• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

3090 zotac Trinity Oc dead or psu weak?

dzmdd

New Member
Joined
Feb 20, 2024
Messages
10 (0.02/day)
Hi guys, I purchased a second hand GPU. It's a 3090 Trinity Oc from Zotac. Sadly this card isn't stable at all, indeed my screens flickering multiple times and many others time he purely crash . I purchased it for a project in AI. I use ComfyUI for make videos and images. When card is in full load Power board draw reach 369 W. My setup is as follows.
- i5 - 12400f
- Seasonic Gc 750 W gold
- 64 gb DDR4 3200mhz
- 1 fan for case, 1 for CPU
- 1 USB dongle bluetooth
- 1 USB dongle ethernet 2.5 gb
- 3 SSD 2.5" / 1 Nvme
- 2 display one in HDMI other in Display Port.
- Zotac 3090 Trinity Oc
I ask myself if GPU is faulty or if Psu is too weak for this setup. As i saw 750W should be okay. When I purchased this GPU I was with the guy we were running OCCT test for 20 mn and I saw him running BO6 really well for check stability and game performance . I don't know to do have you some idea guys for fix crash flickering please ?
Here you can see some GPU-Z sensors details and more details.
Edit : I've changed and cleaned GPU Core thermal paste temp are really since but flickering and crash still here.
 

Attachments

  • IMG_20250207_090813.jpg
    IMG_20250207_090813.jpg
    4.7 MB · Views: 101
  • IMG_20250206_181004.jpg
    IMG_20250206_181004.jpg
    1.9 MB · Views: 98
  • IMG_20250206_180241.jpg
    IMG_20250206_180241.jpg
    4.4 MB · Views: 102
  • IMG_20250209_015057.jpg
    IMG_20250209_015057.jpg
    2 MB · Views: 104
  • IMG_20250209_043145.jpg
    IMG_20250209_043145.jpg
    3 MB · Views: 82
  • Snapchat-1196134766.jpg
    Snapchat-1196134766.jpg
    1 MB · Views: 98
Can't say for sure. Most likely it's caused by OCP tripping due to power spikes. I'd suggest testing with an overkill PSU if you can get one without too much trouble.

Right now, you can lower the power limit and decrease the VRAM clocks and see if it makes any difference.
 
The seller probably had a better quality PSU. Something like a RM850x will be solid and have room to spare for ease of mind.
 
Thanks, I'll try to try that ^^ I mean for try on a bigger psu. Actually I use Core clock at his boost default clock 1710 Mhz and I lowered limit power by - 20 %. I had the illusion it was better but finally problem isn't fixed at all. I hope one of my friend could help me with a bigger PSU for check if GPU is faulty. Seller had same PSU than me Toothless ^^ he have a 750 w Cooler Master. Thanks in all case dude.
Edit : For Vram clock I'll try that I hope it will be good my bad I was talking about Core clock ^^
 
Last edited:
62 Amps on 12V line for output of 744 Watts
 
The seller probably had a better quality PSU. Something like a RM850x will be solid and have room to spare for ease of mind.
Apparently the SeaSonic G12 is being made by 1ST Player/Helly







That cards Target tdp is 350W, max is 385W apparently...

Which would leave 365 Watts remaining for rest of your system.

Please can you give me more details about this please. I really need to solve this problem
Well since the person who sold you the card had the same psu as you did and they claimed it was working then, I would just return it for money back and look at a brand new in box card.
 
Last edited:
Indeed it's true but when I lowered power limit this happened. Maybe power spikes is too high I don't know I saw in TechPowerup review some minute power spikes go to 402 W in furmark.
I think I'll probably return it yes. Or I will use his guarantee. I'm really disappointed. I thought this guy was really serious but sadly as I saw I'm not really sure now. :(
 
Last edited:
Apparently the SeaSonic G12 is being made by 1ST Player/Helly







That cards Target tdp is 350W, max is 385W apparently...

Which would leave 365 Watts remaining for rest of your system.


Well since the person who sold you the card had the same psu as you did and they claimed it was working then, I would just return it for money back and look at a brand new in box card.
Age could've made the caps a little less durable on spikes. Max power output could be half of total system use and still not handle GPU power spikes if not built well enough or aged enough.

