Roguey's avatar Roguey's blog United Kingdom
Level badge Trueguey (22)
Joined: 2nd January, 2010

Blog » Stay away from Zotac cards

Posted by Roguey on Saturday 9th April, 2016   ·   Comment 2 comments

Awhile back, you may have seen that I brought a new Zotac 980 Ti Amp. Before purchasing the card I did a lot of research to work out which one I should buy. In the end it came down to price and 5 year warranty. I wish I would of picked another brand now.

Blog picture


Before Christmas, I had the card running at 1080p. 1080p doesnt put much stress on the card, as 60fps isnt hard for a card of this power. For a Christmas present I got a 4k screen, which meant increasing the resolution to 4k. This is when I noticed some big problems.

Immediately the card got a lot louder, from barely audible to the loudest part in the PC. If you ever got 3 high speed fans and turned them up to their maximum you may get the idea; I could hear the PC from downstairs! With what I read about 4k - I thought this was normal operation, but clearly it wasnt. I should of picked up on this faster.

A few days ago I noticed the card was especially loud; so loud it felt that something was wrong. After checking GPU-Z I saw a temperature of 90c, with a fan speed of 99%. 90c is only 10c away from boiling water - clearly something was wrong. I have never seen a GPU hit that temperature ever before. I started doing some more tests to see if that was a one off, however in quite a few games I could reach 85c, sometimes 88c. When looking at the card, the cooler is massive and shouldnt of been struggling;

Blog picture


Thinking now this was a real problem, I contacted Zotac to see what was going to happen. Where I got some strange and poorly worded replies. They passed the buck to the reseller (overclockers.co.uk); who suggested this was normal operation - 85%+ fan speed, 83c+ with constant thermal throttling. I am disappointed in the reseller also.

At this point I had 2 choices;

  • Put up with it (live with the very loud card, reaching stupid temps - maybe even try and break it so it would be RMA'ed - which is kind-of stupid),
  • Or try and fix it (thinking it was something simple like too much thermal paste),


Ive built PC's for many years, so I know how to put them together and take them apart. Graphic cards however... not much; normally I buy them already overclocked with good coolers on them. So all I need to do is plug them in. I was pretty sure the problem was something simple, like too much thermal paste between the cooler and the GPU. Its not something you would expect on a £500 graphic card! However it wouldnt be the first time a company have cocked-up things like this...

So hesitantly I removed the cooler, undoing the screws - which felt a little lose to be honest. At this point I think my warranty may be void, but im a believer if you can fix something yourself, you should! Why try and break it, so you can RMA it? The cooler came off pretty easy to be honest, where I saw what was wrong;

Blog picture


The thermal paste between the cooler and the GPU was extremely poorly applied. It looked like a big blob had been squirted onto the top left (and way too much of it), and then when the cooler was screwed on, it had been tighten up in the bottom right corner first - forcing the paste to pushed out of the top left (meaning there was little/no paste what-so-ever on the bottom right part of the gpu). This is one the worse applications ive seen from a company.

So now I cleaned up the GPU and removed all the cheap gunk. With the GPU cleaned, I replied a small blob of MX-4 (in the centre, like how it should of been done). So next it was the matter of screwing it back together, being careful not to over tighten any of the screws (this could potentially crack the gpu). This didnt take long, so it was a matter of sticking it back in the PC and trying it.

Straight away I saw a drop, my idle temp when down from 40-50c to 33c! The lowest I ever seen - good sign! So the next thing was to check its gaming performance, this is when things really matter. In low-loads the fans remain off, so doesnt really give you a true idea on its performance. So now I tried FurMark and started the test. After 5-10 minutes the highest temp I saw was 70-71c, with a fan speed of 57%;

Blog picture


This felt very strange - the card was being taxed at 100% usage, but was more or less silent. This is the card I should of had at the start! This was eerie - the lack of noise. So I checked a few more games... still nothing, even in 4k! I couldnt get this card pass 71c, or make much noise. It looks my repair has knocked off up-to 18-19c!

So my advice; stay away from Zotac cards; the paste was extremely poorly applied on this card, and it should of been RMA. I was told by reseller this was normal operations. Both Zotac and reseller would not help, so I had to take matters into my own hands - which may mean the warranty is void. I would be extremely interested if Zotac would comment on this post. This repair shouldnt make the warranty void!

Social media link: The Facebook post on this story

[update 1] the reseller was overclockers.co.uk. Zotac had told me (On the Friday);

If the GPU temperature is still nearer 90 degree, you may contact the card seller for RMA service.
- Zotac


to which they replied;

I understand it may seem high, but those temperatures are completely within tolerance and is expected for a card of that power.
- overclockers.co.uk


[Update 2] For anyone interested, you can download and view my GPU-Z logs for here. Red is before, Green is after.

Fallout 4, Before and after; (avg. 12c drop, although games vary alot)

Blog picture


FurMark, Before and after; (18c drop! with a massive fan speed drop)

Blog picture


[Update 3] Ive been told by Zotac thats the reseller responsibility to RMA the card (on the Monday);

We do see the card fault, but the GPU-Z temperature is higher than expected.
Please contact the card seller for RMA service.
- Zotac


Which is all too late, as card was already repaired (Friday night).

[Update 4] So I wrote on Twitter how I was disappointed in both Zotac and Overclockers. Overclockers responded and said to contact their customer services. After so, I got the following reply;

Under full load the card will happily sit at around 90 degrees as per Nvidias specifications listed here: http://www.geforce.co.uk/hardware/desktop-gpus/geforce-gtx-980-ti/specifications

Unless it is hitting above nvidias recommended specs or the card is underclocking we are unable to offer a resolution as the card is operating within designed parameters.
- overclockers.co.uk


So according to them, hitting 90c isnt a problem (its happy to sit at those temps, even although NVidia state 92c is its maximum). Zotac have stated it should of been RMA by the reseller. I have to contact Zotac to see if they will atleast honor my repair, so I can keep my warranty.

[Update 5] Zotac arent understanding at all;

Follow current Zotac policy, remove the fansink will void the graphic card warranty.
If you did not remove the fainsink, that will not affect the warranty.


So basically I should have tapped up the fans and tried to burn out the card, so I could RMA. Pretty poor English too.

[Update 6] A few days later I posted this story on trust-pilot. The next day James Bailey contacted me to sort-out the problem. The blog continues here.


Comments

Avatar To post comments you need to register and log-in.
⇊ Load more comments ⇊