EVGA Finishes Issuing BIOS Updates GTX 1080/70/60 FTW Line of Cards

[Raevenlord]

Just a slight update on our reported stories of overheating issues with EVGA's FTW line of NVIDIA graphics cards, with the company having finished rolling out fan-curve adjusting BIOS updates for the affected cards.

The issue affects EVGA's line of FTW graphics cards ranging through the GTX 1060, GTX 1070, and GTX 1080, sporting the ACX 3.0 solution. As a casual reminder, you can view the list of affected cards below, for which BIOS updates have been issued in a blog post, along with a small FAQ and some instructions on how to perform the update.



View at TechPowerUp Main Site

 

[ssdpro]

Product is marketed, reviewed, and sold as X, which is faulty due to improper cooling. Card receives a BIOS update to medicate the problem with fan ramp. Card X is now card Y, a louder card that may or may not be overheating still. I would have zero confidence EVGA has patched this issue if they didn't understand or didn't care the card was overheating in the first place.

This is what they call a fail.

 

[the54thvoid]

Product is marketed, reviewed, and sold as X, which is faulty due to improper cooling. Card receives a BIOS update to medicate the problem with fan ramp. Card X is now card Y, a louder card that may or may not be overheating still. I would have zero confidence EVGA has patched this issue if they didn't understand or didn't care the card was overheating in the first place.

This is what they call a fail.

It's called a mistake and lots of companies make them. It's how they rectify it that counts. My past 3 cards have all been EVGA and I wouldn't hesitate to buy from them again.
Then again, I don't play Furmark.

 

[phanbuey]

I wonder if this affects overclocking or if its just a fan curve update.

 

[TheLaughingMan]

Just a fan curve update. They didn't help me one bit when the BIOS update didn't work on my system. Their forums claimed it was because my platform is old, which I could understand, but their support should know that and how to work around it.

I fixed my card with MSI Afterburner and a custom fan curve. no BIOS update. The custom fan curve stopped all issues I was having instantly. Card runs now at a happy 62 C max, its still not loud enough to hear in my case, and games like a champ.

Even with all of that, I will still work with EVGA in the future.

 

[phanbuey]

Just a fan curve update. They didn't help me one bit when the BIOS update didn't work on my system. Their forums claimed it was because my platform is old, which I could understand, but their support should know that and how to work around it.

I fixed my card with MSI Afterburner and a custom fan curve. no BIOS update. The custom fan curve stopped all issues I was having instantly. Card runs now at a happy 62 C max, its still not loud enough to hear in my case, and games like a champ.

Even with all of that, I will still work with EVGA in the future.

It does seem like a fluke issue - I have the 1080SC card, and it runs awesome, no issues and it cost less than the FE by quite a bit. I don't really push it - just a mild oc that keeps it at 2025-2012Mhz constant during gaming, and if it dies I know I can complain and have them replace it.

I would be a bit pissed if I had the Classified though, since you pay to have an 'Overclockers card'. (Then again all of them top out at around the same 2050 - 2150 MHz core, so getting one of those is depressing in itself.)

 

[Ungari]

It's called a mistake and lots of companies make them. It's how they rectify it that counts. My past 3 cards have all been EVGA and I wouldn't hesitate to buy from them again.
Then again, I don't play Furmark.

Calling this a "mistake" comes directly from EVGA's statement to the media, and makes you sound like a PR Rep.
A mistake is an unintended error, but that is not what happened here. What EVGA did was make an calculated decision to cut corners on high end products that retail for $750 and $440 for the 1080 FTW and 1070 FTW.
Cutting corners is not what customers expect from premium tier custom PCBs like the FTW. The FTW edition is only succeeded by the Classified, and only two years ago many 980 FTW were sold that were unable to maintain stability at Factory OC.
EVGA has seriously damaged their brand.

 

[the54thvoid]

Calling this a "mistake" comes directly from EVGA's statement to the media, and makes you sound like a PR Rep..

 

[cdawall]

Product is marketed, reviewed, and sold as X, which is faulty due to improper cooling. Card receives a BIOS update to medicate the problem with fan ramp. Card X is now card Y, a louder card that may or may not be overheating still. I would have zero confidence EVGA has patched this issue if they didn't understand or didn't care the card was overheating in the first place.

This is what they call a fail.

You do know how they cool mosfets correct? With air, quite a few cards don't even have a heatsink on the mosfets. I strongly doubt these were actually exceeding case temp, but they were heating the surrounding area pretty bad. Couple of people on the page have already posted that after the update the card was able to maintain a higher boost clock as well as not being noticeably louder.

