Zotac 980ti AMP! Extreme not working properly

[nihan2k]

Hi,

I own a Zotac 980 Ti AMP Extreme that I had purchased last year and it worked flawlessly till last month.
I noticed that my temps were idling at 52deg C @ 1400RPM and I was getting very low FPS (my FireStrike score is 3100!!). Previously it was averaging 14000 on FireStrike. On checking GPUShark, I noticed that the GPU Voltage limit has been reached message being shown. I have not overclocked the GPU at all. I have the latest drivers on NVIDIA installed as well. Could you please help me out.

Thanks.

 

[junkeymonkey]

well there was a issue with the latest NVidia drivers but I thought I affected 10 series cards [seemed to cause cards not to boost] . I still use old 255.82 driver with my 980ti ??

fire strike will max the voltage to the limits that's a v- bios set thing . so don't be too alarmed by that from it or any ''stress'' test

this maybe odd ''I noticed that my temps were idling at 52deg C @ 1400RPM '' with the 900 cards having that -0- fan thing where the cards fans may not come on in till say up to 60c , but yours is spimnning at 14000 ??

about all you can do is test the card in another system [with different drivers to test with ] to see if it does the same or not if not may need to look at something like maybe your PSU is not any longer delivering stable power to the card ??


first 5 or 6 reviews from here show died in less then a year ??

http://www.newegg.com/Product/Product.aspx?Item=N82E16814500379

 

[nihan2k]

Thanks for your quick response!

I use 375.70 Nvidia drivers. I initially thought it was the drivers and hence used DDU and reverted to a version I had when everything seemed to be working fine, but to no avail. So i doubt it could be NVIDIA drivers.

I am not alarmed by Firestrike maxing the voltage. My concern is, why is still showing GPU VOLTAGE LIMIT REACHED on GPUShark. Even Afterburner shows that the Voltage limit to be stuck at 1.

The RPM is what I sent on Afterburner. I have a custom fan setting. When everything was working fine, at 1400RPM, the temps would be idling at 35DegC.

So in my quest to figure this out, I flashed the bios to the latest version, but no avail. I got a custom bios from overclock.net. BUT when i tried to flash the bios, i did not use the tweaked version of nvflash. So after i put the command of --protectoff and proceeded to flash the bios, it obviously didn't take it. But surprisingly, it solved the issue albeit temporarily. My FPS were back to normal and everything seemed OK until i restarted the PC. And it went back to the same issue.

So then i realized that i needed the tweaked version of nvflash, i used that and flashed the bios which was made to remove the voltage limit for the GPU. it worked for others but did not work for me.

The 980ti AMP! Extreme apparently suffers from fan RPM spikes. I notice that on mine too but its not a big issue for me.

My PSU is the Corsair 750i and everything seems to be working fine from it so I would rule it out. I do not have access to another PC where I can test the GPU, so probably another method would have be suggested.

If you need screenshots of any values as such, let me know and I can put them up here.

Hope this helps in troubleshooting it.

 

[junkeymonkey]

''My concern is, why is still showing GPU VOLTAGE LIMIT REACHED on GPUShark'' [never used it and first time I ever heard of it , must not be too poplar a tool ]

but the bios for it claims Board power limit Target: 390.0 W Limit: 435.0 W [that's pretty excessive and I feel zotac trys to over do things on there cards that's asking a heck of a lot of a out of the box card everyday use to me ??] that's 36 amps just for the card on a psu's 12v rail just for the card at full load a 750w psu should be 62 amps at the 12v+

https://www.techpowerup.com/vgabios/173848/zotac-gtx980ti-6144-150602

my 980ti is only rated for 275w ??
https://www.techpowerup.com/vgabios/179138/evga-gtx980ti-6144-151104

''My PSU is the Corsair 750i and everything seems to be working fine from it so I would rule it out''

even the best of something can fail ? then say maybe the voltage limit is not due from the card and maybe the psu is not providing the voltage needed anymore ?

then if not I guess you can be down to a RMA of the card and hope the replacement is working back as correct ?

thing is with out a backup or test computer /parts that you can try the card in and see you just guess with out some kind of process of elimination


I wonder though the stress put on everything in the build to maintain that Board power limit Target: 390.0 W Limit: 435.0 W long term hours of gaming at full load ??

I don't know ? if you feel it something gone out on the card then I guess try a RMA and hope the new replacement card works as expected . if not then you looking at something else like the psu or maybe the boards slot , vrm's or maybe software ?

give any system specs or screenshots you feel is worth looking at maybe something will pop up form that

 

[nihan2k]

Thanks again for your response. I doubt I can RMA the card since I purchased it in the US but I live in United Arab Emirates. So I doubt that is an option.

I can use a multimeter and check my PSU and see if that is an issue.

