Question 750W PSU can't handle RTX 3080?

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.
Sep 27, 2020
16
3
15
Hi! So like the title states, I'm under the impression that my Corsair RM750x is not able to keep up with the addition of my new Zotac RTX 3080 Trinity. I've posted this issue in other forums but haven't gotten to a reasonable conclusion as of yet. The main issue I'm facing is that my computer will completely power cycle at random when under load. I've attempted to solve this with fresh driver installations, Bios settings reset and more. I tried plugging my computer directly into the wall because I thought maybe my surge protector was getting saturated even though it's rated for 1875 Watts. At first I thought this solved my issue as I went a full 24 hours without a shutdown, then it happened again... I tried reproducing the issue to test certain fixes but I have been unable to do so. My computer has only shutdown during gaming at random times. I cannot get it to shutdown in a benchmark. For reference, these are the benchmarks I've run: Time Spy Extreme, Heaven, FurMark, OCCT Power Test, and I also ran Cinebench and Prime95 to make sure I wasn't imagining it was a GPU related issue. Like previously stated, I couldn't get any of these to shutdown my PC. So, part of me is unsure if this is even a PSU related issue. The solutions I have been recommended have all been for the case of people BSODing or crashing to desktop for I suppose GPU instability at 2+ GHz or driver issues. I haven't had this issue so I don't believe I'm facing the same problem as those folks. I would also Like to note that my PC previously worked fine with my Gigabyte RTX 2080 Gaming OC. Here is a list of my core specs:
  • CPU: I7-9700k (stock clocks)
  • AIO: Corsair H100i RGB Platinum SE
  • GPU: Zotac RTX 3080 Trinity (Stock Clocks)
  • RAM: 48 GB Corsair Vengeance RGB Pro @ 3200 MHz
  • MOBO: MSI MPG Z390M Gaming Edge AC
  • PSU: Corsair RM750x
I should also note that I don't think it's an overheating issue considering that with a good aggressive fan curve, my 3080 doesn't go above 74 C (which still baffles me as some folks claim they don't go above 65 C with stock settings on the same card as me). My CPU temps I believe are also not the issue, although it does run quite hot (70 - 85 C under load) despite regular remounting of the water block with thermal paste reapplication, but I suppose that's an issue for another thread. If you have any insight into this issue beyond the basics, I would greatly appreciate your help, I would like to know as definitively as I can if it IS or IS NOT my PSU that is causing the issue before I go out and replace it. I know it can't be said for sure if that's so, but I'm sure my problem can be narrowed down greatly.
 
  • Like
Reactions: joshuacant
Oct 2, 2020
4
7
15
I have seen the exact same issues. Recently installed a 3080 FTW into my rig. I'm running a Seasonic 750w Prime Titanium PSU. SSR-750 to be specific which was purchased about 3 years ago. After installing the 3080 I've experienced a few PC restarts where the system just turns off/on. It seems like the PSU is hitting some limit and resetting itself. I can't pin it down however. Doesn't happen in all games and happens infrequently enough (once a day) that I can't reproduce it.

I even went as far as to load up a loop of Cinebench while looping 3D Mark to put both my CPU and GPU into a max power state. I was pulling 690W at the wall with with this setup, but I ran the loop for 30 minutes and no issues. I then load up a game where the CPU is around 40% and GPU is sitting at 70-90% and about an hour in it power cycles. Other times, I can play a game with the card around 99% and no issue. System was rock solid with the old 2080 Ti.

It is interesting to see others having this random problem. I have a new 1000w PSU on the way to test with but I am extremely curious as to way these failures are happening. With the power draw at the wall, the PSU should have enough power to make it without issue, I wonder if there is something else at play such as significantly high transient power spikes.

I'll update once I do testing with the new PSU.
 

Gussv

Commendable
Apr 4, 2019
23
3
1,515
I'll do much more testing tomorrow and over the weekend. If I get a shutdown, I'll reply here immediately. Otherwise look for another post from me by the end of Sunday with more info.
Will look closely for your new findings.