 

[kaellar]

All graphics cards shipped from EVGA after 11/1/2016 will have the VBIOS update applied

Of course, who the hell needs thermal pads preinstalled? Everyone wants the card that has to spin its fans at 2200 rpm to operate properly. Honestly, screw EVGA this time around.

 

[cdawall]

Of course, who the hell needs thermal pads preinstalled? Everyone wants the card that has to spin its fans at 2200 rpm to operate properly. Honestly, screw EVGA this time around.

They didn't change the fan profile that much. Holy assumptions. Do you even know where the thermal pads they are offering go on the card?

 

[kaellar]

They didn't change the fan profile that much. Holy assumptions.

sure they did. go check the source thoroughly.

 

[cdawall]

sure they did. go check the source thoroughly.

I have flashed a 1070 with it already and didn't notice a massive ramp up on the fan speed during testing.

 

[kaellar]

My roommate's PC has evga sc 1070 and it definitely became way loudier after an update. So I 100% believe gamersnexus's article is absolutely right about vbios changes: http://www.gamersnexus.net/news-pc/2666-evga-heat-solution-thermal-imaging-of-vrm-1080-ftw

 

[newtekie1]

Calling this a "mistake" comes directly from EVGA's statement to the media, and makes you sound like a PR Rep.
A mistake is an unintended error, but that is not what happened here. What EVGA did was make an calculated decision to cut corners on high end products that retail for $750 and $440 for the 1080 FTW and 1070 FTW.
Cutting corners is not what customers expect from premium tier custom PCBs like the FTW. The FTW edition is only succeeded by the Classified, and only two years ago many 980 FTW were sold that were unable to maintain stability at Factory OC.
EVGA has seriously damaged their brand.

I can't really by that when the rubber stand-offs on the backplate probably cost just as much as the thermal pad they replaced. It wasn't really a cost cutting measure, I just think they didn't believe the thermal pad was necessary, and it hurt the aesthetic of the card(because the thermal pad can be seen through the backplate grill). This was a mistake.

I also think they are handling it well. They are issuing a BIOS fix, but also replacing cards if the customer wants. They are also offering to send out the thermal pad if the customer wants to do it themselves. Also, if you opt for a replacement card, they are cross-shipping the replacement, so you have no down time. I don't think they could handle the situation any better.

 

[Smanci]

Couple of people on the page have already posted that after the update the card was able to maintain a higher boost clock as well as not being noticeably louder.

Every fact-based review shows very noticeable noise increase. Confirmation bias is a very real thing and so is the brand loyalty that's the backbone of EVGA's business.

 

[cdawall]

My roommate's PC has evga sc 1070 and it definitely became way loudier after an update. So I 100% believe gamersnexus's article is absolutely right about vbios changes: http://www.gamersnexus.net/news-pc/2666-evga-heat-solution-thermal-imaging-of-vrm-1080-ftw

Ah we are playing furmark again that's probably why I didn't notice it. I just spun up fallout 4 and metro LL, it was a customers machine with a black screen complaint. (it was an unrelated XMP issue)

Every observation-based review shows very noticeable noise increase. Confirmation bias is a very real thing and so is the brand loyalty that's the backbone of EVGA's business.

I don't own an EVGA card, most of the reviews keep playing furmark (which again I couldn't care less about) and noticeable noise increase is in the ears of the listener I might be a bad example having worked around the loudest aircraft in the Air Force's inventory for the past 6 years.

 

[Smanci]

noticeable noise increase is in the ears of the listener

Valid point but if the cards turn from quiet to noisy in comparison with other manufacturers' cards, that should tell something. Furmark is unrealistic load but heavy gaming power consumption (GPU only) hovers within just a few watts from that. Total system power draw will be higher. Also people who run their cards at hotter environments or in less well ventilated cases will be clearly affected.
There's no way getting around that these are rather disappointing cards unless you want to take advantage of being able to replace the cooler (cooling solution if you wish) with something more functional.

 

[cdawall]

Valid point but if the cards turn from quiet to noisy in comparison with other manufacturers' cards, that should tell something. Furmark is unrealistic load but heavy gaming power consumption hovers within just a few watts from that. Also people who run their cards at hotter environments or in less well ventilated cases will be affected.
There's no way getting around that these are rather disappointing cards unless you want to take advantage of being able to replace the cooler with something more functional.

It isn't the cooler, the VRM just needs to be thermal padded to the backplate. Like was said that doesn't look as pretty.