"This GPU ships with a 258W TDP which you should never reach and is more than enough (a lie - both of my test units hit this limit under some scenarios without any overclock." This seems to be an issue with the card as shown here - https://www.reddit.com/r/nvidia/comments/46qo9z/avoid_zotac_amp_extreme_series_if_you_plan_on/

Here are links to screen shots of GPUShark and MSI Afterburner. I have not ran the GPU through any stress testing and these were taken on an idle PC.

https://s11.postimg.org/vsjzjet7n/Afterburner2.jpg
https://s11.postimg.org/szqrzdsv7/Afterburner1.jpg
https://s11.postimg.org/wxe1osfoj/GPUShark.jpg

 

[nihan2k]

My Rig:Intel Core i5 6600K @ 3.5Ghz cooled by H110GTX
Asus Z170-A
16GB RAM
Zotac 980Ti AMP! Extreme
H750i PSU
Win 10 x64

 

[nihan2k]

I took some more screen shots today.

These were when I restarted my PC :
https://s21.postimg.org/jas8gc6zb/At_Powerup.jpg
https://s21.postimg.org/ijziak4lz/At_Powerup_2.jpg

Then I immediately ran Furmark at 4k :
https://s21.postimg.org/dk222lyzb/Furmark.jpg
https://s21.postimg.org/6sbmzr9zr/Furmark_2.jpg

After than I ran FireStrike and got a score of 13288 :
https://s21.postimg.org/tsiac37tj/3d_Mark_Firestrike.jpg

Not soon later, the GPU VOLTAGE LIMIT REACHED message hit and these are the stats :
https://s21.postimg.org/z2inpn2on/Voltage_limit_reached.jpg
https://s21.postimg.org/qvqny2clz/Voltage_Limit_reached_2.jpg

 

[junkeymonkey]

should not matter where you bought it from zotac warrantee should be under a global policy ? just getting it to them and back from where you live maybe the biggest issue [customs -tax- shipping fees ,ect... ] if you had to send it to say UK

see things as this from your area ?

http://www.abudhabicityinfo.com/news/pr_P788N1E8CASR2JMV00A3HHT1L82Q4P8B.aspx


you could try and use hwinfo64 and run it in the background when testing and look at any min-max readings like psu gpu and all that shows to se if anything looks out of whack ?

[examples from mine after a benchmark run to see min and max useages ?]

http://imgur.com/AvRnRgU

http://imgur.com/N4V6t5Q

http://imgur.com/Ak24Pbi

or like gpu-z shows where the card was hitting its power limits

http://imgur.com/YDTq3pV

my card running valley bench and gpu-z render test at the same time

http://imgur.com/CpI1F2t


too bad you don't have a buddy with I nice gaming rig to test your card in .

 

[nihan2k]

Hi,

http://gpuz.techpowerup.com/16/11/30/kba.png

That is a screenshot of my GPUz.

I have not overclocked my GPU at all but it is showing VDDC as 1.274V but yours is showing 1.1870V. Probably this is where the issue lies.

The PerfCap reason is VRel.

This is my afterburner settings : https://s21.postimg.org/riuo7uurr/afterburner.jpg

I shall take screenshots using hwinfo64 when I get back from work.

Thanks again for the replies.

 

[junkeymonkey]

''1.274V '' well that can be set by the manufacture in the cards bios should be no big deal

heres what I don't like you got that higher voltage but you not gaining anything from it if all your card is boosting is 1253 ?? my card stock as you can see boosts to 1354 my card is speced for 1140 /1228 boost ? [evga 980ti hybrid ]

look at this screen shot ?

https://www.bjorn3d.com/2015/08/zotac-gtx-980-ti-amp-extreme-the-fastest-980-ti-available/

http://www.legitreviews.com/nvidia-geforce-gtx-980-ti-video-card-roundup-asus-msi-and-zotac_176648/11

notice from legitreview the lower voltage and higher boost and all ?

then I don't know what bios flashing you did or what went on there .

don't know ? may want to dig into that RMA and try a replacement card ? just be sure you flash the original bios back

or try the same from the card legitreviews amp extreme has on it ? [84.00.36.00.61 ??

https://www.techpowerup.com/vgabios/?architecture=NVIDIA&manufacturer=Zotac&model=&interface=&memType=&memSize=&did=10de-17c8--&since=

 

[nihan2k]

Hi JM,

i reverted back to the stock bios that you mentioned. I noticed a bit of an improvement. firstly the max VDDC is 1.193V.

