EVGA has always been overrated in my opinion...like you said their GPU board designs have never been that innovative...slapping an ACX cooler on their cards is the extent of their custom designs...they've always gotten a pass because of their Step Up program which makes buying from them a high value proposition
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
EVGA screwed up ACX on the 970.
- Thread starter TaintedSquirrel
- Start date
EVGA has always been overrated in my opinion...like you said their GPU board designs have never been that innovative...slapping an ACX cooler on their cards is the extent of their custom designs...they've always gotten a pass because of their Step Up program which makes buying from them a high value proposition
Spot on.
Just to add the 970 FTW doesn't have any heatsink contact with either the vram or the VRMs. vram I can give a pass, but VRMs? Are you fucking kidding me here?
Not to mention the 970 FTW is as expensive as the Gigabyte 970 G1, but only comes with 2x6 pin connectors, which makes it overclock like a dud in comparison to the offerings by MSI and Gigabyte. And the MSI 970 is cheaper to boot.
Personally if I were ever to buy EVGA it would be solely for their service. Their board and heatsink designs are decidedly middle of the road, yet their cards cost as much or even more than their competition. Although one could probably argue you're paying extra for good service.
flopticalcube
Limp Gawd
- Joined
- Oct 2, 2011
- Messages
- 325
On the internet, everyone is a thermal design expert.
You really ought to read 3DGuru's review of the Asus and MSI 970s. Even with a small heatsink the VRMs heat up past 80C when cooled entirely by the airflow from the fans. It doesn't take a thermal design experrt to figure out that the VRMs will run toasty without any sort of heatsink to increase the effective surface area for heat dissipation. Hey but you keep believing what you want if it makes you happy.
People don't like hardware running hot in general period. If thermals didn't matter as long as they didn't affect stability then why bother with good cooling at all?
Most of the criticism here is directed at the fact that EVGA offers a worse cooling solution for the same or more amount of money relative to the competition.
Most of the criticism here is directed at the fact that EVGA offers a worse cooling solution for the same or more amount of money relative to the competition.
flopticalcube
Limp Gawd
- Joined
- Oct 2, 2011
- Messages
- 325
Anandtech didn't seem to think it was a problem on the FTW.
Usually, speaking as an EE, as long as the thermals are within the operating range of the component there should not be a worry. My one worry with the VRM on the FTW is that should there be a fan failure, the VRM might overheat faster than the GPU can shutdown. I'm am sure, however, that EVGA has tested this. I don't think they want loads of boards returned with burned out VRMs.
The thing about the 970 is that even though its a mid-level enthusiast card, its power draw is more modest than those of the past. This allows some liberties in the design of the card that would not have been possible before. I am not saying that EVGA should have necessarily done this, its just not the huge failure that some are purporting it to be.
What did they do wrong? The optics look bad even if the engineering doesn't. They probably should have added some VRM heatsinks to the FTW and a chunkier heatsink to the ACX 2.0 models. They should have probably not have bothered with the ACX 1.0 models. They should probably have had the 0db firmware from day 1. Would they have had better temps with a bigger heatsink like the MSI/Gigabyte and ASUS? Probably but then they would also have a bigger card. Would it have made a difference in overall performance, probably not.
The free backplate and sudden firmware upgrade also smack a little of corporate back-peddling. Then again, some web sites should do some back-peddling of their own for claiming about misaligned heatpipes and design flaws when no such thing existed.
Usually, speaking as an EE, as long as the thermals are within the operating range of the component there should not be a worry. My one worry with the VRM on the FTW is that should there be a fan failure, the VRM might overheat faster than the GPU can shutdown. I'm am sure, however, that EVGA has tested this. I don't think they want loads of boards returned with burned out VRMs.
The thing about the 970 is that even though its a mid-level enthusiast card, its power draw is more modest than those of the past. This allows some liberties in the design of the card that would not have been possible before. I am not saying that EVGA should have necessarily done this, its just not the huge failure that some are purporting it to be.
What did they do wrong? The optics look bad even if the engineering doesn't. They probably should have added some VRM heatsinks to the FTW and a chunkier heatsink to the ACX 2.0 models. They should have probably not have bothered with the ACX 1.0 models. They should probably have had the 0db firmware from day 1. Would they have had better temps with a bigger heatsink like the MSI/Gigabyte and ASUS? Probably but then they would also have a bigger card. Would it have made a difference in overall performance, probably not.