Kinda like when my GT1300w couldn't handle my card when the HX850 still holds up.
 
Age could've made the caps a little less durable on spikes. Max power output could be half of total system use and still not handle GPU power spikes if not built well enough or aged enough.

Kinda like when my GT1300w couldn't handle my card when the HX850 still holds up.
I'm believing it's really not the psu but another case of a dishonest seller of a gpu. Also from feature set the psu does have ocp and ovp amongst others.

GT1300, who is it by?
 
Hi guys, I purchased a second hand GPU. It's a 3090 Trinity Oc from Zotac. Sadly this card isn't stable at all, indeed my screens flickering multiple times and many others time he purely crash . I purchased it for a project in AI. I use ComfyUI for make videos and images. When card is in full load Power board draw reach 369 W. My setup is as follows.
- i5 - 12400f
- Seasonic Gc 750 W gold
- 64 gb DDR4 3200mhz
- 1 fan for case, 1 for CPU
- 1 USB dongle bluetooth
- 1 USB dongle ethernet 2.5 gb
- 3 SSD 2.5" / 1 Nvme
- 2 display one in HDMI other in Display Port.
- Zotac 3090 Trinity Oc
I ask myself if GPU is faulty or if Psu is too weak for this setup. As i saw 750W should be okay. When I purchased this GPU I was with the guy we were running OCCT test for 20 mn and I saw him running BO6 really well for check stability and game performance . I don't know to do have you some idea guys for fix crash flickering please ?
Here you can see some GPU-Z sensors details and more details.
Edit : I've changed and cleaned GPU Core thermal paste temp are really since but flickering and crash still here.
You have power problems. Your PSU is of good quality but might have a problem. You should likely replace it and get a 850w minimum. If the card was the problem it would be exhibiting other issues as well, problems you would have noticed and mentioned.
 
Last edited:
Recent graphic card tests recommend 1050Watt psu. If i were you I would not buy something below 1050 Watts ATX 3.1 PSU.
It's up to.

When I purchased this GPU I was with the guy we were running OCCT test for 20 mn and I saw him running BO6 really well for check stability and game performance .

When that'S valid ... ? I would not buy a graphic card which costs more than 30 € used.

I'm not fond of prime mainboards - #1 indicates a Asus prime mainboard, one of those screenshots
 
I'm believing it's really not the psu but another case of a dishonest seller of a gpu. Also from feature set the psu does have ocp and ovp amongst others.

GT1300, who is it by?
EVGA.

Recent graphic card tests recommend 1050Watt psu. If i were you I would not buy something below 1050 Watts ATX 3.1 PSU.
It's up to.



When that'S valid ... ? I would not buy a graphic card which costs more than 30 € used.

I'm not fond of prime mainboards - #1 indicates a Asus prime mainboard, one of those screenshots
Not everyone is buying a 5090. Saying a 1050w is needed is absurd.
 
Saying a 1050w is needed is absurd.

And why does the poster has issues? I also had a faulty trash corsair rm750 psu - which was hard to see.
Someone who has the cash for a 3090 should have cash for a proper mainboard and proper psu. high end parts with low end parts do not really fit.

edit: I'm glad I got rid of " Corsair RM750 PC-Netzteil (Voll-Modulares Kabelmanagement, 80 Plus Gold, 750 Watt, EU) schwarz" -> total trash with random crashes (purchase Tue, 17 Aug 2021 -> dead after 1 year some months) ---- just saying random issues - psu is nasty to find as root cause

- Seasonic Gc 750 W gold
 
Last edited:
And why does the poster has issues? I also had a faulty trash corsair rm750 psu - which was hard to see.
Someone who has the cash for a 3090 should have cash for a proper mainboard and proper psu. high end parts with low end parts do not really fit.
RM750 but which? Of the RM line the RM---x is the better quality, then it goes to HX which would absolutely handle a 3090.

Also we don't know how much they paid for it. Maybe it was cheap because it actually has issues or whichever.
 
Edit : I've changed and cleaned GPU Core thermal paste temp are really since but flickering and crash still here.