I ran the stress test and here are the screenshots :
https://s13.postimg.org/vnqkjebav/stress_test_1.jpg
https://s13.postimg.org/woqt8isaf/Stress_test_2.png

The 12V rails works perfectly fine. what i found is, no matter the bios, the card runs up to the max VDDC and stays there and that is causing the issue. everytime i restart the PC, the GPU works fine for 15-20mins and then the voltage rises up and does not come down (does not matter if the GPU is idle or being stressed - it rises up none the less)

and yes it kinda sucks that there is no other PC that i can test it out on...i shall try a different PCI-E slot and see if that helps (not that i am keeping my hopes up). i am in touch with ZOTAC..let me see if anything can be done about RMA-ing card even in this country.

do you have any other ideas that i might have missed?

 

[junkeymonkey]

wow that thing sucks some power . almost 352w max draw and only 1400mhz clock ? my card at 1400mhz gets up to 265w if I recall .. got to say that 68c looks nice for a air cooled card for sure ,sweet .

your voltage in that shows 1.193 not that 1.274V you reported but when I run my cards overclock I do have that bumped up to 1.2?? v as well .

your screen shots look good to me I think its giving you that Voltage limit has been reached message just because , all that is set in the cards bios and it triggers that when you get like with in 5/10% of the bios set limit .

looks to me like your card is doing a good job just that higher GPU power of 350w but you got a ''aftermarket custom '' pcb card as well mmy card is a reference pcb card so that's something zotac just bumped up to a higher limit .


in this screen shot here did it not finish or what ? [where it seemed to take a nose dive ]

https://s13.postimg.org/woqt8isaf/Stress_test_2.png

now I do know with my card I could not pass that test unless I opened like afterburner ] and just used the voltage % slider and put it up to 110% .

see at stock it was hitting the power limit but adding that 10% buffer kept that ''down clock'' event in check .

see that NVidia GPU boost 2.0 is funny

another thing I thought was odd from that screen shot is when the FPS dropped at that point under the temp graph you see your blue line got cooler but that other line [red / cpu ? ] got hotter ???

now I aint too hip on skylake and I have seen guys have issues and after replacing everything but the CPU and as last ditch RMA there skylake COU back to intel and found that was there issue all along [long shot / opinion of what I seen and read ] so I drawed a conclusion that theres a risk of getting a not so good skylake cpu .

you see how stable mine stayed

http://imgur.com/Ak24Pbi


now one thing you did not add was this screen shot of how your psu voltage on the 12v when the card was at full load

http://imgur.com/AvRnRgU



.

 

[nihan2k]

Hi JM,

Thanks again for solely replying to my cause.

The reason the voltage is low this time around is because I reverted back to stock bios (84.00.36.00.61) as per your previous post and I believe the voltage limit on that bios is set to 1.193V. Prior to this I was using a modded bios which had set the limit higher to 1.274V.

I have set the power limit to 111% and so far everything is going good...but we shall see.

I did notice the CPU temp go higher but unless I have another PC to install this GPU, I cannot rule that out.

With regards to the 12V reading, it was shown on the screenshot : https://s13.postimg.org/vnqkjebav/stress_test_1.jpg
It is at the top left corner and if you see the Stress Test in the background, you would notice the FPS at 17.2. And if you notice the graphs at the corner, you would see the GPU power drop. That was the moment when the Voltage Limit was reached and the stress test tanked. Initially i was getting 98FPS avg, but when the limit was reached, those were the indications.

I am still waiting on a reply from Zotac.

 

[junkeymonkey]

that sounds right when the voltage [at stock] comes with in say 95% of 1.193v the card due to NVidia gpu boost 2.0 as part of how the cards v-bios is

it will start to throttle . with that slider bumped up to say 110% that adds a bit more head room so the with that the card can go up to say it 100% of stock set voltage

I ran firestrike with my overclock profile closest to what your card is [notice I bumped my voltage to 1.230v]

http://imgur.com/kpno2wn

another thing I noticed is how much higher your cards memory stuff is compared to mine [ I'm using regular firestrike for 1080 @60hz [default] not the extreme or ultra ] that may make a difference as well between us ??


heres the thing Im guessing

seeing how your card just drops out as it is from one of your first screenshots and this last one you posted it looks to me your card is not able to hold the factory overclock and is kicking the card to a failsafe ?

it looks like when you do a bad overclock and run a bench/ stress test and the card crashes on it to a failsafe to prevent the card from damage ..

heres what I would try use the NVidia debug mode and with it on run another test and see if your card stays stable all the way ?

http://www.evga.com/support/faq/afmviewfaq.aspx?faqid=59594



dude I just don't know I ran at stock with that graph and don't look right ?? its all over the place up and down , I stay stable no drops during the bench run ? not be 90fps then drop to 17, it all stays constant through out ??? each one of my ''dips'' corresponds each time the firestrike test changes gpu test 1 then test 2 then physics then combined tests ??

http://imgur.com/OufJXAm

in a bit i'll redo that, but I will run the stress test part and see .............

well passed and only drop you see is when the test was over and the time it took to screen shot it ?

http://imgur.com/9p6QuyK

I guess try for that RMA ? maybe the card developed a bad part on it like a voltage regulator or its getting hot or mis reading the sensor that controls it

overall your car kinda looks good nice clocks good temps [68c is nice for a air card and its not like you live in the arctic tundra ,lol...]

