PSU trouble

[Deleted member 211755]

Hello everyone at techpowerup forums!

I RMA'd my main PSU - TX850M -
because it started making wierd noises
so I had CX750M still around so I replaced it
until my TXM series is fixed.

After using it for a few days I noticed some strange things though.
My GPU fans rev up every time it works hard, like 3D gaming.
They go to 100 percent in a second and my GPU performance suffers,
but if I use lower details everything is normal.
And just yesterday my PC froze while not running anything demanding
I was just browsing.
I think that my old CX750M is the cause of this
because I never had these problems with my TX850M.

I just wanna see what you guys think because
what better place to come for info and advice then techpowerup.

best regards everyone!
spanjaman

edit:
This CX750M is the first model made in 2014 or even earlier. It was bought around that time and not used for a long time.
So I believe it is it's fault my system in not responding like it should because that's the only component I replaced in the last week
because my TXM series is at repairs.
The TX850M became noisy and it made strange sounds so it could be just the ventilator which is at fault
because I remember playing Witcher 3 on ultra with no problems with fans on GPU
nor stuttering like with this CX750M.

Temperatures are in check, CPU around 55 celsius and GPU around 75
but I don't know the temps of CX750M
so if it's overheating it could be the PSU.

 

[joemama]

System specs please.

Your GPU fan behavior doesn't seem like a PSU issue, PSUs wouldn't affect your GPU fan speed. Maybe your GPU heatsink loosened and needs a repaste.

 

[FreedomEclipse]

The CX lines of PSU arent great in general i dont know why. Ive had to replace a fair few of them in recent years. Unless your pc is for basic tasks, Id avoid CX power supplies. Your TX is infinitely better.

 

[Deleted member 211755]

Ryzen 5 3600 + Noctua NH-U12S
MSI GTX 1660 Gaming X
B450 MSI Tomahawk Max
2x8GB Trident Z Neo
Nvme A2000 Kingston (Windows 10 Edu)
A400 Kingston SSD PCIe
Samsung Evo 850 PCIe
CX750M (Gonna be replaced with RM650x)

Thanks for your replies.
@FreedomEclipse Yes this is a bad one... My problems started after I installed this PSU

@joemama I know, that was my suspicion for a while, but just after I installed CX750M to replace TX850M which is RMA'd
all this had started
I am ordering RM650x tomorrow and will update this post if my PC continues to behave this way after replacing the CX750M or if it fixes the situation

edit:
I truly believe this is all CX750M's fault because the second I installed it all this began happening
When I was running TX850M none of this was happening
The GPU would be at 80 celsius and it wouldn't change fan control I setup in MSI Afterburner
After installing CX750M I found out that the fans don't rev up if my GPU temp is below 75 celsius

 

[The red spirit]

The CX lines of PSU arent great in general i dont know why. Ive had to replace a fair few of them in recent years. Unless your pc is for basic tasks, Id avoid CX power supplies. Your TX is infinitely better.

They might be lower end, but under any circumstances they shouldn't be straight up malfunctional.

 

[eidairaman1]

They might be lower end, but under any circumstances they shouldn't be straight up malfunctional.

Unfortunately they are very low quality.

 

[Frick]

Do you have a specific model number for the CX? There are a few of them around.

 

[The red spirit]

Unfortunately they are very low quality.

I very highly doubt that. I still remember Corsair VS series, those were okay, if you don't mind their relatively poor efficiency. CX ones shouldn't be any worse than VS. And PSU in question should supply 750 watts without problems, meanwhile this whole computer shouldn't even consume 300 watts. I think that his unit is just defective somewhere or his hardware somehow failed exactly after PSU swap.

 

[Deleted member 211755]

It is the very first model of CX750M (the one with the green letters)
It's pretty old, but I'm replacing it with RM650x ASAP
I didn't game with it long. Because of the situation I never play anymore, but I need my PC to work with online courses
I just hope that it won't fail other components until I swap it with RMx
Fortunately, my PC isn't a power hog so I believe it won't be a problem and since it is on the bottom of my case,
after my Windows 10 kind of froze, I put my case 6 cm higher so it has more fresh air to work with
It is very hot here during the summer and this is a very, very old unit
It could be overheating

@The red spirit I get what you mean, but when you replace the PSU unit and things like ramping fans start happening while under load
and the Windows 10 freezes, you have no other option then to blame the PSU
Especially if nothing like this ever happened with TXM series during two years of use

