Need help replacing my potential bomb of a PSU

[Zaakuro]

My new graphics card produces a lot of coil whine, and I suspect my PSU might be the cause. I checked the voltages with HWiNFO, and it shows a reading of ~9/10V on the +12V rail. From what I understand, that’s bad and could potentially damage my components over time. My system consists of a 5700X3D, 32 GB of RAM, and a 7800 XT. The PSU I’m currently using is a Tacens Radix VII AG 700W, which I’ve read has a bad reputation online.

That said, mine has worked fine for 4 years, even surviving several outages. However, with the additional power demand from the 7800 XT, I’m worried it’s not up to the task anymore. I’m looking for a budget-friendly but reliable PSU to replace it. After some research, I found two options that seem popular and affordable: the Corsair CX 750 and the MSI MAG A750BN.

I mainly play Microsoft Flight Simulator, and my sessions can last up to 12 hours, keeping the GPU at nearly 100% usage the entire time. This puts my PC under a lot of stress, so I need something dependable for this kind of workload.

Here are the ones I’m considering:

• Corsair CX 750
• MSI MAG A750BN

Do these seem like good choices for my setup? Would they be able to handle my usage scenario?

 

[Harlow]

it shows a reading of ~9/10V on the +12V rail.

If it truly is 9V/10V, switch it off and don't touch it again unti;l you've got a new PSU.

I wouldn't rate any CX particularly highly. Aim for an RM750 or similar. The only time I use CX is in very low power old builds.

The MAG750BN looks marginally better, but I'd still be inclined to look for a modular PSU.

Just my 2 centswporth.

 

[Shrek]

Do you have a voltmeter to confirm the voltage?

The Tacens Radix VII AG 700M has a 10 year warrantee.

 

[Bill_Bright]

~9/10V on the +12V rail.

The allowed tolerances for the +12VDC rail for ATX 3.0 and before PSUs are ±5%. So the allowed voltages are +11.40VDC to +12.60VDC. For ATX 3.1 the allowed low is -7% but that still allows for nothing lower than +11.16VDC.

I agree with Harlow. If your PSU is only outputting ~10V or less, stop using that PSU. If nothing else, it is forcing the regulator circuits and devices on your motherboard and attached devices to work considerably harder to compensate. That excess strain creates excess heat and may result in accelerated aging and even permanent damage.

Do you have a voltmeter to confirm the voltage?

Shrek is right. You need to verify this voltage. HWiNFO is a decent HW monitor, but it is simply reading a sensor. And those sensors are VERY low-tech, inexpensive, and often very inaccurate. If you don't have a meter, then at least try a different HW monitor.

 

[Bobaganoosh]

Go into your BIOS and see if there's a hardware monitor tab with the +12v reading. Some motherboards (like biostar for example) don't have sensors that read correctly in software.

As for budget replacements, I'd consider the MSI MAG A850GL PCIE5 or Be quiet! Pure Power 12M 750W. There better supplies a bit more expensive, but those the cheapest ones I'd look at.

 

[9087125]

You have evidence of coil while which by the sounds wasn't normal? and voltages that are low, so this would warrant measuring the voltage. But you have to be carful who you listen to on the net, telling one to measure the voltage could easily give the wrong idea. Dont get the idea that you just measure the output/open-circuit voltage as that tells you nothing. A battery can have a full healthy open-circuit voltage and still be completely dead, like wise any open circuit voltage in a psu arises fundamentally from charge separation (the electric field across the two poles) caused by the emf and IS NOT what pushes or imparts energy to the electrons. So you must test the unit with an inline volt meter of some sort while under load. Listen to physics.

Not sure what you mean by surviving a few outages? My PSUs in my systems have survived literally thousands of abrupt outages they don't care about it, same as my many mains battery chargers. My 1200W charger has been shut off hundreds of times abruptly while at full load and still works fine. My new AM5 system has already been exposed to dozens of outages. They must be made to handle that as outages are quite common.

 

[DirtyDingusMcgee]

If you are playing 12 hour sessions, I'd get as nice a power supply as you can possibly afford.

 

[dirtyferret]

Do you have a voltmeter to confirm the voltage?

this, a lot of those apps are inaccurate. I have a back up corsair CX that didn't displayed accurate 12v on two separate mobos although the voltmeter said otherwise (as did performance)

 

[eidairaman1]

EVGA, Seasonic, Superflower, Corsair

 

[R-T-B]

If it truly is 9V/10V

The computer would most likely not start. Its not the best psu, but this strikes me as a bad sensor reading.