I guess just sems like something went haywire with the cards voltage control ?? that's the best I can tell you ..

just for fun why not try this driver out ??

win-10 [I guess you got 10?]

http://www.nvidia.com/download/driverresults.aspx/90500/en-us

win-vista-8.1

http://www.nvidia.com/download/driverresults.aspx/90494/en-us

or something older / different jut to see if its not another NVidia broke something as they have been , but 355.82 iis what I use with win-7 [ seems NVidia has struggled with drivers fix one thing brake something else??]

 

[nihan2k]

Hi JM,

I got a RMA request approved. Now I just have to ship the GPU over there.....its gonna cost be 250$ !!!!! that too one way!!!! lol...

I tried looking for an option here in UAE, but cant seem to fine any

 

[junkeymonkey]

OUCH!!!! I guess like I helped a guy over there with his card and we look at all the manufactures and where they got rma centers seemed like MSI had then more local located in a region ??

so that helped hi on whos card to try to buy so if it went bad he did not have to pay shipping around the world or customs taxes ??

maybe something you should look in to next time , just a thought to ponder

it all ways says notebook no matter msi product I use to find out ?

https://us.msi.com/page/service-location/

AAE-JUMBO
Jumbo Electronics Co. Ltd (LLC)
Address: Jebel Ali Industrial Area 2, Plot No. B136, 7th Interchange
PO Box 3426,Dubai, UAE
TEL: +971 4 884 9074 EXT 435
FAX: +971 4 8849 796

I guess you could call them just for fun and see what they really do if you did have a MSI card ??

post back after you get the replacement card on how things go with it

I guess like evga you would have to send it to the EU ?? $$

http://forums.evga.com/EVGA-Warranty-Dubai-m2503629.aspx

they sure don't like to give up any real info on any of that then looks like its all dealer location thing not manufacture handled ..

 

[nihan2k]

I have emailed Zotac to see if they have an alternate (read: cheaper) option. But surprisingly now, the card works perfectly fine. Had a fun session of Deus Ex :Mankind divided with no hiccups.

I did the stress test again and it passed with no issues. The only thing i did was increase the power limit to 111% but I am certain it was at that point when the whole issue started. Now i am afraid to restart the PC cos i think it might go back to its old self. lol

Here are screenshots for you to check out: https://s16.postimg.org/9u5bgg7vp/Stress_test_passed_main.png
https://s16.postimg.org/7bjmfrm5h/Stress_test_passed.jpg

 

[junkeymonkey]

well my card would not pass the firestrike stress test at first I was scratching my head as well but I opened up evga precision x and slid up that slider to 110% [nothing else ] and re ran the stress test and it passed , but funny is now after doing that I put it back to default [stock] and it still passes with out the power limit increase when it would not before ?????? crazy

well I guess it looks good , and run it and enjoy. I like to say these 900 cards can be ''funny''

good luck

for fun?

http://imgur.com/oXBQD9B

http://imgur.com/BLA7QfW


one more thing I was thinking about was that GPU power from one of your screen shots showing almost 352w

https://s13.postimg.org/vnqkjebav/stress_test_1.jpg

figure your psu and the card needs a pci-e slot power at 75w then 2 8-pin from the psu at 150w each that's 375w

pci-e slot = 75w 6-pin psu = 75w 8 pin psu = 150w

that's getting close and I would guess your PSU or any PSU would have to be spot on to hold up

I don't see how zotac figured up all that like I showed from here ?

https://www.techpowerup.com/vgabios/173848/zotac-gtx980ti-6144-150602

Board power limit
Target: 390.0 W
Limit: 435.0 W
Adj. Range: -74%, +12%


the cards target power is more then the pci-e slot and 2- 8 pin psu can even deliver then add that +12% ?????

maybe I need to get that better explained to me . seems like the card would need a added 6-pin at 75w on top of thee rest to meet that demand ??

maybe like said here ??

''PCIe-PEG Specs state 75w for 6 Pin, and 150w for 8Pin.

Though, a PCIe 8pin connector has the same number of 12v wires (3) and just 2 more ground wires than a PCIe 6pin connector. So right there, theoretically, a 6pin could provide 150w

'http://www.jonnyguru.com/forums/archive/index.php/t-7212.html

 

[nihan2k]

Hi JM! Hope you are having a great new year.

I just wanted to update you that i have received a new card via the RMA process and everything works perfectly fine! it is such a relief.

Zotac analyzed the card that I shipped but they found no fault. Yet they chose to send another card instead. Works fine for me

Thanks for all your help!