Question 750W PSU can't handle RTX 3080?

[brandonc0431]

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.

 

[RealBeast]

While you cannot know for sure until you try, 750W is the minimum recommended PSU.

 

[brandonc0431]

While you cannot know for sure until you try, 750W is the minimum recommended PSU.

I would Like to get a new 850 Watt PSU to see for myself, but I don't have a reliable testing point at which my PC, or perhaps PSU, fails.

 

[RealBeast]

I would Like to get a new 850 Watt PSU to see for myself, but I don't have a reliable testing point at which my PC, or perhaps PSU, fails.

Did you read the link at 750W for background on their power issue?

 

[brandonc0431]

Did you read the link at 750W for background on their power issue?

Yes, and according to the table provided by Asus in that article I should be okay. But, that evidently is maybe not the case. The fact that I can't get my PC to power cycle under prolonged full synthetic load also greatly confuses me.

 

[sizzling]

Yes, and according to the table provided by Asus in that article I should be okay. But, that evidently is maybe not the case. The fact that I can't get my PC to power cycle under prolonged full synthetic load also greatly confuses me.

Synthetic benchmarks often only load the gpu and not the rest of the system. An aging or faulty psu could definitely be the problem. However it’s wrong to say the minimum NVidia recommend is a 750w, against the 750w rating they also say “2 - Requirement is made based on PC configured with an Intel Core i9-10900K processor. A lower power rating may work depending on system configuration.” The 10900k is a cpu that can peak at over 250w.

 

[brandonc0431]

Synthetic benchmarks often only load the gpu and not the rest of the system. An aging or faulty psu could definitely be the problem. However it’s wrong to say the minimum NVidia recommend is a 750w, against the 750w rating they also say “2 - Requirement is made based on PC configured with an Intel Core i9-10900K processor. A lower power rating may work depending on system configuration.” The 10900k is a cpu that can peak at over 250w.

That's a fair point for most benchmarks I ran, but I've also been running OCCT's power test which puts full load on the whole system, which is more stress than it sees while playing any games...

 

[sizzling]

That's a fair point for most benchmarks I ran, but I've also been running OCCT's power test which puts full load on the whole system, which is more stress than it sees while playing any games...

That is odd, could be problem with the swings in load that real world gaming puts on the psu. A good 750w should have plenty of headroom for your system but the description of the fault immediately shouts a psu problem. It could be something else but if my system I’d start with the psu.

 

[brandonc0431]

That is odd, could be problem with the swings in load that real world gaming puts on the psu. A good 750w should have plenty of headroom for your system but the description of the fault immediately shouts a psu problem. It could be something else but if my system I’d start with the psu.

So I suppose my best option is to get new PSU, most likely 850 W, and see how it goes. I can always send it back if it isn't the issue. This is just a little tricky though as sometimes I don't see any issues. I Played RDR2 for a total of 3 - 4 Hours twice just yesterday, but booted up FS 2020 around 30 minutes later just to see it shutdown after 10 minutes in the main menu. Really unpredictable it is.

 

[sizzling]

I’d do the same. Not sure what else to suggest. If it’s not the psu I’d start to suspect the motherboard.

 

[ajr1775]

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
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.

If the power supply is the issue it's not because of its power rating. It may likely have an issue. Many users running 750W power supplies without issue. Looking at your build it doesn't look like the other components are eating up that much power. I think you have either a bad power supply or a bad Zotac 3080. What kind of of caps are on the back of the GPU processor?

 

[joshuacant]

So I suppose my best option is to get new PSU, most likely 850 W, and see how it goes. I can always send it back if it isn't the issue. This is just a little tricky though as sometimes I don't see any issues. I Played RDR2 for a total of 3 - 4 Hours twice just yesterday, but booted up FS 2020 around 30 minutes later just to see it shutdown after 10 minutes in the main menu. Really unpredictable it is.

