PC Graphics

[grantelb4rt]

Hello Guys and Girls,

is it possible for my Graphicscard to just have suddenly died? I was casually browsing the Internet, suddenly both my Monitors went dark, the PC shut down and restarted itself. First thought was a Powerloss, but the Monitors and all other Electronics were still running. After that there was no more Video signal from my R9 Fury Nitro. But I heard the Windows welcome sound, so I plugged my Monitor into the onboard GPU where I got a Video Signal.

In the Device-Manager the only Videocard shown was the Intel iGPU one. Same with the BIOS, the Mainboard did not even recognize that there was a Card installed in the PCI-E Slot.

So naturally i tried the following:

1. Reseating Power Cables and Videocard
2. clear CMOS
3. flash new BIOS
4. switching V-BIOS (BIOS Button on Card)
5. second PCI-E slot
6. different Video output
7. only Keyboard plugged into PC
8. checking PSU Voltage on PCI-E Connectors (was at 12.17V)

But all of those things didnt have any success.

The GPU in a different PC yields the same results. A different Videocard is recognized, so the Mainboard and CPU are not at Fault.

I had issues with the DP Video out since I had the Card. Could this be a late effect of that Problem? (Video of the Issue: Sapphire R9 Fury Nitro Screen Flickering - YouTube )

Anybody else had similar Problems with any Graphicscard? Really would be a shame if that card died on me after 1.5 years, especially considering the current Videocard market...

I have the feeling that there is nothing I can do, but I thought I might as well ask for some other Opinions

System specs:

• i5 3570k
• Asus P8Z77-V
• 4x4GB 1333Mhz DDR3
• Superflower Amptec 750W
• Windows 10 64bit
• Benq GW2765

[elstaci]

Sounds like your GPU card Overheated. When a Computer shuts down like you pressed the "restart" button on your computer case. It generally indicates either something overheated (CPU or GPU) or the PSU (Weak or failing ) didn't give enough power to a component shutting down the computer.

If you tried all the Output Ports on the card with different cables and on another PC with the same results, then most likely your card got damaged when it overheated. Maybe the Fan stopped working (For whatever reason, jammed or burned out) at the moment the GPU was under heavy load and temperature spiked suddenly extremely high before giving the GPU a chance to throttle or shut down the computer. (My hypothesis)

Most Manufacturer's give more than a year of warranty for GPU cards. I have a Asus GPU card with a three year Warranty. Contact the GPU card Manufacturer and open a RMA with Support. They will let you know if it is under Warranty for what happened.

Went to Sapphire Support. It doesn't mention Warranty length. But under Amazon.com it does . Supposingly it has a TWO YEAR Warranty according to this PDF from AMAZON from SAPPHIRE: https://images-na.ssl-images-amazon.com/images/I/61sYFfVoR0S.pdf  and this is the only information Sapphire Support gives about RMA/WARRANTY :http://www.sapphiretech.com/support.asp?lang=eng

[grantelb4rt]

I dont think it overheated, I was only browsing the Web... The cooler on that card is so massive that the Fans are stopped by Factory on low loads and Temperatures below 60° (or something around that area).

PSU could be another thing, but I dont have another PSU laying around with 2x 8-Pin PCI-E connectors

I already contacted Sapphire, they just got back to me. They said they wont do the RMA and want me to send the card to the Store where I bought it. The Store told me that they would send that card in for Repair, but if they cant repair it I would get the Time Value of the card... Which is only 150€, from that money I cant get anything close to a Fury performance wise

[plumboby]

How old is your PSU might be an idea if you have pulled your card trying to get a RMA to maybe just invest in a brand new PSU for future proofing & a safe guard incase it was the PSU that cooked the card. IS your mobo firing up on intel chipset drivers if it is that's a bonus. Tho if something has died or PSU is failing I would advise in a upgrade for good measure if its not you card & gives a chance for a new card if they honour by giving you a new gen card.

[grantelb4rt]

Sooo, I bought a new PSU to be sure that that is not the Problem. Still no change. Seems that my Card is completely dead. I guess I will send it in, in hope of being repairable. If its not repairable I will need to save up a few Months to buy a new card with the prices right now.