The free backplate and sudden firmware upgrade also smack a little of corporate back-peddling. Then again, some web sites should do some back-peddling of their own for claiming about misaligned heatpipes and design flaws when no such thing existed.
Ocellaris
Fully [H]
- Joined
- Jan 1, 2008
- Messages
- 19,077
Anandtech generally gives a glowing review to everything they touch.
Marketing.
Take a look at most motherboard VRM heatsinks. Hugely stylised, in bright colours and unusual shapes, often with some variety of cosmetic shrouding. Designed primarily for appearance, with heat sinking being a distant second. Notice how business boards that aren't marketed to consumers often lack heatsinks altogether, and even server boards, with tighter specifications and harsher validation,often don't bother with heatsinks or appear to have very 'undersized' heatsinks.
Then there's RAM heatspreaders, which are 100% cosmetic.
A lot of the time heatsinks are added because consumers have been told to expect them, rather than actually being needed. Silicon chips don't "wear" from sustained high temperatures. As long as they do not get hot enough that the packaging is damaged, the solder melts, or the transistors themselves begin to tunnel, the chip will continue to work to spec.
Unknown-One
[H]F Junkie
- Joined
- Mar 5, 2005
- Messages
- 8,909
Not always true. The manufacturers producing enthusiast products have to account for the fact that many users will water-cool their systems, leading to a massive reduction in airflow around motherboard power regulation circuitry.
This generally isn't a problem in normal air-cooled workstations and servers.
Servers move ridiculous volumes of air front-to-back, channeled over every component. Cooling is already assured, so there's no need to over-provision the system in case of low-airflow installations.
That I'll agree with. Had to remove the heatspreader from some Corsair Vengeance to get my Megahalems installed, and stability has not been effected. The RAM barely even gets warm, according to my IR temp gun.
limitedaccess
Supreme [H]ardness
- Joined
- May 10, 2010
- Messages
- 7,594
Semiconductor lifespan and efficiency is affected by temperatures. It isn't just a fail/pass situation. I believe the literature for mosfets (which is what VRM sinks are actually covering) suggests a factor in 2 change in average lifespan for each 10C change in operating temperature.
This is also actually how manufacturers do stress testing and reliability testings. They run at much hotter than regular environments to essentially speed time up, otherwise testings would not conclude until the products themselves are obsolete. For example if you remember Intel's Cougar Point recall (Sandybridge chipset) this is how they uncovered the bug which otherwise would not on average show up in the wild until ~1year later.
Of course we don't exactly have the actual technical data available to judge how much of an impact in real terms in theory (eg. what the components are actually specced for). It'd also need more in-depth hands on testing that isn't available. Likewise we don't actually know what the VRM temperature difference would be since Nvidia, unlike AMD, does not have monitoring. But on AMD products the operating temperature of VRMs have been known to vary from 90C to 130C, with the latter for products with problematic VRM cooling.
This is also actually how manufacturers do stress testing and reliability testings. They run at much hotter than regular environments to essentially speed time up, otherwise testings would not conclude until the products themselves are obsolete. For example if you remember Intel's Cougar Point recall (Sandybridge chipset) this is how they uncovered the bug which otherwise would not on average show up in the wild until ~1year later.
Of course we don't exactly have the actual technical data available to judge how much of an impact in real terms in theory (eg. what the components are actually specced for). It'd also need more in-depth hands on testing that isn't available. Likewise we don't actually know what the VRM temperature difference would be since Nvidia, unlike AMD, does not have monitoring. But on AMD products the operating temperature of VRMs have been known to vary from 90C to 130C, with the latter for products with problematic VRM cooling.
theelviscerator
Gawd
- Joined
- Jan 26, 2004
- Messages
- 959
Speaking from experience, my ACX 2.0 cards run quite cool, high 50s to low 60s C, under full load.
CtrlAltDefeat
n00b
- Joined
- Dec 21, 2010
- Messages
- 40
X2. My evga 970 ssc runs cool also. It is nice with the standard profile that shuts the fans off for silent mode.
With that said I went ahead and made a custom fan profile to keep it nice and cool while gaming. Still the fans only get around 55% and my temperature is around 55C.
With that said I went ahead and made a custom fan profile to keep it nice and cool while gaming. Still the fans only get around 55% and my temperature is around 55C.
