDFORCE 3X OC (rev 2.0) cards from different sellers on eBay. The first R9 280X works great and I haven't come across any major issues. The second card isn't working properly. Whenever I try to play a DirectX title, the game crashes to the desktop or freezes and forces me to reboot. So far, I've tried Battlefield 4, Crysis 3 and Metro: Last Light. None of those games work and if it crashes to the desktop, I'm met with an error along the lines of "DXGI_ERROR_DEVICE_REMOVED" (Crysis 3) or "DXGI_ERROR_DEVICE_HUNG" (Battlefield 4)".

 

I took a screenshot of both cards in GPU-Z:

 

R9 280X #1 (works):

 

NFrre60.png

 

 

R9 280X #2 (doesn't work):

 

JrdAoPx.png

 

I noticed that the memory type has a different name: Elpida instead of Hynix. And the BIOS version seems to be older. Is there anything I can do try to fix the card? I've uninstalled and reinstalled my drivers properly. I even ran a driver cleaner program to remove any leftover registry entries. Also, I've ruled out my PSU after I tried it in my brother's PC.



Best Answer Yusuf M. , 17 June 2014 - 07:56

If it's been undervolted then they've modified the BIOS. Modifying the BIOS to something custom is really daft and could have caused other problems especially if they've not taken the different specifications of the memory etc. into consideration. Don't count on that card lasting long, it might just suddenly die completely out of the blue.

I figured the BIOS was modified because I wasn't able to change the voltage with OC Guru II. Anyway, I found a solution to my problem. I re-flashed the latest BIOS for the Elipda card (F60) and it's back to a max voltage of 1.200v. It's running fine but I'm going to switch back to the Hynix card after a couple of days of gaming. If it runs fine, I'll put an ad up for it on Kijiji. At the moment, I don't need two video cards and I'd rather keep the one that performs better.

Go to the full post



#2 Scorpian8867

Scorpian8867

    Neowinian

  • Joined: 04-May 09
  • OS: 8 x64
  • Phone: Samsung Galaxy SIII

Posted 17 June 2014 - 05:37

You can try switching to the secondary BIOS on the card. There should be a little switch next to crossfire fingers on the top. Just make sure the PC is off, or it won't take effect until you restart.



#3 Raa

Raa

    Resident president

  • Tech Issues Solved: 5
  • Joined: 03-April 02
  • Location: NSW, Australia

Posted 17 June 2014 - 05:55

It might be worth having a look for a bios update, but be careful they have different versions for different memory types (at least Gigabyte has done this to me before with nVidia cards [recommended]).

 

Failing that, it does sound like it's a dodgy card...



#4 Axel

Axel

    --[Est. 1986]--

  • Joined: 05-August 03
  • Location: Milton Keynes, UK

Posted 17 June 2014 - 05:58

Highly likely the card was ragged for mining! Have you checked if it is overheating?

#5 OP Yusuf M.

Yusuf M.

  • Tech Issues Solved: 1
  • Joined: 25-May 04
  • Location: Toronto, ON
  • OS: Windows 8.1 Pro
  • Phone: OnePlus One 64GB

Posted 17 June 2014 - 06:55

Thanks for the replies, guys. I really appreciate it.

 

You can try switching to the secondary BIOS on the card. There should be a little switch next to crossfire fingers on the top. Just make sure the PC is off, or it won't take effect until you restart.

Thank you! That seems to have done the trick.

 

BIOS switch position 1:

 

kvZKm11.png

 

BIOS switch position 2:

 

7n2WooF.png

 

The difference between the two is the max voltage. It went from 1.075v to 1.200v. I swapped it out with the first card and changing the position of the switch didn't change the max voltage. I guess that's the difference between Elpida GDDR5 memory and Hynix. From what I've read online, Hynix is better. And the benchmarks I've run so far seem to prove that. I get about 2 FPS more with the Hynix card. It's not a big difference but it's also more stable. Performance seems to be more jumpy with the Elpida card.

 

It might be worth having a look for a bios update, but be careful they have different versions for different memory types (at least Gigabyte has done this to me before with nVidia cards [recommended]).

 

Failing that, it does sound like it's a dodgy card...

That's the first thing I did after ruling out drivers and my PSU. I'm already using the latest available version for the card that didn't work (but works now with the physical BIOS switch in position 2).

 

Highly likely the card was ragged for mining! Have you checked if it is overheating?

According to the seller, the card wasn't used for more than a week. The first card was used for mining for 2-3 months and that one ran fine. As for temperatures, both cards seem to hit a max of 71C.

 

So, any ideas as to why I can't run it at 1.075v max? If my 600W PSU is the reason, then I'll upgrade it.