I have this exact same issue with an EVGA 3080 XC3. I have a Seasonic PRIME PX-750 SSR-750PD2 that I bought in January and your description of events matches mine perfectly.

Never shuts down when doing normal tasks. Never shuts down during benchmarks. Sometimes a game will run for quite a while with no issues, sometimes it will shut down in the menu. I used my phone to record the output of hwinfo (multiple times) while the game was running, so I could see exactly what the watts/etc were in the last frame before the PC shut down. Nothing looks out of the ordinary. Watt readings for CPU and GPU are fine. CPU, GPU, RAM, and chipset temps all fine. GPU temp in the low 70s, everything else in the 50s.

On paper all the math says a 750 should be able to handle my total system wattage. I have the PC plugged into an APC UPS with an LCD screen, and I watched the total wattage load while playing a game, and it never went higher than 600w. Seasonic Prime supplies are very high quality and it never gave me any issues all year long with a 1080 TI FTW3. I am not running any overclocks on CPU or GPU, just the XMP profile for my RAM, which has been perfectly stable all year long.

I'm not happy about it, but I ordered a new Seasonic PRIME TX-1000 SSR-1000TR to see if more watts fixes the problem. If that doesn't fix it, then I guess the problem could be the card itself, but I'll cross that bridge when I come to it. The new PSU should be here Wednesday of this week, and I'll reply here with the results.

 

[squanky]

I have this exact same issue with an EVGA 3080 XC3. I have a Seasonic PRIME PX-750 SSR-750PD2 that I bought in January and your description of events matches mine perfectly.

Never shuts down when doing normal tasks. Never shuts down during benchmarks. Sometimes a game will run for quite a while with no issues, sometimes it will shut down in the menu. I used my phone to record the output of hwinfo (multiple times) while the game was running, so I could see exactly what the watts/etc were in the last frame before the PC shut down. Nothing looks out of the ordinary. Watt readings for CPU and GPU are fine. CPU, GPU, RAM, and chipset temps all fine. GPU temp in the low 70s, everything else in the 50s.

On paper all the math says a 750 should be able to handle my total system wattage. I have the PC plugged into an APC UPS with an LCD screen, and I watched the total wattage load while playing a game, and it never went higher than 600w. Seasonic Prime supplies are very high quality and it never gave me any issues all year long with a 1080 TI FTW3. I am not running any overclocks on CPU or GPU, just the XMP profile for my RAM, which has been perfectly stable all year long.

I'm not happy about it, but I ordered a new Seasonic PRIME TX-1000 SSR-1000TR to see if more watts fixes the problem. If that doesn't fix it, then I guess the problem could be the card itself, but I'll cross that bridge when I come to it. The new PSU should be here Wednesday of this week, and I'll reply here with the results.

I'm having exact same issue. Thermalake 750W with an EVGA XC3 and an I7-9700K. Gaming at 4k will cause my computer to power down and cycle back on. 1440p i havent had an issue yet but we will see. Have a 1000W superflower psu coming (I know risky taking a third party brand but we will see)

 

[johnsoner13]

48 GB Corsair Vengeance RGB Pro @ 3200 MHz

Weird amount. Try with just one matching kit of ram installed, problems arise when you mix n’ match

 

[brandonc0431]

I have this exact same issue with an EVGA 3080 XC3. I have a Seasonic PRIME PX-750 SSR-750PD2 that I bought in January and your description of events matches mine perfectly.

Never shuts down when doing normal tasks. Never shuts down during benchmarks. Sometimes a game will run for quite a while with no issues, sometimes it will shut down in the menu. I used my phone to record the output of hwinfo (multiple times) while the game was running, so I could see exactly what the watts/etc were in the last frame before the PC shut down. Nothing looks out of the ordinary. Watt readings for CPU and GPU are fine. CPU, GPU, RAM, and chipset temps all fine. GPU temp in the low 70s, everything else in the 50s.