Well seller sold "a good graphic card" which was verified by the seller and buyer with "proedure" 20 min of testing some software. Regardless if valid or not.

Buyer "opened" the graphic card after second hand purchase and has issues. 3rd root cause may be opening the graphic card and not proper handling it. (no offense)

Seller had same PSU than me Toothless ^^ he have a 750 w Cooler Master.

*different* brand power supply
It's not all about the maximum Wattage of a psu.
 
Folks, a 750w should be enough as suggested by TPU's own specs page.
Even Zotac's own specs page states 750w;

However, that is the recommended MINIMUM. Where PSU's are concerned, going with the minimum is effectively asking for future problems.
In the case of the OP, they stated they only have problems under load, therefore, the card is not getting the power it needs.
The problem here is clearly the PSU. The OP needs a better PSU.
My above recommendation of a 850w unit is for a reasonable margin of error. A better margin of safely would be a 900w, 950w or 1000w unit.

It's not all about the maximum Wattage of a psu.
It can be and frequently is.
 
Thanks for responding guys. Indeed I use a Asus PRIME 610H. This mobo cost me something like 90 euros. I know it's a cheap mobo now. When I purchased it I wasn't really good for recognize low tier, mid tier and high end mobo now i know. _Roman_ I understand but I've built this pc little by little. I'm still learning and you know I play only World of Warcraft xD. 3090 will be used only for creative AI stuff. . My problems appears before I opened card, many times it happened on full load when I want create a video or image. I done repast many times on my own gpu. Many times on my old gpu nvidia 1060,3060,6800xt and on Cpu. I sold 1060 and 3060 without any problem. I'll try to convince my best friend uncle to let me test this 3090 on his setup. Now I hope he will have a bigger psu than me ^^. Thanks guys a lot.
If you have some idea about a good PSU brand tell me please. I think about Corsair but some people talk about multi rail security issue.
 
Last edited:
You really don't need to worry about multiple rails. You're running one GPU.
 
This problem is really weird. When I run OCCT Power test everything is oley but when I generate image it crash... I asked my neighborhoods son if he could test this card he work in this domain. I tried lower memory clock by 219 - Boost clock by - 250, limit power at -80 %. Same result crash in loop. Maybe Vram is faulty Vram isn't loaded at all on Power OCCT test so maybe it don't provoke error in power test. I think more and more GPU is fried, my RX6800xt overclocked with MPT use even more watts and it don't crash at all during heavy load (350-360 watts pick sometimes). Today I get a BSOD testing to generate video too.
 
Use the 6800xt then send that 3090 back
 
GPU stats appear fine based on GPU-z shown.
It's either a PSU issue or a bottleneck from the i5 CPU.
 
Hi guys, I purchased a second hand GPU. It's a 3090 Trinity Oc from Zotac. Sadly this card isn't stable at all, indeed my screens flickering multiple times and many others time he purely crash .
I ask myself if GPU is faulty or if Psu is too weak for this setup.I don't know to do have you some idea guys for fix crash flickering please ?
Here you can see some GPU-Z sensors details and more details.
Edit : I've changed and cleaned GPU Core thermal paste temp are really since but flickering and crash still here.
If the PSU is too weak, the PC should shut down or restart.
What is this crash you are speaking of? I've seen crashes refer to shut down, restart, bsod, driver timeout, etc. Try to be more specific.
 
Recent graphic card tests recommend 1050Watt psu. If i were you I would not buy something below 1050 Watts ATX 3.1 PSU.
It's up to.
Which "recent graphic card tests"? You can always run your own power station. Just to be on the safe side.
 
try looking into 3.3V monitoring on the Hwinfo64.

I have a combo of Seasonic PSU and Asus MoBo (several years older then yours though) and running in the issue of the 3.3V is running as low as 2.9 during 350W+ GPU load. MoBo detects it and shuts down the system.

The actual reason is quite quircky, since there is NO any high load on 3.3V line. I suppose that the PSU tries to use too clever wire-voltage-drop compensation for 12V that somehow negatively affect the 3.3 levels.

When I got those low 3.3 readings - I reseat the main 24-pin connector several times (both sides, MoBo and PSU) and after this the issue disappeared
 
Back
Top