 

[dirtyferret]

Its not the best psu

it's actually a decent resonant LLC platform (unlike the modular version) and better performing than people give it credit for considering the cost. It's obviously not a RMx, Focus plus, core reactor II, etc., but those people are not spending the money to get those if they are looking at the CX series.

Pros​

• +
  Affordable
• +
  Full power at 41 degrees Celsius
• +
  High overall performance
• +
  Tight load regulation
• +
  Efficient
• +
  Highly efficient 5VSB rail
• +
  Low leakage current
• +
  Low vampire power
• +
  Long cables

Cons​

• -
  Very low hold-up time
• -
  Transient response at +12V could be better
• -
  Pretty high inrush current with 230V
• -
  Non-modular cables
• -
  A second EPS or an ATX12V would be ideal
• -
  Small distance between peripheral connectors
• -
  The fan speed profile could be a bit less aggressive

Corsair CX650 Power Supply Review

The Corsair CX650 power supply is affordable yet full capable of supporting a mid-range system.

www.tomshardware.com

 

[Bill_Bright]

But you have to be carful who you listen to on the net

Right. And I'm a certified master electronics tech and sorry, 9087125 but the rest of your post is pretty much gibberish. Who did you expect to understand that?

And sorry again, but it is not even entirely correct!

You said,

So you must test the unit with an inline volt meter of some sort

Huh? "Inline" volt meter of some sort? I've never even seen an "inline" voltmeter for computers, except, kinda, for USB ports and USB does not even carry +12V. They only carry +5V. So how does that help the OP? Not to mention, he has not even come back to tell us if he has a meter or not.

while under load.

Now that is correct. To measure accurately, the power source needs to have a realistic load placed on it. And what is the easiest and safest way to measure a PSU's output voltages with a multimeter? By testing through an unused power lead from the PSU. This way you don't even have to worry about sticking highly conductive probes on the into the heart of the motherboard, and computer is still placing a very realistic load on the supply.

Yes, there are PSU testers, but they are hardly conclusive because they only have a tiny 10Ω dummy load. They are best suited to verify missing voltages, not voltage accuracy. Not to mention, they cannot test for ripple and other anomalies - but that's for a different discussion.

But again, we don't even know if the OP has access to a meter so yes, he needs to be careful who he listens to.

@Zaakuro - I don't understand what you mean by "surviving a few outages" either. Outages don't hurt electronics. A sudden outage without properly shutting down a computer may result in corrupt files/data stored on a drive, but it will not damage the drive, PSU or other electronics.

The manner in which power is restored, however, can be problematic. Sadly, when unexpected power outages occur - from bad weather, for example - very often it is not restored "cleanly". It may "flicker" on and off, or come with sags (opposite of surges), dips (opposite of spikes) as well as excessive surges and spikes as 100s of air conditioners, refrigerators and other big appliances and 1000s of lights in your segment of the power grid all try to start up at once. It may take a few seconds for the grid to become stable. These anomalies take a toll on power supplies and just another reason to have your sensitive and expensive electronics supported by a "good" UPS with AVR (automatic voltage regulation). But that's also for a different discussion.

 

[Bobaganoosh]

We just went through this on another post recently. The motherboard sensors are what is being reported in hwinfo and it's not terribly uncommon that the sensors are wrong. Usually the motherboard does a better job in the BIOS. Go in there and check.

 

[Bill_Bright]

Usually the motherboard does a better job in the BIOS

I agree but the motherboard (chipset) is reading those same, very low-tech, very inexpensive sensors. The sensors simply convert the voltage (same with temps) into a hexadecimal value and the BIOS (or HW monitoring program) then converts it back into a regular decimal value we can understand. Should be easy but for some reason, some programs seem to be less accurate than others.

In any event, do not go by the reading of a single program, whether that be the BIOS or an installed program. Verify it, at least once, with a multimeter if something looks suspicious or off.

BTW - if the voltages appear correct, it is my experience they are correct. But if they appear to be out of tolerance AND the computer appears to still be working fine, then most likely the sensors are bad.

And as a side note and speaking of temps - note programs take a "sample", usually every few seconds. And the timer starts when the program starts. Also, a CPU can go from cool to overheated in just a few clock cycles and back down almost as quickly. And at 3GHz, that's 3 billion clock cycles every second. So it is common for different HW monitoring programs to show different temps when the load varies. Voltages, on the other hand, should be pretty stable and consistent.