On paper all the math says a 750 should be able to handle my total system wattage. I have the PC plugged into an APC UPS with an LCD screen, and I watched the total wattage load while playing a game, and it never went higher than 600w. Seasonic Prime supplies are very high quality and it never gave me any issues all year long with a 1080 TI FTW3. I am not running any overclocks on CPU or GPU, just the XMP profile for my RAM, which has been perfectly stable all year long.

I'm not happy about it, but I ordered a new Seasonic PRIME TX-1000 SSR-1000TR to see if more watts fixes the problem. If that doesn't fix it, then I guess the problem could be the card itself, but I'll cross that bridge when I come to it. The new PSU should be here Wednesday of this week, and I'll reply here with the results.

Sorry to see that you're facing the same issue as me, but it's comforting to know I'm not the only one. It's also quite helpful to know that your having this issue on a different AIB. This strengthens the case for the PSU being the issue, assuming it's not a coincidence. I unfortunately can't measure the power draw of my system because I don't have a UPS or tool to measure it, at least for the GPU that is. But, I'm quite certain I'm not pulling more than 700 Watts. I would love to hear back from you when you get your new PSU if the issue has been resolved. I know it will be difficult to test if your shutdowns are anything like mine, being completely random without repeatability. If the PSU replacement solves the issue, I will gladly purchase a new 850+ Watt PSU.

 

[rcald2000]

brandonc0431 Your Corsair RM750x is enough power for your new RTX 3080. Reboots are normally caused by one of three issues: (1) power, (2) heat, or (3) software. I assume that your system was stable before the new PSU, so that likely eliminates (3) software. As for heat, assuming that your CPU temp under gaming loads, is accurate, it shouldn't be the CPU thermal throttling to the point of reboot. If you want to be certain that it's not heat related, then while the computer is on, open the side of your case. Then turn the computer on it's side, so that the motherboard is facing the ceiling. This will allow excess heat to escape. Then see if the reboot problem is repeatable.

What I think the most likely cause is: I think your power supply is malfunctioning. Not because of excessive power draw, but simply malfunctioning. I once had a EVGA P2 850 reboot intermittently about ten minutes after starting a Fortnite update. The only components in the system were a Ryzen 5 2600 and GTX 1660 Ti. I'm aware of the current issue with RTX 3080 cards. But it's my understanding that those situations only involve games crashing to desktop, not rebooting the entire system. If it were my system, and if I believed the power supply were the likely culprit, I would then contact Corsair and seek a "advanced RMA replacement". This means that they send you a replacement power supply before you send the damaged one back to them. They'll need to request your credit card number. If you don't send your damaged PSU back to them within the set time period, then they charge you for the replacement power supply:

Advanced Replacement
If you want to speed up the time it takes to get your replacement, you can order an advanced replacement. If you provide your credit card details to Corsair, Corsair will ship the replacement to your designated location using express courier service at Corsair’s expense.

When you order an advanced replacement, Corsair will place a charge on your credit card equal to the current cost of the replacement product sent to you. To remove this charge from your credit card, return the original product to Corsair within 14 days of receiving your replacement. Make sure you follow Corsair’s instructions for returning the original product and:

• Send back the original product in the package containing the replacement
• Affix the prepaid return shipping label on the outside of the package
• Clearly write or display your RMA number on the outside of the package and include a copy of your RMA confirmation email inside your package.

If Corsair does not receive the original product within 14 days, the charge on your credit card will remain.

If Corsair receives the original product within 14 days, the charge on your credit card will be removed if the product is eligible for warranty service. If the product is ineligible for warranty service, the charge on your credit card will remain. Products ineligible for warranty service include products rendered unusable by unauthorized modifications or catastrophic damage.

** Addendum **
Your Corsair RMx power supply has a ten year warranty. I strongly suggest that you take advantage of it. Good luck and please keep us in the loop, as it would be helpful for future issues.