Thank you guys for the replies, I appreciate it a lot
I will update you when I get the new RMx power supply
People say they are good so I want the 650W unit in case I ever decide to upgrade to something more power hungry then 1660

 

[The red spirit]

@The red spirit I get what you mean, but when you replace the PSU unit and things like ramping fans start happening while under load
and the Windows 10 freezes, you have no other option then to blame the PSU
Especially if nothing like this ever happened with TXM series during two years of use

You know, I remembered one thing. If it's really old it might not be "Haswell Ready" or some stuff like that. I may remember it incorrectly, but maybe ATX revision was updated and bunch of C states were added. There's a chance that it doesn't support that, but double check that I may be wrong. If it doesn't support that stuff, then you may need to disable C states, "ErP Ready" and enable "power loading" in BIOS.

 

[Deleted member 211755]

@The red spirit That just might help me! I know my C-states are already disabled, but I'm not sure about other two options
I will restart now into BIOS to search for them
Thanks so much, this could be it because it is almost a 10 year old unit, at least 8 years I'm sure of it
Will update

 

[Shrek]

Your GPU fan behavior doesn't seem like a PSU issue, PSUs wouldn't affect your GPU fan speed. Maybe your GPU heatsink loosened and needs a repaste.

The power supply is part of the cooling system; maybe the CX is not drawing air out like the TX

just a thought

 

[Deleted member 211755]

@The red spirit So I went into BIOS and found out that c-states were enabled because a while back I was trying out CTR 2.0 to undervolt my CPU
So I disabled PSS, C-states. spread spectrum, performance regulator, PBO, LN2, core performance boost, I set idle control to typical current and I disabled CPPC and CPPC preferred cores.
I noticed my LED on motherboard started like they did with TXM series PSU and my system booted faster.

Thank you The red spirit for remembering me to check my BIOS
I forgot all about my settings when I still used TXM series

I will try running Witcher 3 on ultra to see if the GPU fans still ramp up under high load
Will update
Thanks again

@The red spirit It still happens, but it don't matter. I will replace my PSU soon with RMx series and the problem will go away I'm sure
I know for a fact new BIOS settings will help me when my PC is under light load or in idle
I read somewhere how old PSUs can give not enough amps and stuff like ramped fans on GPU or CPU happen
One guy even replaced the PSU and the problem went away
Since it was just a few days after I replaced my PSU with this old hog and my problems started
it's only logical that a new PSU will fix the situation

 

[joemama]

The power supply is part of the cooling system; maybe the CX is not drawing air out like the TX

just a thought

But even hot air inside the case wouldn't make the GPU fan speed ramp up to 100% within 1 second when applying load.

Temperatures are in check, CPU around 55 celsius and GPU around 75
but I don't know the temps of CX750M
so if it's overheating it could be the PSU.

That should be the temperatures under load right? Could you check the GPU core frequency under load when it stutters?

 

[The red spirit]

@The red spirit It still happens, but it don't matter. I will replace my PSU soon with RMx series and the problem will go away I'm sure
I know for a fact new BIOS settings will help me when my PC is under light load or in idle
I read somewhere how old PSUs can give not enough amps and stuff like ramped fans on GPU or CPU happen
One guy even replaced the PSU and the problem went away
Since it was just a few days after I replaced my PSU with this old hog and my problems started
it's only logical that a new PSU will fix the situation

You could try monitoring 12V, 5V, 3.3V rail voltages with HwInfo64 when gaming and report lowest values.

 

[Chrispy_]

Reseat the GPU and reseat the 24-pin ATX connector, and any PCIe connectors going into the graphics card. If they're particularly stiff cables going to the GPU, plug those in first, and then insert the card into the slot so that you're not wiggling the card around in the slot as you plug the power connectors into the graphics card.

It could well be a failing PSU that exhibits just enough voltage drop under load to make the graphics card panic, but the behaviour you're describing sounds more like a bad connection to the PCIe slot and it doesn't cost you anything to reseat the graphics card.

