PC Crashed. PSU 5VSB LED is always off now. What caused?

[robotsalad]

Sorry to anyone who tries to help because i made this a horrible puzzle by buying a motherboard, gpu, and cpu used. The Power supply (Rosewill Quark 550) never had problems before but was running a xeon e3-1240 v3 and a gtx 1060 3gb. The "new" components were Asus prime z370-a, i5-8600k, and rx-580 8gb. I have not been able to test the gpu in another system yet, but i have tried another pcie vga adapter to determine that it at least is not the pcie slot that is broken. When i first started using the new mobo and cpu (before gpu arrived) there was a frequent reving of something in the system which i assumed was the PSU fan just working harder because of the higher TDP CPU (OC at 4.4GHz default voltage).

 

tldr; I want to assume the PSU is just inadequate and the motherboard did not cause the instability in the system. But maybe the GPU died and damaged the PSU. Or maybe the motherboard killed the GPU and damaged the PSU. I still have to test the GPU in another system.

 

on the plus side, my goal was to buy used parts so my first time overclocking didn't kill my new parts, so if my 4.4 OC is to blame then that part of the plan worked.

[Spotty]

What happens when you try to turn the computer on? Nothing?

[robotsalad]

Yeah i can power the pc back on after jumping the power supply. It will not boot with the graphics card plugged and it trips the PSU when I try. I haven't been able to test the GPU in another system yet. So yeah it runs, but the 5vsb light (that was always on even when pc was off) is never on anymore.

[Spotty]

2 minutes ago, robotsalad said:

Yeah i can power the pc back on after jumping the power supply. It will not boot with the graphics card plugged and it trips the PSU when I try. I haven't been able to test the GPU in another system yet. So yeah it runs, but the 5vsb light (that was always on even when pc was off) is never on anymore.

And if you use the 8600k integrated graphics? (Remove graphics card, Plug monitor in to motherboard).

[robotsalad]

Yeah that's what I'm doing now. I almost definitely shouldn't be because I know something is wrong with the PSU but I want to get answers and the parts are used anyways.

[J.Poppe]

 

1 minute ago, Spotty said:

And if you use the 8600k integrated graphics? (Remove graphics card, Plug monitor in to motherboard).

i would also clear cmos since you purchased it used, if you haven't already.

[mariushm]

PSU 5v sb off means that either the power supply is dead, or the motherboard doesn't initialize at all (bios completely damaged, chipset faulty etc)

 

The power supply is basically made out of two separate power supplies, the big one that powers all the components and a very small one that outputs only 5v.

The small one is basically like a usb phone charger, it outputs 5v all the time, even when your pc is turned off, and it's only designed to output a small amount of power, enough to keep the chipset on the motherboard powered and optionally power small devices like modem, ethernet card - the idea is that if the chipset it powered (and therefore bios code is running), it can monitor a modem or ethernet card or a keyboard and if a certain command is received through such devices, it can command the whole pc to turn on.

 

So if 5v stand by is not working, the chipset is unlikely to work so you can't turn on the pc because the bios can't send the "turn on the main power supply" signal to the power supply when you push the button on the case - the button is connected to the chipset on the motherboard, not to the power supply. 

 

The 5v standby power supply can often die or become bad in cheaper power supplies, because that's part of the power supply is always on, 24/7, and often the components in that section don't have adequate cooling (it's also not helping that psu fan doesn't move air when pc is turned off)

 

Some basic diagnostics you can do... check if the power supply outputs 5v on the 5v standby pin.

You need a multimeter. Disconnect the 24pin atx cable from the motherboard, and put the multimeter on the DC voltage measurement and put one probe on any GROUND wires (usually black) and the other probe on the 5v SB wire (usually purple) - you should see 5v or -5v or something close to these numbers (depending on what probe you put on ground)

If you don't have a multimeter, you could still check for some voltage there with let's say a 6v or 12v incandescent bulb for cars, bikes etc ... even though it's only 5v, the filament should still warm up or even produce light, so you'll know there's energy there.  Or if you have a led and a small resistor (like let's say 100 ... 1000 ohm), you could connect the resistor in series with the led and then connect the two between the 5v stand by and ground and the led should light up (leds care about polarity so you may have to flip the led around to make it light, basically if it doesn't light regardless of polarity you know the psu doesn't output voltage or led is dead)

 

You can also attempt to power the power supply by faking the signal sent by the chipset - you can do that by connecting a wire between the PS-ON contact (green) and any ground contact (black wire , black in the picture below)

When a wire creates connection between PS-ON and Ground, the power supply starts completely and sends 3.3v , 5v and 12v to the other wires in the connectors. So you could use a multimeter again to check the other voltages. Just keep a probe in one of the ground contacts and use the other one on orange (3.3v) , red (5v) and yellow (12v). -12v (blue) may or may not be present, and it wouldn't be a problem if you don't actually measure -12v, could be as low as -8.. -10v and it would still be fine ... it's practically no longer used in computers.