 

[Juular]

It's not a PSU issue. If it were PSU, i.e if it tripped it's over-current protection then it wouldn't have let you to power PC on until you flick the switch on it. That's probably related to an issue with power delivery on some AIB cards.

 

[DSzymborski]

I'm having exact same issue. Thermalake 750W with an EVGA XC3 and an I7-9700K. Gaming at 4k will cause my computer to power down and cycle back on. 1440p i havent had an issue yet but we will see. Have a 1000W superflower psu coming (I know risky taking a third party brand but we will see)

Super Flower is not a third-party brand. And there's no Thermaltake-branded PSU in existence I'd put clearly above a higher-end Super Flower.

 

[Juular]

And there's no Thermaltake-branded PSU in existence I'd put clearly above a higher-end Super Flower.

Well, Thermaltake's CWT CST and GPU based units are fairly good - TP GF1, PF1, iRGB, DPS G RGB, some of them are loud tho but otherwise their performance is pretty much on par with Leadex II/III.

 

[brandonc0431]

It's not a PSU issue. If it were PSU, i.e if it tripped it's over-current protection then it wouldn't have let you to power PC on until you flick the switch on it. That's probably related to an issue with power delivery on some AIB cards.

There is a setting in the BIOS that enables this feature, I'm not currently using it. I do also believe that the PSU is not being overstressed. I think it's just faulty and certain types of loads cause it to shutdown.

 

[Juular]

There is a setting in the BIOS that enables this feature, I'm not currently using it. I do also believe that the PSU is not being overstressed. I think it's just faulty and certain types of loads cause it to shutdown.

There's no BIOS setting to disable OCP or OPP. A faulty PSU is possible but you're definitely not getting these restarts because PSU can't output enough power, otherwise it would've been latching off.

 

[joshuacant]

Sorry to see that you're facing the same issue as me, but it's comforting to know I'm not the only one.

I will definitely post a follow up after I get the new PSU installed.

I did find this thread, however, which is a bunch of people chasing the exact same symptoms (albeit with 2080ti cards) and all of them did what we're doing (new PSU) and all of them found that did not fix the issue, only a new GPU did. Give it a read if you have a moment: https://forums.evga.com/EVGA-2080-Ti-XC-Ultra-tripping-OCP-on-Seasonic-PSU-m2908567.aspx

I sure hope that's not the case here for a bunch of reasons, lol. I guess I'll know in a few days either way.

 

[brandonc0431]

There's no BIOS setting to disable OCP or OPP. A faulty PSU is possible but you're definitely not getting these restarts because PSU can't output enough power, otherwise it would've been latching off.

Yeah, my mistake, I didn't mean OCP for the BIOS setting, but there is a setting that prevents the computer from trying to power on again when there is a power fault, which I have disabled. I definitely don't think my PSU is shutting down from too much power draw, but hopefully it's the root of my problems as getting a replacement 3080 won't be easy right now...

 

[joshuacant]

@brandonc0431

My PSU was delayed by a day and didn't arrive until a few hours ago. I installed it and and have only done a little testing so far. Not enough for me to feel 100% certain but...

No overload shut downs so far, and I have run the card much harder (even a small OC) than I did previously. The shutdowns were random, but given what was happening before, I should have seen several of them by now. 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.

As a side note, earlier this week I did upgrade to the new driver that fixes the CTD issue. I still got the overload shut downs, so that driver had no impact in my case.

 

[brandonc0431]

@brandonc0431

My PSU was delayed by a day and didn't arrive until a few hours ago. I installed it and and have only done a little testing so far. Not enough for me to feel 100% certain but...

No overload shut downs so far, and I have run the card much harder (even a small OC) than I did previously. The shutdowns were random, but given what was happening before, I should have seen several of them by now. 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.

As a side note, earlier this week I did upgrade to the new driver that fixes the CTD issue. I still got the overload shut downs, so that driver had no impact in my case.

I'll be on the look out. As for me, my CTDs actually increased significantly after updating to the new driver unfortunately...