MSIGTX650Ti
n00b
- Joined
- Jan 4, 2014
- Messages
- 31
Sorry to dig up old conversations, But people on this forum seem much less narcissistic than others, And you seem to care about truth more than other forums,
I bought a EVGA GTX 970 SC ACX 2.0, (Yes I know, I messed up) But there was only one place I could buy a 970 with cash, And this model was the only one reasonably priced, And I desperately needed a upgrade because my HIS R9 270 fell behind real quick. Sad thing is I bet the HIS IceQ X² mini twin cooler would do a better job LOL, Anyway, I have no coil whine, So I lucked out on that, But I did get crashes in unigine heaven and 3dmark sky diver unless I underclocked it, But again I got lucky and re-installing Windows 8.1 solved the crashing problem. So I plan keeping the card awhile, But since the quality is so so, Would it be worth it to invest in a water cooler for it, And just do away with the ACX 2.0 cooler? And if so what water coolers do you guys recommend? If possible I would prefer to have the V-Ram cooled too.
I bought a EVGA GTX 970 SC ACX 2.0, (Yes I know, I messed up) But there was only one place I could buy a 970 with cash, And this model was the only one reasonably priced, And I desperately needed a upgrade because my HIS R9 270 fell behind real quick. Sad thing is I bet the HIS IceQ X² mini twin cooler would do a better job LOL, Anyway, I have no coil whine, So I lucked out on that, But I did get crashes in unigine heaven and 3dmark sky diver unless I underclocked it, But again I got lucky and re-installing Windows 8.1 solved the crashing problem. So I plan keeping the card awhile, But since the quality is so so, Would it be worth it to invest in a water cooler for it, And just do away with the ACX 2.0 cooler? And if so what water coolers do you guys recommend? If possible I would prefer to have the V-Ram cooled too.
Ocellaris
Fully [H]
- Joined
- Jan 1, 2008
- Messages
- 19,077
The ACX 2 cooler is fine. If your card is crashing due to heat issues, something is defective on the card. Have you checked to see what the temps are reaching under load?
Indeed... It seems that a large majority of people are not seeing the big picture here... and overclocking potential is not the only factor. (time for a soap-box moment)
Three other very significant factors are:
Value (bang for buck) - I regularly see the ASUS Strix priced higher than the EVGA cards... and the factory clock is higher on the EVGA SSC & SC compared to the Strix (1342 & 1317 vs 1253, respectively). So you want to overclock..., let's say that everybody gets the results found here on HardOCP with a 1473 core OC and 7832 memory OC... and let's say that the EVGA only gets the avg OC of say 1425, and 7700... That difference MIGHT be an average of 2-3 frames at a game that already averages around that magic 60fps... 2-3 frames on tat is nothing, and imperceptable by the human brain... Is that really worth the potential extra cost of $30 - $100??? For some people.. sure, for the extremists... and for those who get suckered into the hype driven by threads like this one.
Reliability (including stability) - EVGA consistently provides a solid product that does what it says it can do. Again, most people do not overclock.
& Support - I don't believe I've seen a bad review for EVGA's support (though, I'm sure it's happened... we are still dealing with human beings on the other end of that support call)
The Big Picture
And, at the end of the day... does it work or not? Is it effective, or not? Based on all of the reviews I have seen, I would say, yes...most definitely, yes. Now, if they had designed a new cooler with direct contact for all three pipes, it would probably be better, but anybody who has ever been involved in a business decision (not many people here) would understand that there were likely a lot of financial factors involved with EVGA's decision to use the existing ACX heatsink. That decision surely included extensive testing to verify that the cooler was sufficient to provide the advertised performance for the models which it was applied to. It probably also included an engineer (or several) on their design team saying "this is a bad idea and we should design a better cooler" (I have been that engineer in my profession before)... But the truth is that the cooler works for those models and the manufacturing cost was more economic for EVGA to go ahead and use that existing design, which in turn, keeps overall production costs down and allows for higher performing models to go on sale, or to be sold at more competitive MSRPs.
FYI... here is the ACX 2.0 that ships with the FTW edition which makes full contact with the GPU via a full aluminum plate... and the cooling is sufficient enough to overclock even higher than the factory clock of 1367 (just set mine to 1450/7700 today)
http://images.anandtech.com/doci/8568/HSF.jpg
And it also seems that there is a new ACX 2.0+ revision... though, it also uses a large aluminum plate to make full contact with the GPU, but still has copper heat pipes.