 

 

[kingknightrider]

I personally prefer to use the formula:

(3/5)*LISTED PSU WATTAGE=Total Probable Constant Output of a degraded PSU.

So in this case: (3/5)*550=330W probable constant output of a degraded PSU.

So 330W Probable Constant output of the PSU.

Note: The formula is just a approximation of constant output -it can deviate above or below based on the quality of unit of PSU.

I choose to err on the side of caution when messing with power supplies.

 

When dealing with finding out if a power supply will work long term in a system always go for the constant output.

A PSU can't push its max 550W indefinitely -especially if it is one of half and half PSU's.

 

With the total load being 400W by the CM Calculator website but recommending 450 W PSU -though some information to truly calculate it is missing.

 

IF the PSU has aged and can't push the 330W sustainable or is unreliable unit it is much more need a stronger unit.

[Spotty]

2 hours ago, kingknightrider said:

I personally prefer to use the formula:

(4/5)*LISTED PSU WATTAGE=Total Sustainable Output

So in this case: (3/5)*550=330W

What? 

[kingknightrider]

4 minutes ago, Spotty said:

What? 

Sorry my post got terminated early.

PSU being used is a 550W.

Thank you for the notification.

Also maybe should have used probable than just total constant.

[Spotty]

4 minutes ago, kingknightrider said:

Sorry my post got terminated early.

PSU being used is a 550W.

Thank you for the notification.

Also maybe should have used probable than just total constant.

 

2 hours ago, kingknightrider said:

I personally prefer to use the formula:

(4/5)*LISTED PSU WATTAGE=Total Probable Constant Output

So in this case: (3/5)*550=330W probable constant output

So 330W Probable Constant output of the PSU.

Note: The formula is just a approximation of constant output -it can deviate above or below based on the quality of unit of PSU.

I choose to err on the side of caution when messing with power supplies.

 

When dealing with finding out if a power supply will work long term in a system always go for the constant output.

A PSU can't push its max 550W indefinitely -especially if it is one of half and half PSU's.

 

With the total load being 400W by the CM Calculator website but recommending 450 W PSU -though some information to truly calculate it is missing.

 

IF the PSU has aged and can't push the 330W sustainable or is unreliable unit it is much more need a stronger unit.

I have no idea what language that is being translated from, but in English it makes absolutely no sense.

A 550W power supply can output 550W sustained load. That's why it's why it's labelled as a 550W power supply.

[kingknightrider]

48 minutes ago, Spotty said:

 

I have no idea what language that is being translated from, but in English it makes absolutely no sense.

A 550W power supply can output 550W sustained load. That's why it's why it's labelled as a 550W power supply.

Sigh.

A label can say anything that does not mean there is truth in the label.

Is this 600W or 250W?

Spoiler

300W or 350W I believe but I don't trust that label.

I refer to the efficiency of the PSU over time -years and not just the sustained output at this day or moment.

No PSU even a brand new one will be able to put out its labeled output for ten years exactly dead on every moment of every second of every day without some degradation.

 

I would find it doubtful using a 550W PSU in a system that needs 550W will work for nearly as long as a 750W in the same system without some issues of some kind eventually being known. Trouble in the system would occur with a 550W long before the same system with a 750W would cause any issues. This taking into effect that both PSU's would be of the same caliber construction.

 

https://www.silverstonetek.com/techtalk_cont.php?area=en&tid=wh10_005

 

Especially if the unit is in a damaged state and has degraded over time and only puts out 4/5 or even 3/5 its max labeled load.

Then again a new 550W -that is actually really good could probably even put out 560W maybe even 570W for a while.

[Spotty]

Just now, kingknightrider said:

Is this 600W or 250W?

That's a fire hazard.
But to answer your question, the model name is "600W", but the max rated output is 220W which is shown on the label. You can call the model number "10000000W" and that doesn't mean anything other than what the model name is - it technically has no bearing on the rated wattage of the unit. You need to look at the max output that's listed.