I find it hard to believe the issue was caused by your PSU. Seasonic is a pretty respectable brand and 750W should be enough for a 3080. I reckon it must be the cap thing in the AIB cards.
 
Last edited:

Juular

Respectable
BANNED
Mar 14, 2020
1,061
258
1,940
After installing the 3080 I've experienced a few PC restarts where the system just turns off/on. It seems like the PSU is hitting some limit and resetting itself.
If the system restarts then it's not a PSU but something else, did you update nVidia drivers to recent version ? It's supposed to fix Ampere instability issues.
 
Oct 2, 2020
4
7
15
Swapped in the 1000w unit yesterday. Still too early to call it, but I've been able to game for much longer sessions than before and have not seen a power cycle happen. Going to do more testing over the coming week, but it does appear the 3080s can get quite power hungry and 750w PSUs, even high end ones, may not cut it at all times depending on the rest of the system.
 
Sep 27, 2020
7
6
15
Not a single shut down all weekend long. I haven't done any real marathon 4+ hour long sessions but I used it all 3 days in various games. Today I loaded up a game (Metro Exodus), set the graphics settings to max out the GPU where the framerate was bouncing between 50 and 60. I left it running while I was doing other things around the house, occasionally playing for a few minutes to move the game into new areas/cutscenes and hopefully spike CPU/GPU in random ways. It was up and stable for well over 2 hours.

Watts and temps don't look any different from before, all perfectly within acceptable limits. The only difference is the PSU no longer trips and shuts everything down at random.

I opened an RMA with Seasonic on my SSR-750PD2, explained the situation, and they accepted it. I'm going to mail it back to them and see if they find any faults or return it as-is saying it's perfectly fine. I'm genuinely curious.

@brandonc0431 and @demingo I'll keep watching this thread to see if replacement PSUs solve the problem for you as well.
 
  • Like
Reactions: brandonc0431
Sep 27, 2020
7
6
15
@demingo we must have been typing our replies at the same time, lol. I also plan on doing more testing, I don't think I'll feel like the problem is truly behind me until I've got at least 100 hours of gaming between me and the last random shut down. That said, given the frequency with which I was seeing them on my previous PSU, I should have had 10 to 20 of them in the last 3 days.

What CPU do you have? Do you have any other hardware that might be drawing lots of power or drawing power in unusual ways? I have a 3950X with no overclocks, no platter drives, and nothing else that's power hungry...
 
Sep 27, 2020
16
3
15
Not a single shut down all weekend long. I haven't done any real marathon 4+ hour long sessions but I used it all 3 days in various games. Today I loaded up a game (Metro Exodus), set the graphics settings to max out the GPU where the framerate was bouncing between 50 and 60. I left it running while I was doing other things around the house, occasionally playing for a few minutes to move the game into new areas/cutscenes and hopefully spike CPU/GPU in random ways. It was up and stable for well over 2 hours.

Watts and temps don't look any different from before, all perfectly within acceptable limits. The only difference is the PSU no longer trips and shuts everything down at random.

I opened an RMA with Seasonic on my SSR-750PD2, explained the situation, and they accepted it. I'm going to mail it back to them and see if they find any faults or return it as-is saying it's perfectly fine. I'm genuinely curious.

@brandonc0431 and @demingo I'll keep watching this thread to see if replacement PSUs solve the problem for you as well.
I've just put in an order for Corsair's RM850x through BestBuy (It was $100 more on Amazon which is insanity). I will report back here once I install the new PSU and do some testing. I've found that I see at least one shutdown per day, hopefully that will change. I also hope that Corsairs modular cables are the same between products so I can just leave the old cables... I really don't want to route the cables again.
 