[elstaci]

very poor Warranty by Sapphire. Most Warranties, at least here in the US, either Repairs or Replaces the defective card. Granted the replacement Card would probably be a Refurbished one but at least you will have a working card again.

I would look at Asus or some other Manufacturer with a better Warranty than Sapphire. Asus has a three year Warranty, at least here in the US. You need to research before you purchase a new card what the Warranty is. Visiontek has a limited Lifetime Warranty when I purchased a AMD HD 7850 card from them about five years ago. In Europe, Warranties are probably different than here in the States since you posted the cash in Euro and not dollars.

[grantelb4rt]

My thoughts exactly. Will be the last Sapphire card I ever buy.

Fury was AMDs top of the Line until Vega came out, which was not even a year ago when I remember correctly... Having AMDs most high end Card of the "last" Generation, and getting no Support at all is massively disappointing.

[hardcoregames_]

grantelb4rt wrote:

Hello Guys and Girls,

is it possible for my Graphicscard to just have suddenly died? I was casually browsing the Internet, suddenly both my Monitors went dark, the PC shut down and restarted itself. First thought was a Powerloss, but the Monitors and all other Electronics were still running. After that there was no more Video signal from my R9 Fury Nitro. But I heard the Windows welcome sound, so I plugged my Monitor into the onboard GPU where I got a Video Signal.

In the Device-Manager the only Videocard shown was the Intel iGPU one. Same with the BIOS, the Mainboard did not even recognize that there was a Card installed in the PCI-E Slot.

So naturally i tried the following:

1. Reseating Power Cables and Videocard
2. clear CMOS
3. flash new BIOS
4. switching V-BIOS (BIOS Button on Card)
5. second PCI-E slot
6. different Video output
7. only Keyboard plugged into PC
8. checking PSU Voltage on PCI-E Connectors (was at 12.17V)

But all of those things didnt have any success.

The GPU in a different PC yields the same results. A different Videocard is recognized, so the Mainboard and CPU are not at Fault.

I had issues with the DP Video out since I had the Card. Could this be a late effect of that Problem? (Video of the Issue: Sapphire R9 Fury Nitro Screen Flickering - YouTube )

Anybody else had similar Problems with any Graphicscard? Really would be a shame if that card died on me after 1.5 years, especially considering the current Videocard market...

I have the feeling that there is nothing I can do, but I thought I might as well ask for some other Opinions

System specs:

• i5 3570k
• Asus P8Z77-V
• 4x4GB 1333Mhz DDR3
• Superflower Amptec 750W
• Windows 10 64bit
• Benq GW2765

Time to break open the piggy bank and spend some $ on a new video card

sucks that digital coin miners are hogging all the video cards too

[grantelb4rt]

Sadly my Piggybank is as empty as it can be.

Also I dont see a point in spending more money on a slower card now, then I did almost 2 years ago.

[elstaci]

This User had something similar happen to his Sapphire R9 card in this thread : My sapphire r390 just die after drivers update

[grantelb4rt]

Spontaneous combustion of Sapphire Cards. Nice.

I will just wait until the GPU prices are bearable again.

[hardcoregames_]

grantelb4rt wrote:

Spontaneous combustion of Sapphire Cards. Nice.

I will just wait until the GPU prices are bearable again.

That might be a while, digital coin enthusiasts have hogged the cards by the 6 pack

[elstaci]

How can these Miners afford to spend several thousand dollars on GPU cards, Risers, and USB cables and God knows what other computer components they require to run the Mining Computer setup.

I barely could afford a top end GPU card.

[hardcoregames_]

I have been watching market and evidently a lot of speculators are using credit cards to buy cards galore and gambling on selling digital coins for more than they invested

the problem is that the digital coins value is very fickle and I avoid them like ebola

[pokester]

That's why the saying "It takes money to make money" exists for many I think it is nothing more that a rich guys hobby.

[hardcoregames_]

pokester wrote:

That's why the saying "It takes money to make money" exists for many I think it is nothing more that a rich guys hobby.

unfortunately that is true with the gambling over digital coins which has brought countless new buyers into the video card markery

[grantelb4rt]

I know..

