Question The 3.3V rail is below the minus threshold, can this effect an NVMe SSD when in gaming ?

[csm101]

Hi All

i have my 3.3 V rail reading as 2.968V. can this effect the read write operations of a NVMe SSD when it comes to gaming?

I also create content where the video is about 40GB for a 1 hour long video for YT. i use Davinci resolve and when this operation happens I don't get any issues. i only get this when playing a game that needs lot of reading from the nvme like wukong, indiana jones or SOTTR. i get the BSOD with that windows hardware uncorrectable error message. when check the logs its pointing to a PCI root hub 3. sometime the game crash and the ssd is visible in the explorer but as cannot open any folders as it says path is not valid. shutting down and booting up fix that problem until it happens again.

the 12v goes as 11.952v and 5v goes as 4.95v
the details are below

OS -win 10 latest
motherboard :- gigabyte B660M
cpu :- i5 12400
ram :- DDR 4 16gb*2
os is on one nvme (256gb) and gaming is done on another nvme (1tb)
psu:- corsair RM750x

i do have a HDD and its full of movies and i can watch then without an issue as well. my gut feeling is its time to replace the PSU . should i replace it?

 

[kanewolf]

Hi All

i have my 3.3 V rail reading as 2.968V. can this effect the read write operations of a NVME SSD when it comes to gaming? i also create content where the video is about 40GB for a 1 hour long video for YT. i use Davinci resolve and when this operation happens I don't get any issues. i only get this when playing a game that needs lot of reading from the nvme like wukong, indiana jones or SOTTR. i get the BSOD with that windows hardware uncorrectable error message. when check the logs its pointing to a PCI root hub 3. sometime the game crash and the ssd is visible in the explorer but as cannot open any folders as it says path is not valid. shutting down and booting up fix that problem until it happens again.

the 12v goes as 11.952v and 5v goes as 4.95v
the details are below

OS -win 10 latest
motherboard :- gigabyte B660M
cpu :- i5 12400
ram :- DDR 4 16gb*2
os is on one nvme (256gb) and gaming is done on another nvme (1tb)
psu:- corsair RM750x

i do have a HDD and its full of movies and i can watch then without an issue as well. my gut feeling is its time to replace the PSU . should i replace it?

You have to question how that voltage is being read. What is the accuracy of the sensor. If you said 2.9xxx with a recently calibrated 4 digit accuracy digital volt meter, then I would have more faith in that value.

 

[csm101]

its what bios reports and its also what the gigabyte provided s/w also reports. im planning to test psu using a multimeter. if your system have that value will you replace the PSU?

 

[Misgar]

Is this any help? You might need to push down firmly on the 24-way ATX PSU connector to improve connection with the motherboard.
https://www.techpowerup.com/forums/threads/low-3-3v-rail-resolved.229795/

 

[csm101]

so after much discussion with other people as well as AI ! 😀 i decided to replace the PSUand so far the system has not crashed. but the 3.3 rail value is still below . this means the censor is not reliable at all. i will test the old psu via multimeter and see the rails. the conclusion is that the psu is not stable under different loads. the games that i was playing was wukong and SOTTR. so obviously these two need different power loads. but the psu is stable enough to do rendering work. i will post my multimeter test results as well and i will do more tests on the system simply playing the games that i have. hopefully the issue is resolved.

 

[Misgar]

the 3.3 rail value is still below . this means the censor is not reliable at all.

Sensors should be used as a guide, not treated as 100% accurate.

i will test the old psu via multimeter and see the rails.

That's more likely to give you a more accurate reading, provided the meter has been returned from a Calibration Laboratory within the last 12 months (as per @kanewolf). With a $3 AliExpress multimeter of questionable quality and calibration, I wouldn't bet on the readings being spectacularly accurate.

the conclusion is that the psu is not stable under different loads.

I wouldn't draw that conclusion unless I had subjected the PSU to a wide range of loads. Unless you've monitored the voltage throughout the tests, it's just theory as to what's causing the problem.

The thread I posted yesterday pointed to a poor connection (high resistance) in the user's 24-way motherboard power connector.

You might measure 3.3 Volts on the top of PSU plug, but if the internal contacts are loose or dirty, you might see only 2.9 Volts on the motherboard. Nothing wrong with the PSU voltage regulation, just a poor electrical connection.

i use Davinci resolve and when this operation happens I don't get any issues

I'm not sure what GPU you're using but if it's NVidia, are you using the Studio Driver or the Gaming Driver?

I'm running Adobe, Topaz, DaVinci and I use the NVidia Studio Driver which is supposed to be more stable. AMD supply similar professional/gaming driver choices for their graphics cards.

Since your system seems stable in DaVinci, I suspect the problem in games is not associated with PSU "stability" but something else.

 

[kanewolf]

That's more likely to give you a more accurate reading, provided the meter has been returned from a Calibration Laboratory within the last 12 months (as per @kanewolf). With a $3 AliExpress multimeter of questionable quality and calibration, I wouldn't bet on the readings being spectacularly accurate.

AND a meter ONLY is a no-load voltage. That is not a valid condition.

 

[SteJBorchard]

Its probably been stated already but you can't trust built in sensors to the motherboard and Operating shystem. Gotta get yourself a decent digital Multimeter, learn how to use it and physically do it yourself.

And if any circuit in the power supply or anything AC/DC or otherwise is more than 5% out of spec it will definitely effect things. 5% is the standard for home and small business electronics. The varriance is smaller for expensive/propreitary devices or military grade.

Sorry how old is this power supply again?

 

[Misgar]

AND a meter ONLY is a no-load voltage. That is not a valid condition.

A meter can also be used to measure on-load voltages.

That's why I suggested performing a series of measurements at different loads to see if the OP's PSU is faulty.

All the DC voltages at the output of an ATX PSU are well below the SELV limit.
https://www.electricalsafetyfirst.o...is-the-difference-between-selv-pelv-and-felv/

For anyone interested, here is the spec for the maximum and minimum votages on an ATX v2.2 PSU.

Power Supply Voltage Tolerances (ATX v2.2)​

PSU Tolerance Table
Voltage Rail	Tolerance	Minimum Voltage	Maximum Voltage
+3.3VDC	± 5%	+3.135 VDC	+3.465 VDC
+5VDC	± 5%	+4.750 VDC	+5.250 VDC
+5VSB	± 5%	+4.750 VDC	+5.250 VDC
-5VDC (if used)	± 10%	-4.500 VDC	-5.500 VDC
+12VDC	± 5%	+11.400 VDC	+12.600 VDC
-12VDC	± 10%	-10.800 VDC	- 13.200 VDC