Sep 27, 2020
7
6
15
I also hope that Corsairs modular cables are the same between products so I can just leave the old cables... I really don't want to route the cables again.
I fired off an email to Seasonic asking them the same thing and got a very helpful reply back before the new one arrived. Maybe you could do the same with Corsair, it's not one of those things you want to risk, lol.
 
Sep 27, 2020
16
3
15
I fired off an email to Seasonic asking them the same thing and got a very helpful reply back before the new one arrived. Maybe you could do the same with Corsair, it's not one of those things you want to risk, lol.
Based on this chart (https://www.corsair.com/us/en/psu-cable-compatibility) it looks like all RMx PSUs use the same cables/connectors. I'll leave it here for anyone else who wishes to know.
 
  • Like
Reactions: joshuacant
Oct 2, 2020
4
7
15
@demingo we must have been typing our replies at the same time, lol. I also plan on doing more testing, I don't think I'll feel like the problem is truly behind me until I've got at least 100 hours of gaming between me and the last random shut down. That said, given the frequency with which I was seeing them on my previous PSU, I should have had 10 to 20 of them in the last 3 days.

What CPU do you have? Do you have any other hardware that might be drawing lots of power or drawing power in unusual ways? I have a 3950X with no overclocks, no platter drives, and nothing else that's power hungry...

I'm running a 8086k at 5ghz with 1.325vcore. Other than that, just a few fans and and a pump for the WC loop. I'm definitely leaning toward the 3080 just needing more power at times. I had this exact same config running with a 2080ti for 2 years without so much as a hiccup. I just think the extra power the 3080 is sucking down stressed the PSU too much. I do have an open ticket with Seasonic to see what they say. Still waiting on a response from them.
 
  • Like
Reactions: joshuacant
Oct 5, 2020
4
2
15
I've just put in an order for Corsair's RM850x through BestBuy (It was $100 more on Amazon which is insanity). I will report back here once I install the new PSU and do some testing. I've found that I see at least one shutdown per day, hopefully that will change. I also hope that Corsairs modular cables are the same between products so I can just leave the old cables... I really don't want to route the cables again.

I guess i am in the same boat as all you, bought a 750 evga gold supernova GA for this 3080 and i am getting hard reboots with ffxv and witcher 3. I ordered the same replacement psu as you as as my evga psu is still within a return window. My specs are:
I5 8600k oced to 4.8ghz at 1.3v
3080 fe
750 evga gold GA.
System was running fine before on a 1080ti and a cx750m (green label)

I will keep an eye out on your results
 
  • Like
Reactions: joshuacant
Oct 2, 2020
4
7
15
Just got done working with Seasonic support. They are approving an RMA to replace the PSU with a "newer model". I'll update with the specific changes once I complete the RMA. It's looking more and more like there is some kind of compatibility issue between some PSUs and the 3080s.
 
Oct 5, 2020
4
2
15
rmx850 got here today. I maxed out power limit on afterburner because that would crash my pc instantaneously upon opening any game and i did not want to wait and see when i would crash. So far i have ran FFXV benchmark 6 times on max power limit and have not crashed once when before it would load for a literal second before it would completely reboot my pc. I will respond if i encounter a crash but things are looking better..
 
  • Like
Reactions: joshuacant

shrapsky

Prominent
Oct 5, 2020
19
6
525
I'm running into the same problem with a RTX 3090 FE. It only seems to occur when playing Fortnite. I've tried playing some COD (~1 hour straight) and had no issues. But randomly while playing Fortnite I get a hard shut down of the entire computer. No indications it's occurring, and temps/usages all seem to be normal throughout gaming, even up to the crash. I've tried updating all drivers and Bios, resetting RAM, and uninstalling/reinstalling Fortnite. Nothing has corrected the issue.

I'm unable to cause it to happen, it's totally random. Sometimes it's immediately upon starting to play, sometimes it's an hour or so into gaming.

One thing I've noticed that I'm not sure is normal/correct, but on the task manager, it's showing my GPU usage as 0-2% when playing Fortnite.