Thats why I wont be a gamer for the next few Months, or even years, eventhough they "care about the gamers" so much

[plumboby]

Only reason why the Miners go for AMD cards for the hash rates to cost of what they can mine effectively unfortunately AMD policy with Open source support has made them Mining friendly with drivers which are supported. What AMD needs to do is needs to make up to the Gamers, enthusiests, users & community of gamers a limit or build a specific card only for mining.Never know what might have with Vega /Navi  suppose to be a few being made & tested on 7nm end of year early next. Hopefully a game changer  hopefully they go with Gddr6 ram with 12 G vram  equivalent in my eys give the 2080 Nvidia is going with on Gddr6 ram. If your a gamer save up as they probly be releasing the the 7nm Vega/Navi   to a simular time as Nvdias new chips.

[colesdav]

Hi,

(1). I just had a failure on an R9 Fury X - looks like the fan controller was broken / possibly modded in that case, however I had just powered down my PC and changed the DisplayPort output from DisplayPort2 to Display Port 3 before booting up and the card failed.

Request for help to identify a component on an R9 Fury X please.
Symptoms are similar, namely no display output at all, and the card is not seen at all by the PC.

(2). Although it may seem like your card has gone to GPU heaven, there may still be hope.
I had an apparently broken Sapphire HD7970 OC 6GB that I managed to bring back to life from total zombie state.
You can see how I fixed it and the parts I used to do it at the end of the following thread here:
HD7970 not working after 17.4.3 update
The card is still running great, every day used for blender or gaming since I repaired it.

For you, (2) above, might be worth a try to fix your own card rather than taking the risk and hassle of buying a second hand card and
going through the problems I had to get my money back in (1) above.

I am looking at saving up for a new GPU, I think it very likely I have to go Nvidia for my next card.
I have been purchasing second hand cards for years from various sites without any problems until now.
I think purchasing any second hand high end GPU right now is riskier than it used to be because of mining.
If you do go for a second hand card beware anything that looks like it has had the fans modded or replaced in any way.
Fan MTBF means the fans should last a card that has been used for gaming for years. Mining 24/7 will likely kill the fans or fan controller early.
Note a second hand card can look in perfect external physical condition but still the components on the PCB can be simply worn out or degraded electrically.

I forgot to mention - if you do remove the Sapphire R9 Fury Nitro be very careful not to touch the interposer between the HBM memory and the GPU itself.
I will look for an article showing a teardown of your card.

Good luck if you attempt to fix your card using Solder Reflow described in (2). above.

Bye.

[colesdav]

Here are a couple of videos showing how to tear down your card and discussing the functions of various parts of that card:

Teardown SAPPHIRE R9 FURY NITRO - YouTube

PCB Breakdown: Sapphire R9 Fury Nitro - YouTube

Bye.

[colesdav]

Hi,

Looking at the second video above, Sapphire  have used electrolytic capacitors on the VRM of the Sapphire R9 Fury Nitro unlike the caps used on the FuryX.
You might want to inspect those electrolytic capacitors and check if any of them look "blown" or if any of them look like they are leaking fluid.
It should be easy to find new electrolytics and replace them.
One failure mode for those caps can start out as intermittent shorting / medium resistance between the capacitor terminals before catastrophic failure (permanent short). This results in voltage glitching on the VRM DC supply line. I looked at your video showing the glitching video display. That *might* indicate what was going on.

Bye.

[grantelb4rt]

Thank you very much for your huge amount of Information. At the point where you wrote all of that my card was already send in for repair. If this is really the Problem, they might be able to find and repair it.

If not I will have a look at the card myself and try to find the Issue on my own.

But the fact that they cheaped out on the capacitors is kind of disappointing..

[colesdav]

Hi,

Thank you.

I am sorry I replied to your thread so late.

As you might see I was/am having my own problems with sourcing good R9 Nanos/Fury/FuryX's for a project I am working on.

Please forward the information to the people repairing the GPU for you. It might help.

Note if the GPU had dry / failing solder joints to the PCB then that could also have been the cause of the glitching on your card.