#6 John.D

John.D

    Neowinian

  • Tech Issues Solved: 2
  • Joined: 10-November 09

Posted 17 June 2014 - 06:59

Looks like that card had a few probs

 

And here saying update the BIOS on it



#7 Soldiers33

Soldiers33

    Neowinian Senior

  • Joined: 01-September 06
  • Location: London
  • OS: Windows 7 Professional

Posted 17 June 2014 - 07:17

Unless you doing graphic intensive stuff i dont think your cards should be heating up to that temp. Havr you tried clearing them out or perhaps getting coolers for them? I normally use thermalright coolers drop temps by at least 20 degrees.



#8 Raa

Raa

    Resident president

  • Tech Issues Solved: 5
  • Joined: 03-April 02
  • Location: NSW, Australia

Posted 17 June 2014 - 07:19

If you've switched it to BIOS 2 and it works, but not on BIOS 1, while it works and that's great - it's definitely a faulty card.



#9 OP Yusuf M.

Yusuf M.

  • Tech Issues Solved: 1
  • Joined: 25-May 04
  • Location: Toronto, ON
  • OS: Windows 8.1 Pro
  • Phone: OnePlus One 64GB

Posted 17 June 2014 - 07:28

I saw the first link before I made this thread and it turns out, both cards have the latest BIOS. F70 for the Hynix card and F60 for the Elpida card.

 

Unless you doing graphic intensive stuff i dont think your cards should be heating up to that temp. Havr you tried clearing them out or perhaps getting coolers for them? I normally use thermalright coolers drop temps by at least 20 degrees.

The 71C temperature I mentioned was under heavy load (e.g. FurMark, Unigine Valley benchmark, etc). It's in line with what reviewers have reported. The WINDFORCE 3X cooler by Gigabyte does a good job at keeping the card cool. It's better than my friend's ASUS R9 280X DirectCU II TOP.

 

If you've switched it to BIOS 2 and it works, but not on BIOS 1, while it works and that's great - it's definitely a faulty card.

I've run a lot of tests on it and it's working fine for now. It seems to have been undervolted to 1.075v for better cryptocurrency mining efficiency. I'm wondering if there's a way to return it to 1.200v.



#10 n_K

n_K

    Neowinian Senior

  • Tech Issues Solved: 3
  • Joined: 19-March 06
  • Location: here.
  • OS: FreeDOS
  • Phone: Nokia 3315

Posted 17 June 2014 - 07:47

If it's been undervolted then they've modified the BIOS. Modifying the BIOS to something custom is really daft and could have caused other problems especially if they've not taken the different specifications of the memory etc. into consideration. Don't count on that card lasting long, it might just suddenly die completely out of the blue.



#11 OP Yusuf M.

Yusuf M.

  • Tech Issues Solved: 1
  • Joined: 25-May 04
  • Location: Toronto, ON
  • OS: Windows 8.1 Pro
  • Phone: OnePlus One 64GB

Posted 17 June 2014 - 07:56   Best Answer

If it's been undervolted then they've modified the BIOS. Modifying the BIOS to something custom is really daft and could have caused other problems especially if they've not taken the different specifications of the memory etc. into consideration. Don't count on that card lasting long, it might just suddenly die completely out of the blue.

I figured the BIOS was modified because I wasn't able to change the voltage with OC Guru II. Anyway, I found a solution to my problem. I re-flashed the latest BIOS for the Elipda card (F60) and it's back to a max voltage of 1.200v. It's running fine but I'm going to switch back to the Hynix card after a couple of days of gaming. If it runs fine, I'll put an ad up for it on Kijiji. At the moment, I don't need two video cards and I'd rather keep the one that performs better.



#12 Raa

Raa

    Resident president

  • Tech Issues Solved: 5
  • Joined: 03-April 02
  • Location: NSW, Australia

Posted 17 June 2014 - 08:11

<snip>  I'll put an ad up for it on Kijiji. At the moment, I don't need two video cards and I'd rather keep the one that performs better.

Well then, that was a fun exercise. :laugh:



#13 OP Yusuf M.

Yusuf M.

  • Tech Issues Solved: 1
  • Joined: 25-May 04
  • Location: Toronto, ON
  • OS: Windows 8.1 Pro
  • Phone: OnePlus One 64GB

Posted 17 June 2014 - 08:22

Well then, that was a fun exercise. :laugh:

My original plan was to sell one card while keeping the seemingly newer one. After this "fun exercise", I'll most likely sell the previously-undervolted card.

 

1Y7B3zt.png



#14 tim_s

tim_s

    Default

  • Joined: 07-January 13
  • OS: OSX (Macbook Pro i7), Windows 7 (Gaming), Gentoo
  • Phone: Samsung Galaxy SIII, iPhone 4s

Posted 20 June 2014 - 01:14

Great job everyone - seems the culprit was the librarian in the kitchen with a chainsaw.  This concludes our Graphics card mystery.

 

.

.

.

.

 

No ... really glad you got it sorted and I would sell the card for more than you paid and invest in another without issues.