Any decent quality PSU from a reputable manufacturer will be able to output at its rated wattage. Obviously this doesn't include Chinese garbage that lie on the label.
 

3 hours ago, kingknightrider said:

I personally prefer to use the formula:

(4/5)*LISTED PSU WATTAGE=Total Probable Constant Output

So in this case: (3/5)*550=330W probable constant output

So 330W Probable Constant output of the PSU.

Your calculation of "You have to multiply the labels wattage by 3/5ths to get the true wattage" bullshit is just that; bullshit.


PS. What's a "Bule" cable? @jonnyGURU do you know what a "Bule" cable is?

 

Anyway... @robotsalad looks like your PSU is dead. Contact Rosewill or the merchant you bought it from for a return under warranty. Your unit has a 5 year manufacturers warranty so it will still be covered.

[kingknightrider]

39 minutes ago, Spotty said:

That's a fire hazard.
But to answer your question, the model name is "600W", but the max rated output is 220W which is shown on the label. You can call the model number "10000000W" and that doesn't mean anything other than what the model name is - it technically has no bearing on the rated wattage of the unit. You need to look at the max output that's listed.

Any decent quality PSU from a reputable manufacturer will be able to output at its rated wattage. Obviously this doesn't include Chinese garbage that lie on the label.
 

Your calculation of "You have to multiply the labels wattage by 3/5ths to get the true wattage" bullshit is just that; bullshit.

 

Old Blue reminds me of something else but I digress...

Yes it is a fire hazard and yes you need to look at the rating on the box and indeed having a quality PSU is a must.

Though not everyone is going to be as reliable and do any of that and more likely than not just come to some forum and ask for help.

 

However IF you end up working with customers who do not do these things then you take precautions -including developing methods and formulas that err on the side of caution when dealing with people who don't.

 

Also I didn't say to get its "true wattage" I said and that it is an APPROXIMATION of the Total PROBABLE Constant Output that is meant for a PSU that is in a degraded state anyway. Though your right the formula is misleading -though if you don't understand it -that is fine -it is a business formula I should have probably have left off -one that has paid dividends more times than not.

[mariushm]

33 minutes ago, kingknightrider said:

I refer to the efficiency of the PSU over time -years and not just the sustained output at this day or moment.

No PSU even a brand new one will be able to put out its labeled output for ten years exactly dead on every moment of every second of every day without some degradation.

 

I would find it doubtful using a 550W PSU in a system that needs 550W will work for nearly as long as a 750W in the same system without some issues of some kind eventually being known. Trouble in the system would occur with a 550W long before the same system with a 750W would cause any issues. This taking into effect that both PSU's would be of the same caliber construction.

 

 

That's kind of false.

 

A power supply correctly rated for 550w will output 550w even after 10 years or 20 years.

In fact, a quality power supply is often overrated, and there's at least 5-10% of extra capability, basically a good 550w power supply should be able to output 580w-650w for short periods of time if needed, before the protections kick in and shut down the power supply gracefully.

 

Degradation in capacitors will typically result in a more unstable voltage, higher ripple, more fluctuations in the output voltage. Capacitors will heat more due to increased ESR inside them, due to higher current ripple, which in turn can heat other components inside to a higher degree.

Also with age, heatsinks inside can accumulate dust, fans can lose lubrication and may spin slower if the power supply uses a fixed rpm fan instead of using a fan controller or some basic adjustment of fan speed with the internal heat generated.

So with age, components inside may heat more, but thermal sensors normally would catch that and cause the fan to spin faster, which means power supply may become noisier but it would still produce the amount of power it's rated for.

You may not be able to overclock components as much (because the fluctuations in voltage may cause the VRMs on components to not work as well, so with noisier power components won't reach as high speeds/clocks) but raw power wise, a power supply will still produce the advertised values even after 10 years.

 

There are some observations though

 

Some power supplies these days are only rated for 40c or 30c - see for example some Bronze efficiency EVGA units, or Corsair VS for example ... such power supplies are often derated by 20-25% when the ambient temperature inside the case or around the power supply gets close to those values. Basically, the power supply is not designed to cool itself sufficiently to handle advertised values if temperature is too much.

 

So for example, an EVGA 500w BR series is only capable of 500w  at 30c ambient temperature. If the temperature inside or outside gets close to 30c, derate it by around 20%... so you actually have a 450w power supply, or something like that.