If the people doing the repair are a professional outfit they should be able to remove the GPU IC, remove the old and apply new solder balls underneath the IC, and refit the GPU properly, not just use the Solder Reflow method I described. They should have a proper solder Surface Mount Reflow / Rework Station if they are repairing GPU's. Note that just heating the GPU die during Solder Reflow / Rework, might also repair the GPU temporarily.

They might tell you it is not worth the cost, but it should only take 1 hour of work to do it.

Electrolytic capacitors  are often used because they have higher capacitance than available on tantalum capacitors, for the same cost.
More capacitance will generally mean less ripple on the VRM DC supply and more stable clocks and better GPU performance, provided they use a smaller tantalum capacitor in parallel with the electrolytic capacitor. I don't know if they did that on your card. So Sapphire were not necessarily "cheaped out". 
They generally make good cards. Those Sapphire Fury Nitros are only a little slower than an R9 FuryX in gaming, even though they have a cut down Fiji GPU.

Bye.

[hardcoregames_]

I have not run into bad capacitors for quite a while since the bad electrolyte problem. That was problem way back in 2003, this is 2018.

TEAPO CAPACITORS – HARDCORE GAMES™

[colesdav]

Thanks for the information. I know this is 2018.

I have seen blown electrolytics on motherboards and cards manufactured as recently as 2-3 years ago.
It still happens.
Cheers.

[hardcoregames_]

colesdav wrote:

Thanks for the information. I know this is 2018.

I have seen blown electrolytics on motherboards and cards manufactured as recently as 2-3 years ago.
It still happens.
Cheers.

I can imagine that the flakey caps are so pervasive that discovery may be ongoing for another decade or more

capacitors are a very mature technology, so when something like this happens the damage can be so pervasive it defies imagination

I updated my article a few times, when I found Dell had to plink down over $400 million for bad motherboards, ouch

[colesdav]

Hi,

Interesting article about the capacitor plague, I remember it was not good.

New, high quality electrolytic capacitors used on GPU's should have a life span at least 6 years at rated temps, but the failure rate does increase quite quickly with even a 10'C temperature rise. So maybe people who overclock their cards or OC edition of cards might hit a problem early.
I agree that the flakey capacitors that caused the problem form 2003-2007 may well still be around.
Alternatively sometimes components with manufacturing defects do get through and make it onto cards anyhow.

Bye.

[pokester]

Can't speak to this issue specifically but I have always found it crazy that they just don't use the best capacitors available to begin with unless it really is planned failure. The difference in price is minimal, likely wouldn't make even a $5 difference in price. I have fixed many Motherboards over the years replacing the capacitors but these multilayered pcbs today are getting a bit beyond what I can do at home.

[hardcoregames_]

due to competition, motherboard makers have to get the lowest cost parts they can

this is how bad competition has become

[elstaci]

Do you believe that applies to the high-end Motherboards? I have an Asus TUF Sabertooth 990FX which supposed to be made up to Military specs in components used. It has a Five year Warranty. I have had absolutely no issues with this Motherboard even though it is completely loaded with discrete components and USB connections.

Asus wouldn't give a Five year Warranty on a Motherboard if they used lowest cost components that may fail within the Five year Warranty range.

[colesdav]

Hi

You will likely be fine, however you might be unlucky and a bad component could slip through QA despite all attempts to catch them during testing.
Regarding RMA ... I have no experience with having to send any motherboard back to ASUS within the Warranty Period.
Oldest ASUS MB's I am running at present are Z87 Deluxe Dual's. They had a 3 year warranty and are still fine after ~ 4.5 years of use.  

RE: Wouldn't give a Five year Warranty on a Motherboard if they used lowest cost components that may fail within the Five year Warranty range.

Note I dropped ASUS from your statement. This comment is not about ASUS.

In general, a product could be offered with a 5 year warranty, the product could fail within that period. A Company could deliberately make the RMA process so difficult, slow, and expensive (for example you pay for return postage to another country) that people just give up and do not return a failed product. Offering RMA and RMA process will be seen as a loss making side of the business by accountants so it could be a "don't care" in reality, despite what consumers are told by product marketing before purchase.

Bye.