All of that said.... Two complaints I have with EVGA
1. Coil Whine... my FTW GTX 970 performs awesomely, and I play games with TV sound pumped out to my stereo (stereo HDMI pass thru is broken), so I don't notice it. Audio via HDMI is fine, but coil whine is amplified via standard audio jack.
2. They really do have WAY too many different revisions for every card they manufacture.
For the apologists who will surely find some comment that I didn't research to the n'th degree and begin to respond with what "they" know about it...
1.Fact: Most people do NOT overclock (this is a true statement).
2. Yes, I know that all different mfg models of graphics cards go on sale ALL the time.. which leads to fact #2, most people who do any performance research will buy the fastest card that is available for the best price "at the time they are shopping"... which is typically less than a 24-hour time frame. The point is, most people do not WAIT for the "best" card to go on sale.
3. "Most people" barely do any research at all... they tend to buy what they know and trust, regardless of the overall value.... or take the advice of a friend (even when it's bad advice).
4. Yes, I know that some people think "every frame counts" and are willing to pay hundreds of dollars to go from 60 fps avg to 70 fps avg (which is imperceptible)... As fun as it is to compete in that Schwartz contest.. it's unreasonable and simply bad economics.
5. Does the direct copper contact on the Strix make it overclock higher (not according to most review)? Does it make it operate quieter?? I don't know, and don't care... does it actually matter that much?? Do ANY of the air-coolers run "quietly" on a full load? NO
(does anybody else hear the apologists in their head as they type out their responses here? or is it just me?)
k, I'm done...
Last edited:
MSIGTX650Ti
n00b
- Joined
- Jan 4, 2014
- Messages
- 31
No I am not having crashing problems, (I wish people would read) I had crashing at first because of a software issue, But after a format and re-installing Windows 8.1, I no longer get crashes.
MSIGTX650Ti
n00b
- Joined
- Jan 4, 2014
- Messages
- 31
Again I wish people would read, I have no coil whine, And I do not think my temps are bad, And I am not having crashes, I had crashes after GPU boost would clock the card back down after being stressed, It must have been a hardware conflict, Because after a format and Windows 8.1 re-install, the card can be stressed all day and not crash. All I asked was if water cooling would be a good idea, Because I want to make the card last as long as possible, And if so what would be the best water cooler, I suppose I was wrong about the narcissism of the forum, And on top of that people want to only half read a post, and then ramble on! You know what! Nevermind! I do not need any advice!
My reply was not directed at you... It was directed at the general topic of the thread and all the negative association with the cooler that EVGA chose to use on the GTX 970.
As for coil whine, I have an EVGA 970, and I do have coil whine... which is why I listed it as a complaint.
Retaliator
n00b
- Joined
- May 14, 2015
- Messages
- 10
Had an FTW and SSC version of the 970 - both had coil whine, but when using Vsync at 60hz, its barely audible - without Vsync, its annoying - anyone experience the same?
Ocellaris
Fully [H]
- Joined
- Jan 1, 2008
- Messages
- 19,077
That is pretty par for the course with coil whine. Try 3D Mark where it runs at like 1500 for one if the tests, should make quite a howl
IdiotInCharge
NVIDIA SHILL
- Joined
- Jun 13, 2003
- Messages
- 14,675
Got a FTW in the mail here; I hope I don't have the coil whine issue, but we'll see. Minus that, I have no expected reservations using the card.
Are you using HDMI for audio? My coil whine was only really noticeable when it was amplified though the on-board HD Audio... You can hear it on the card if you get really close... but with HDMI Audi, I never notice it at all...
IdiotInCharge
NVIDIA SHILL
- Joined
- Jun 13, 2003
- Messages
- 14,675
I won't be; DVI or DP here (HP ZR30W), and I use HD555s for sound exclusively. We'll see what happens, should be here Friday.
Retaliator
n00b
- Joined
- May 14, 2015
- Messages
- 10
That is pretty par for the course with coil whine. Try 3D Mark where it runs at like 1500 for one if the tests, should make quite a howl
You're right - coil whine is quite annoying in 3dmark and in Heaven benchmark when the GPU is really stressed