 

[Ungari]

I can't really by that when the rubber stand-offs on the backplate probably cost just as much as the thermal pad they replaced. It wasn't really a cost cutting measure, I just think they didn't believe the thermal pad was necessary, and it hurt the aesthetic of the card(because the thermal pad can be seen through the backplate grill). This was a mistake.

I also think they are handling it well. They are issuing a BIOS fix, but also replacing cards if the customer wants. They are also offering to send out the thermal pad if the customer wants to do it themselves. Also, if you opt for a replacement card, they are cross-shipping the replacement, so you have no down time. I don't think they could handle the situation any better.

The FTW is a custom PCB that runs a higher factory overclock than any other model except the Classified, yet you are suggesting that EVGA in it's design testing did not see that these cards ran at higher temps than normal and could have benefited from TIM on the VRAM and Memory chips?

And so when EVGA tested the cards and found higher temps yet allowed them to be released for retail, was that a "mistake"?

Two years ago EVGA had released for retail GTX 980 FTWs that had to be downclocked to run without crashing due to improper binning of the chips. Was that also a "mistake"?

 

[phanbuey]

The FTW is a custom PCB that runs a higher factory overclock than any other model except the Classified, yet you are suggesting that EVGA in it's design testing did not see that these cards ran at higher temps than normal and could have benefited from TIM on the VRAM and Memory chips?

And so when EVGA tested the cards and found higher temps yet allowed them to be released for retail, was that a "mistake"?

Two years ago EVGA had released for retail GTX 980 FTWs that had to be downclocked to run without crashing due to improper binning of the chips. Was that also a "mistake"?

it wasn't not a mistake.

So you think that eVGA's long term business strategy is to put out bad products and hope that no one notices? You're talking about 1. an unstable factory overclock on the 980FTW, - which I am pretty sure they would replace the card if you opened the ticket. and 2. vrms running hot that after 1.5hrs in Furmark (using modded drivers to make furmark run) will pop, and the company will replace the card.

I think they handled the issues well, but that they also do need better product QA testing prior to release. It's clear that their process needs to be a bit tighter in regards to that.

 

[EarthDog]

LOL, rinse and repeat... same people on both sides too...

Half of you need to go invade Tom's...its a more fitting home.

 

[newtekie1]

The FTW is a custom PCB that runs a higher factory overclock than any other model except the Classified, yet you are suggesting that EVGA in it's design testing did not see that these cards ran at higher temps than normal and could have benefited from TIM on the VRAM and Memory chips?

And so when EVGA tested the cards and found higher temps yet allowed them to be released for retail, was that a "mistake"?

Ok, so this post shows you have no idea what the issue actually is. The VRAM and Memory chips already have TIM on them, and they aren't the parts overheating. Also, the problem isn't the custom PCB either.

The problem is also not just limited to the FTW card, it affects all cards using the ACX cooler. And the problem is a missing thermal pad on the VRM. This is also a part of the card that doesn't have a thermal sensor.

Also, the problem only happens under furmark load for extended periods of time. Remember when AMD's reference design blew VRMs when doing this???

 

[Ungari]

it wasn't not a mistake.

So you think that eVGA's long term business strategy is to put out bad products and hope that no one notices? You're talking about 1. an unstable factory overclock on the 980FTW, - which I am pretty sure they would replace the card if you opened the ticket. and 2. vrms running hot that after 1.5hrs in Furmark (using modded drivers to make furmark run) will pop, and the company will replace the card.

I think they handled the issues well, but that they also do need better product QA testing prior to release. It's clear that their process needs to be a bit tighter in regards to that.

As to the 980 FTW, way too many tickets were opened and EVGA did replace the cards, if they weren't immediately refunded by the seller. As to the Paxwell FTWs, yes they have been responsive to correct the issue.
But this is not about EVGA Customer Service which I understand has an excellent reputation, it's about the need for so many to have to use their customer service;---and why?
Yes, I am saying there is no way that EVGA did not notice these problems in QC, and that there was a deliberate decision to go ahead and release these cards to retailers. To call this a "mistake" is misleading as it was an intentional decision by EVGA.

 

[Ungari]

Ok, so this post shows you have no idea what the issue actually is. The VRAM and Memory chips already have TIM on them, and they aren't the parts overheating. Also, the problem isn't the custom PCB either.

The problem is also not just limited to the FTW card, it affects all cards using the ACX cooler. And the problem is a missing thermal pad on the VRM. This is also a part of the card that doesn't have a thermal sensor.

Also, the problem only happens under furmark load for extended periods of time. Remember when AMD's reference design blew VRMs when doing this???

No, I watched the video showing where to apply the TIM pad strips, and one strip gets used on the memory chips.