[elstaci]

RE: A Company could deliberately make the RMA process so difficult, slow, and expensive

Totally agree with that statement. I have a ACER Monitor that cost over 300.00 from Amazon.com. It has been working great since I purchased it. But unfortunately, at the top upper right corner I am seeing a white streak of failed pixels forming. I realize that eventually a whole area of the screen is going to go bad since I can see a very faint area about the size of a small yellow sticky note forming. I opened a RMA Ticket to ACER to repair/Replace it under Warranty which expires in November 2018. They said to ship the ACER Monitor to their RMA Repair center in Texas. It will cost me anywhere from 80 dollars to over 90 dollars to ship the Monitor to Texas from Florida (shipping cost plus packaging). I chatted with ACER online and mentioned that ACER has a Authorized Repair Center very close to where I live. Can I send the monitor to the ACER Authorized Repair center where I live and have them ship it to Texas or for repairs. They said No, if the Monitor should get damaged while in transit they would be responsible and ACER has a specific Repair Center for RMA which is located only in Texas.

So basically I have to live with this Monitor until it become unviewable since shipping it to Texas will cost around 30% of the Original cost of the Monitor. Never will I purchase a LARGE ITEM from Acer again. Small items I will. like Burners or Mouse etc. which will be relatively cheap to ship under warranty.

[colesdav]

Talk with Amazon about it. They have been pretty good to me about one Warrranty / failed part.
They might talk with ACER on your behalf or have something to say about the postage cost.

My Sapphire HD7970 OC 6GB card failed and I contacted them to check the Warranty on the Card. The Warranty had expired, however they refunded just under 1/3 of the original cost of the card to me (~ price of standard 3GB RX280 / new lower end AMD card with similar gaming performance). They told me to  me keep the failed card,  I could sell it for parts.

Instead of just purchase another card, I purchased tools (through  Amazon) to repair the GPU and attempted to fix it myself, mostly because I know how the card has been used and it has 6GB of VRAM. I also wanted to learn about fixing GPU since I own quite a few working GPU's now. They might fail and need repair in future. I managed to repair the Sapphire HD7970 OC 6GB.

Good Luck.

Thanks.

[colesdav]

Hi,

I think Sapphire and other AMD AIB GPU's do use high quality components in general, you can see on their websites.
They all seem to discuss the quality of the capacitors and chokes used on their cards.

Manufactures and designers of any electronics product all face fierce competition on performance, features and price. It is ruthless in my experience. 
Just like the song "Everything Counts" from an old 80's band,  Depeche Mode, everything counts in large amounts.
If the difference is $5 on each unit sold is $5 extra profit per unit. That could translate to many millions or even billions depending on units sold.

I am not sure "planned failure" is really the term used, however each product designed does have an expected lifetime before obsolescence these days.
That lifetime does seem to be getting generally shorter before things fail, but electronics is much more complex with many more transistors and higher performance and power so failure rates can inherently go up. Sometimes products are obsolete if they are released just a few to 6 months late.

RE: I have fixed many Motherboards over the years replacing the capacitors but these multilayered pcbs today are getting a bit beyond what I can do at home.

Yes I agree with that. Fixing SMD components on newer PCBs is not easy. If something blows taking out a section of multilayer PCB you have very likely have to junk the entire board. I had been looking into  building my own SMD repair / reflow station so I can repair things if needed.

You might want to try joining a Makerspace Club if there is one in your area, as there are often projects running to help people work together to make a SMD Repair and Reflow Station available for members to use. That can often reduce costs and sometimes local Companies donate older equipment. Those clubs are often filled with interesting people who have a wide range of skills and long experience from different technology backgrounds, especially electronics and mechanical design and manufacture.

I have an old HP Envy 15 motherboard to practice on, the caps around the CPU socket need replacing.

Unfortunately I also have a PowerColor Fury X with a hole in the PCB where a component on the fan controller (I think) started to fail or was modified before i purchased it, so I can practice on that as well...  I tried giving it away but not taken, so, I have purchased a Sapphire Fury X. Hopefully that will work fine, and I have the PowerColor Fury X for spares now.

Bye.

