Return to

Troubleshooting PSU



I have a corsair RMi PSU that lights up the mobo’s RGB and has a working fan test button on it but the system does not power on. When pressing the power button on the mobo there is a barely audible click from the PSU but no boot or fan spin on either CPU, chassis or PSU.

I have tried connecting the system to a different PSU and it works great. I also removed all non-essential components like graphics card and even SSD and checked all cables with the believed faulty PSU. I also tried bridging the 28 pin cable with a paper clip without fan spin.

To me it’s a bit confusing that the PSU still provides power to the light the RGB so I figured I’d check if anyone here has any other ideas or if it’s a definite RMA.


Got a multimeter? If so, check the 12V, 5V, and 3.3V rails.

It makes 0 sense that the PSU fan doesn’t spin, but the other fans do, considering they should be both hooked up directly to the 12V rail…

It’s probably the output capacitors ¯\(ツ)/¯ Check those for bulging or leakage. If none, find out what rails works in the PSU, and what doesn’t.


Sniff test. If it smells odd, it is broken.


Well if the system works fine with a different psu,
then i would say for sure that the psu is dud.
Unless its a fully modular psu, then it could also be something with the 8 pin cpu power cable,
or the 24 pin powercable.


It’s dead. Replace it.

Well actually, I’d bet that the 12v rail is dead.
The led uses the 5v or the 3.3v. Therefore it comes on.
Case fans use 12v and 7v ( Electricity is a strange creature. When you add 12v and 5v together you get 7v. ). No case fan(s), 12v is most likely toast.

Another way that electricity is strange, you can actually see/have 12v show on a meter. But there may not be enough/any, watts/amps there to power anything (i.e.: 95W for the CPU ). So you can still “have” 12v and nothing works.

Like I said. It’s dead. Replace it.


Ok, so the fan that spins is the one on the PSU but only when pressing the fan-test button on it. No other fans are spinning with this PSU.

Thanks, should the voltage be measurable even when the PSU is not powered up?

No visible changes, at least not without opening it and I feel the warranty will be less of a hassle if I don’t remove the warranty void stickers.


no smoke or strange smells


Nope, none of those voltages will be there without the PSU powered on. You need to short the PS_ON pin to ground (green wire) for those 3 rails to come alive.

If you’ve still got warranty, then yeah, don’t open it, RMA it!


Yeah and also he has got a famous Corsair RM-i psu series.
I believe there have been some issues with those particular psu’s dying.
But i cannot fully remeber what the exact cause of the issue was.


I’m not sure what causes the RM-i PSU’s to fail, and couldn’t find any news on it. From my experience in fixing PSU’s in the past, a lot of the time it is the output caps, or more often, a glue that is used to keep components more secure onto the PCB, but goes conductive with heat (normally happens after years of usage).


2 things
fuze and relay
fuze failure will be no power output at all and no fan spin on test.
relay (also 2 things)
if relay is not getting supply power but is getting control voltage you will hear the relay click as it closes the switch but will not pass current if it is not present.
this can often happen with blown caps, burned mosfets, cooked resistors, and fried diodes.
damaged relay contacts
if supply current is present but getting no output from the relay the contacts are burned enough they prevent passing current even when the relay is switched on.

replace the psu anyway!

If you don’t have experience with repairing power supply systems
attempting to work on them is dangerous.


Yeah, I got a replacement. The store didn’t really diagnose it but shipped it back to corsair.


that is one issue people face when they use a digital multi-meter
most of your dmm’s are non loading and will display ghost voltages ( induced voltage with little or no current)
an analog meter provides some load and is evident when you apply the probes it will show a voltage but will quickly drop to zero.
(you dont know how many times Ive had to explain this to apprentices over the past 20 years)
usually when detecting a voltage you are unsure of using an analog meter or magnetic solenoid tester(wiggy) will prove out a ghost voltage quickly


good that it got sorted.

A few things that could have failed here, the modular 24pin cable, have a dead one too.
Or indeed something in the psu went bad.
The early RM series psus turned the fan on to late and some parts managed to get to hot.
Fan curve adjustment fixed that in the later ones.
Mine literally popped caps. it still worked, but yea.

The motherboard leds lighting up is because of the allways on, 5V StandBy rail that every psu provides.

By the way, the Psu isn’t necessarily dead.
There are a few phenomena where PSU and board start disliking each other.
One possible Symptom is it not turning on.
Both parts will, in that case, probably work flawless with other parts, just not in that configuration.


the power good signal that is provided by the motherboard back to the power supply is usually the culprit in many psu disliking a board
it may be borderline due to pin oxidation or connector fatigue
loose connections will generate heat and drastically change conductor resistance.
too much resistance and the power good signal may not be getting through to the psu control circuits. this causes immediate shutdown
before replacing a psu I carefully inspect both connectors and test the power supply with an external testor.
pin fatigue is evident by the female connector pins being spread out too wide.
(sometimes this can be corrected by gently bending them back in with a dental pick or large needle)


i thought that the Powergood Signal is coming from the PSU once 5V rail is stable.

but a good tipp on the connector side.


a psu does not have any processing capability so it cant determine whether the power is good or not.
the psu relies on a nested control relay set up to let power through to the motherboard. (generally using opto couplers or ssr (solid state relays)
if the signal voltage is not returning from the control relay on the motherboard the control relay in the power supply does not stay closed and immediately cuts power to the motherboard.
this configuration came out with the advent of the ATX boards.
the earlier at and xt psu did not have this setup and the switches were directly connected to the psu. the drawback of this was it supplied power full time when on and if something failed on the board it could cause cascading and catastrophic damage to the motherboard.

The plus side of it was you could use the psu externally to test and run optical drives, fans and so on.

(retired industrial electrician with many years experience in computer systems and power supply scenarios)