PC Specs -
CPU - Ryzen 9 3900x
GPU - NVIDIA RTX 3090 FE
PSU - EVGA 850w Gold+
RAM - G.Skill Ripjaw V 2x16
Mem - 2x WD Black 750 SSDs
 
  • Like
Reactions: joshuacant
shrapsky Hello. I'm not particularly concerned about task manager showing your GPU usage as 0-2%. In the past I've noticed that task manager incorrect reads GPU usage with newly released GPUs.

As for your power supply, I see that you've listed it as an EVGA 850 watt unit. However, could you please state the full model number/name of the unit, or link to it? EVGA doesn't actually manufacturer their power supplies. They have several different companies manufacturer their many lines of PSUs, therefore different models have vastly different quality standards.

Is this an entirely new build, or it was existing and you simply added the RTX 3090? Did you download Nvidia's fix for crashing-to-desktop? If not, then you should run DDU and reinstall the latest drivers.

What is your specific motherboard model? Check motherboard user guide and verify that you've installed RAM in the indicated slots. What speed are your RAM modules rated for from the manufacturer? If you've enabled XMP, which overclock speed have you set them to? I once mistakenly placed the RAM in the wrong slots, which only became a very bad problem after enabled XMP and overclocked the modules. The system became unstable to the point of no longer posting; I'm not exaggerating.

What BIOS version and date is your motherboard updated to? Did you download and install the latest chipset drivers directly from AMD?

Just to be clear, what precisely is happening with your computer. Does all power to both the computer and monitor turn off? Is only the display turning off but the PC has power? Or is the PC rebooting itself? I once had a problem where one of my computers would restart after ten minutes of simply updating Fortnite, not even running the game. In my specific case, a faulty EVGA P2 850 power supply was the culprit. EVGA shipping me a new unit under warranty solved the issue. Speaking of which, you should register your PSU on the EVGA website, regardless whether it's the cause of your problem or not.

I believe in a shot gun approach to solving issues:

Download fresh copy of the Nvidia game-ready drivers from Nvidia; do not run installation yet.
Download DDU (Display Drivers Uninstaller) from Wagnardsoft (publisher).
Run DDU; preferably in Windows safe mode.
Now install Nvidia game-ready drivers that you downloaded moments ago.

Open your case and take two photos: (1) of the power supply label, and (2) of the entire inside of your build (to show air flow). Then upload those photos to Imgur and share the link here. These are good start, and then we can from there.
 

shrapsky

Prominent
Oct 5, 2020
19
6
525
@rcald2000

PSU - https://www.evga.com/products/product.aspx?pn=220-GA-0850-X1
MOBO - Gigabyte Aorus x570 Pro WiFi - https://www.gigabyte.com/us/Motherboard/X570-AORUS-PRO-WIFI-rev-10#kf
RAM - G.Skill Ripjaw V DDR4 3600 2x16

This is an entirely new build, and I do not have any other GPUs to put into play here. I have the GeForce Experience downloaded, and up to date on all the latest driver's. Was the CTD issue addressed/corrected in a different manner? (Although from what I've read and researched the CTD issue wasn't found in FE models, and what I'm experiencing isn't the CTD that was reported. This is a full shut down).

With the RAM, I have them in the correct slots per the manual. I also have had the issue when XMP was enabled and disabled.

I went through and updated the Bios from Gigabyte the other day and it is the latest version. Chipsets I haven't done and will have to look into checking out if there is any updates for those.

The exact issue - when playing I will randomly get a hard shut down of the computer. There is no error message, or indication this is going to happen. Sometimes it's in the 1st game I'm playing, sometimes it's an hour into playing. Out of nowhere, the entire PC shuts down, monitors, tower, etc. The light around the power button on top of my case stays lit, and a light on the MoBo stays on, but everything else shuts down. I have to push the power button to boot the pc back up, it does not restart on it's own.
 