[pokester]

My very quick point I was trying to make without having to write a page to explain myself was that they could use capacitors that are rated to last say 50 years of 24/7 use for not much more money than what is often used today that often seems to fail within 5 years of heavy use. Many makers are using certain better parts on their boards and they advertise those better parts on their boards packaging and websites, like the solid caps and more recently the ferrite chokes have gotten the "special attention" on boxes now too. But they still stick other junk parts on instead of the better electronics in some cases. Kind of like the old saying a chain is only as strong as its weakest link. Even those better parts often aren't even what the AIB maker thought they were. Most of that stuff on the box today is just marketing band-aids to smooth over the normal practice of sub par compared to what could have been used parts that is common practice in the world today.  I personally would rather have an extra five bucks or more tacked on the price for good parts instead and let obsolescence be dictated by when "I" choose to replace the part over hardware failure. For many years I bought boards that met military specs as they typically did use the best parts. If you can find a board that lists that, it is often using better components. This used to be very common for high end boards before everything became gaming oriented in the market, to have Mil Spec 4 components really meant something. For most people this difference probably is not needed if you are going to replace a machine in 3 to 5 years. For me I tend to have a machine in service for 10-15 years. The good thing is that at least the capacitor are much better than they were during the mid 2000's Chinese counterfeit fiasco when they were exploding after a few months usage. Which brings up another real problem in the market today. Even when a maker thinks they are using the GOOD PARTS the part they are using is often a counterfeit part that to the eye looks the same but would not meet the quality standards it should. My wife actually quality tests electronics for a defense contractor for a living and complains of this being a constant issue and has been for I'd say going on 20 years now. I really do believe that for what we pay though, and how complex the boards have become that we aren't getting ripped off or anything. It is really pretty incredible that a top notch board is only a couple hundred dollars. Heck what I payed for my first IBM XT board was 15 times that, but that machine still fires right up today too. I thinks we all just get frustrated when are part craps out on us before we stopped having use for it. Especially when it is such a cheap component on that part like a cap or resistor that forces you to have to basically buy a whole new computer or video card etc. Most of us seem to have less and less disposable income these days. The saddest thing I see happening today is the denial of warranty going on. The chipmaker, the AIB maker, the computer maker and the seller all shuffle the customer back and forth whenever there is a problem instead taking care of the issue. They do this until the customer gives up out of frustration and just ceases get what they actually are entitled to by warranty. But because so few are ever going to threaten legal action, these companies are doing this more and more. Companies will have statements in their warranty I'll pick on one by name, Cyberpower has it right in their warranty that if unless they or an authorized party, installs a part in your computer you void your warranty on the whole thing. So say you computer has 4 rams slots available and came with 2 chips, you add 2 yourself, you just voided your warranty. This is crazy and is actually illegal by the Magnuson-Moss Warranty Act of 1975 but unless people challenge it in court good luck getting them to do anything different than what they say they will in their warranty as written. By law a company does not even have to offer a warranty at all, but if they choose to they cannot limit service to just the oem or oem parts. You by law have the right to use equivalent parts and put them in or use another 3rd party service to do so without breaking the warranty and can't be denied because you did this. The trick is you likely have to litigate to get them to honor this, and most wont do this. I know I am just getting old, I just remember the days when computer parts were built to last, the boxes they came in were plain cardboard with no marketing in sight and when you had an issue they gave you an RMA with few questions asked. That used to be the industry standard.

[colesdav]

Hi,

I understand your point of view.  I own and run old Vintage Personal Computers that were built in ~ 1982 and they still work fine... Also have PC's built around 1995 still running fine. They are pretty obsolete now though and it is just a hobby.

Thanks.

[grantelb4rt]

I just wanted to update you all...

The seller was kind enough to refund me the full value of the Card (what I payed in 2016).

So I will probably switch to the dark side, since the Vega cards are still way to expensive for the Performance they offer.

At least I will be able to use my PC again without any Performance disadvantages..

[elstaci]

Seems like the seller did the right thing.  If this is normal SOP for this seller concerning Warranty returns, I would keep supporting the seller by purchasing their products in the future.

[grantelb4rt]

Yep, and their prices are alright too, so I plan on purchasing there again.

After all this hassle I will probably buy an EVGA card, since they have a longer guarantee and their support is said to be really good.

[pokester]

Who is the seller if I may ask? Love great referrals like that. We all like to hear about folks that stand behind what they sell. So rare these days.