 

[Deleted member 211755]

@The red spirit They are a bit high on idle +12 - 12.136 V, +3.3V - 3.344 V, +5V - 5.040 V
I will open up Witcher 3 again and report the lowest soon

@Chrispy_ I tried reseating the GPU but not the 24-pin connector. As you said nothing costs me to try again
Will report later

@joemama Yes, those are the temps under load. I shall report the frequency while it stutters
What I can say now is that the GPU just heads on to 100 percent load and fans a second after I up the details in the game
and I know I can handle the game just fine and it doesn't matter which game
If the detals are high enough to make the GPU work above 80 percent, the fans just ramp up and instead of being around 80-95 percent usage
it just goes to 100 usage and the game stutters
Going to report the frequency

Thanks guys!

update:
On high details with hairworks ON it immediately goes to 99 percent usage and fans ramp up
the frequency is okay under the environment (it's almost 40 celsius outside)
Soon as I disable hairworks the load is fine and the fans as well
so my guess would be that after a certain point the PSU just can't get enough
power to the GPU and it behaves like this.

I tried reseating everything and it didn't help
the rails while in game are as follows:
12.192 V min 12.288 max
5.040 min and max
3.328 min 3.344 max

 

[Shrek]

But even hot air inside the case wouldn't make the GPU fan speed ramp up to 100% within 1 second when applying load.

Good point, I missed that.

 

[Deleted member 211755]

Update:
They replaced my failing unit with a new one, but the problem persists.
GPU is still ramping up and my new TXM unit still vibrates sometimes.
I read somewhere how there's a big piece of plastic
dividing the components from the ventilator
and how probably this plastic is what's being hit by a ventilator in the PSU.

At this point, I may have no other choice then RMA the graphics card...
I still have the warranty and I'll probably send it today.
I tried the clean install of drivers to remove software trouble from the equation
and it didn't help.
In OCCT I saw hot spot temperature while under load is above 100 C
so it is definitely overheating.
I remember one of you guys already said how this could be the case.
Well, you're right.

Regards

 

[Deleted member 211755]

Update #2
Well, people from the repairs say my graphics card is not at fault
and now I got a new PSU so I don't know what's going on with my PC
Could be software related or it could be the motherboard, I don't know

Do any of you have any ideas?
In a couple of days I'll get my GPU back and I'll have to do something
They said the fans don't ramp up and the temps are at 70 Celsius
At my room ambient is near 40 Celsius so it could be overheating here
I have an open case, but I ordered a ventilator for outtake and I will install another for intake
so maybe that might help

 

[jaggerwild]

I knew it wasnt the psu just from what the OP stated, Pat's self on back.

 

[Deleted member 211755]

@jaggerwild
Good for you
But it did start at the same time when my PSU started failing

Do you have any ideas where the problem actually lies?

Update #3
The store repairs wanted to screw me over saying how GPU wasn't at fault
but the manufacturer MSI says it is.
After a little argument the store will send it to authorized representative
and I hope they will fix my problem.

 

[Deleted member 211755]

Update #4

The guys at repairs did an extensive testing and my GPU really didn't have any problems
So I got it back today and did 6 things, one or more of which helped me resolve the trouble...

1) Replaced the HDMI cable
2) Installed two Artic case fans (intake/outake)
3) Firmly seated the graphics card in it's place
4) Disconnected cables and held power button for 30 seconds
5) Updated OS
6) Uninstalled GPU drivers then used DDU to clean up then installed them with NVCInstall from techpowerup

Maybe some of these things helped or the fact the PC wasn't even powered on for 15 days
or combination of all, but my GPU fans don't ramp up anymore and the performance is not dropping anymore

Thanks again to a few of you who wanted to help and gave some advice

This matter is resolved now

Regards

 

[Chrispy_]

Glad to hear its fixed. My money is on 3 or 6. Poor device detection by the motherboard+OS or bad drivers commonly cause the symptoms you were experiencing.

Even if you don't know what the exact cause was you've got a set of documented steps that worked so you can repeat them if the problem returns.

 

[eidairaman1]

Update #4

The guys at repairs did an extensive testing and my GPU really didn't have any problems
So I got it back today and did 6 things, one or more of which helped me resolve the trouble...

1) Replaced the HDMI cable
2) Installed two Artic case fans (intake/outake)
3) Firmly seated the graphics card in it's place
4) Disconnected cables and held power button for 30 seconds
5) Updated OS
6) Uninstalled GPU drivers then used DDU to clean up then installed them with NVCInstall from techpowerup

Maybe some of these things helped or the fact the PC wasn't even powered on for 15 days
or combination of all, but my GPU fans don't ramp up anymore and the performance is not dropping anymore

Thanks again to a few of you who wanted to help and gave some advice

This matter is resolved now

Regards

Probably dirty pcie slot.