EVGA SuperNOVA 850 GA: Ah. You may or not believe this, but "just" this morning I was trying to figure out which company manufacturer's the GA line. I'm not sure who it is, but It's definitely not Super Flower.

What I would do: Contact Register your power supply on EVGA's website. Then call EVGA support ASAP and request an "advanced RMA" warranty replacement for your power supply (714) 528-4500

The "advanced" RMA option means that they would send you a replacement (probably refurbished or recertified) power supply first, and then you send your original PSU back to them at your own cost. They will definitely request your credit card number as collateral, in case you don't send them the damaged PSU back. That is what I would do.

The other real possibility is that the motherboard is to blame. Another interesting point: I strongly considered purchasing your exact Mobo, before finally deciding on my MSI board. Yours is very nice though.

** Addendum **

Question: In addition to the 24 pin connector, did you also connect both the 8 pin and 4 pin connectors to the power supply? Ordinarily, only the 8 pin connector is required and the 4 pin is only needed for extreme overclocking. However, I wonder if both connectors are required with the RTX 3080. I just thought it was worth me mentioning.

** Addendum #2 **

shrapsky
In your latest response, somehow I missed the last paragraph of what you wrote until now. I'm now doubting everything I've previously stated about your issue because I'm no longer certain that your system is actually shutting down. Before reaching out to your PSU or motherboard support, do this first: The next time that it "shuts down", open up your computer case and check if the system fans (e.g. cpu and system case fans) are spinning. It's vitally important to determine whether the system is truly shutting off, or simply going into a standby mode, that it can't be awoken from. Also, I strongly recommend that you immediately turn off both hibernation and windows stand by. Then simply "change screen saver" to turn on a blank screen saver after say 30 minutes, or so. If your system is going into standby or hibernation, and can't be awoken out of it, then that's an entirely different issue, and is most likely software related.
 
Last edited:

shrapsky

Prominent
Oct 5, 2020
19
6
525
I have both the 8 and 4 pin connections on, that had not corrected the problem.

When the shut downs occur I'm 99% certain the fans and everything completely stop. I will have to confirm this next time the crash occurs. I will look into the hibernation and standby mode settings as well.


I went to Microcenter and they believed it was a PSU issue as well, so I'm going to try and upgrade it to a 1000w PSU and try that out. I'll let you know if there's any changes.
 
Oct 5, 2020
4
2
15
Im convinced that GA model does not play well with the new 3000 series. I saw at least 2 people on reddit with that psu(both new) that were having the same problem.

I also was using the GA model and was having the same problem.
 
Is Microcenter where you purchased the EVGA GA 850 from? If so, are they going to let you swap to another unit without paying a restocking?

If the power supply is actually the problem, then I recommend that you upgrade to a higher quality model, instead of a higher wattage one. I honestly believe that 750 watts is more than enough wattage, as long as the model chosen is high quality. But if 850 helps you feel more comfortable, then go for it.

** Units I see on Microcenter site **
Seasonic USA Focus Plus 750 Watt
Corsair RM750x 750

** Amazon **

Super Flower has the reputation of being one of the best power supply manufacturers in the world, and they recently started selling in the US for the first time in years. Seasonic also has a stellar reputation.
* FYI - the labels on gold rated super flower psu's are ugly. This wouldn't be an issue for me, but just thought that I would mention.

Super Flower Leadex III Gold 850W 80+ Gold, ECO Fanless & Silent Mode, Full Modular Power Supply, Fluid Dynamic Bearing Fan, SF-850F14HG (850W)
$149.99
https://www.amazon.com/Super-Flower-Fanless-Modular-SF-850F14HG/dp/B07THY11J1

Seasonic FOCUS GX-850, 850W 80+ Gold, Full-Modular, Fan Control in Fanless, Silent, and Cooling Mode, Perfect Power Supply for Gaming and Various Application, SSR-850FX.
$139.99
https://www.amazon.com/dp/B07WVMDZMZ/?coliid=I9L0Y756E6V91&colid=2E863IC0KL4RY&psc=1
 
Last edited:

shrapsky

Prominent
Oct 5, 2020
19
6
525
Is Microcenter where you purchased the EVGA GA 850 from? If so, are they going to let you swap to another unit without paying a restocking?