 

Second, higher end power supplies convert  ALL (or pretty much all) the power to 12v, which is used by the power hungry components (cpu, video card, motors on mechanical drives) and then use very efficient dc-dc converters to produce 3.3v and 5v as needed, from the 12v output. Therefore, if your computer needs only 20 watts on 3.3v and 5v, the power supply will take around 22w from 12v and will produce 20 watts worth of energy on 3.3v and 5v, losing around 2w due to conversion inefficiencies.

On cheaper power supplies, there's an amount of power, usually 100-120 watts, always reserved for 3.3v and 5v - no matter how much power your computer needs that reserve is fixed.

In such power supplies it's not uncommon for the maximum output on 12v to be 50-100w less than the total output power.

 

[robotsalad]

8 hours ago, mariushm said:

PSU 5v sb off means that either the power supply is dead, or the motherboard doesn't initialize at all (bios completely damaged, chipset faulty etc)

 

The power supply is basically made out of two separate power supplies, the big one that powers all the components and a very small one that outputs only 5v.

The small one is basically like a usb phone charger, it outputs 5v all the time, even when your pc is turned off, and it's only designed to output a small amount of power, enough to keep the chipset on the motherboard powered and optionally power small devices like modem, ethernet card - the idea is that if the chipset it powered (and therefore bios code is running), it can monitor a modem or ethernet card or a keyboard and if a certain command is received through such devices, it can command the whole pc to turn on.

 

So if 5v stand by is not working, the chipset is unlikely to work so you can't turn on the pc because the bios can't send the "turn on the main power supply" signal to the power supply when you push the button on the case - the button is connected to the chipset on the motherboard, not to the power supply. 

 

The 5v standby power supply can often die or become bad in cheaper power supplies, because that's part of the power supply is always on, 24/7, and often the components in that section don't have adequate cooling (it's also not helping that psu fan doesn't move air when pc is turned off)

 

Some basic diagnostics you can do... check if the power supply outputs 5v on the 5v standby pin.

You need a multimeter. Disconnect the 24pin atx cable from the motherboard, and put the multimeter on the DC voltage measurement and put one probe on any GROUND wires (usually black) and the other probe on the 5v SB wire (usually purple) - you should see 5v or -5v or something close to these numbers (depending on what probe you put on ground)

If you don't have a multimeter, you could still check for some voltage there with let's say a 6v or 12v incandescent bulb for cars, bikes etc ... even though it's only 5v, the filament should still warm up or even produce light, so you'll know there's energy there.  Or if you have a led and a small resistor (like let's say 100 ... 1000 ohm), you could connect the resistor in series with the led and then connect the two between the 5v stand by and ground and the led should light up (leds care about polarity so you may have to flip the led around to make it light, basically if it doesn't light regardless of polarity you know the psu doesn't output voltage or led is dead)

 

You can also attempt to power the power supply by faking the signal sent by the chipset - you can do that by connecting a wire between the PS-ON contact (green) and any ground contact (black wire , black in the picture below)

When a wire creates connection between PS-ON and Ground, the power supply starts completely and sends 3.3v , 5v and 12v to the other wires in the connectors. So you could use a multimeter again to check the other voltages. Just keep a probe in one of the ground contacts and use the other one on orange (3.3v) , red (5v) and yellow (12v). -12v (blue) may or may not be present, and it wouldn't be a problem if you don't actually measure -12v, could be as low as -8.. -10v and it would still be fine ... it's practically no longer used in computers.

I'm not sure if you saw that part but i have jumped the power supply and the system runs normal without the graphics card. When I plug the graphics card in it will not boot. And the 5VSB light is never on but when the computer is all the way shut down the RGB on the motherboard still glows just like it did before this incident. I am still going to RMA the PSU but i'm just wondering what caused the instability.

 

2 hours ago, kingknightrider said:

Also I didn't say to get its "true wattage" I said and that it is an APPROXIMATION of the Total PROBABLE Constant Output that is meant for a PSU that is in a degraded state anyway.

okay so if you mean that because it seems to have taken damage, i should calculate its future stability that way then i understand and thanks. not a problem though because it is still under warranty.

 

4 hours ago, Spotty said:

 

I have no idea what language that is being translated from, but in English it makes absolutely no sense.

A 550W power supply can output 550W sustained load. That's why it's why it's labelled as a 550W power supply.

yeah i thought i was totally safe going into this with 550W