If the power supply is actually the problem, then I recommend that you upgrade to a higher quality model, instead of a higher wattage one. I honestly believe that 750 watts is more than enough wattage, as long as the model chosen is high quality. But if 850 helps you feel more comfortable, then go for it.

** Units I see on Microcenter site **
Seasonic USA Focus Plus 750 Watt
Corsair RM750x 750

** Amazon **

Super Flower has the reputation of being one of the best power supply manufacturers in the world, and they recently started selling in the US for the first time in years. Seasonic also has a stellar reputation.
* FYI - the labels on gold rated super flower psu's are ugly. This wouldn't be an issue for me, but just thought that I would mention.

Super Flower Leadex III Gold 850W 80+ Gold, ECO Fanless & Silent Mode, Full Modular Power Supply, Fluid Dynamic Bearing Fan, SF-850F14HG (850W)
$149.99
https://www.amazon.com/Super-Flower-Fanless-Modular-SF-850F14HG/dp/B07THY11J1

Seasonic FOCUS GX-850, 850W 80+ Gold, Full-Modular, Fan Control in Fanless, Silent, and Cooling Mode, Perfect Power Supply for Gaming and Various Application, SSR-850FX.
$139.99
https://www.amazon.com/dp/B07WVMDZMZ/?coliid=I9L0Y756E6V91&colid=2E863IC0KL4RY&psc=1


Yes, I bought the 850w at Microcenter too. When I went in and spoke with one of the managers today, he said that the 850w would be strained to try and run the 3090 and Ryzen 9. He said I would need something more in line with a 1000-1200w PSU. So I bought a Corsair HX1000i PSU, since it was the one they had in stock they recommended. They said it would be no problem to bring the 850 back and return it.

After switching the PSUs, I played about 1.5 hours of Fortnite without issue. It's too early to tell if it's 100% fixed until I put some more hours on it.
 
  • Like
Reactions: joshuacant
Sep 27, 2020
7
6
15
Every PSU manufacturer's calculator, and several independent ones, all said that 750w should be enough for my 3950X and 3080. 850w even more so. 1000w is honestly overkill, but that's what I went with due to inventory issues. Not all sizes of good PSUs are in stock everywhere now, so you kinda have to go with what you can get your hands on.

My 750w that failed was a Seasonic Prime Platinum that I bought just 9 months ago. A quality brand, not very old, and worked flawlessly until I upgraded from a 1080ti to 3080, which put an additional 50-60w strain on the PSU.

So either all the calculators are wrong or every manufacturer of PSU has units out there that are falling down under high loads that should be within their spec. I'm inclined to think it's the latter simply because we're not seeing hundreds of people with the issue. There's this thread with a half dozen or so of us, but that's it. Even with the limited supply of 30-series cards, you'd expect more people to be screaming about this.

So far everyone here who has (hopefully) solved the problem replaced their PSU with a larger one, which makes it hard to know if it's the capacity that was the problem. I'm hoping someone does do a like-for-like swap with 750w so we can find out what happens.
 

shrapsky

Prominent
Oct 5, 2020
19
6
525
After a solid 2 days of playing with the new 1000w PSU in place, I'm happy to report that I've had no problems. Played 2 separate sessions of 2.5-3 hours without a restart, which definitely would have had at least 1 with the old PSU during those durations. So I'm going to chalk it up to being a PSU problem. Unsure whether it was a defective 850 that I had, or if the 850 just wasn't powerful enough to handle everything.
 
Status
Not open for further replies.

TRENDING THREADS