Computer powers up but only from no power input

[orangecat]

So I have a friend of mines PC at my house and hes asked me to take a look at it for him and see if I can figure out what the problem is.

 

Known problems:

 

1. The system will only power on if it hasn't been powered with AC mains.

2. if the system powers off it won't power back on until you remove AC mains.

3. If the computer is turned on after being turned off it will power on for half a second then power off and not come back on until AC mains is removed.

 

Things I checked:

 

1. CPU is known good.

2. GPU is known good.

3. Ram is known good.

4. Motherboard is known good.

5. Storage disks are known good.

6. Ran stress tests all night and they all passed. No power down during tests.

7. Cleaned PC.

8. Reseated all components.

9. Used another known good PSU and it works fine.

 

I assume it's a bad PSU but I've never seen a PSU fail in this way before. the system can power on and it works fine when it does but it's only once the system turns off that it can't come back on. I'm not sure if some kind of protection is being tripped or something when it shouldn't be or what. Most likely it's a PSU issue but I'd like to know for sure before I tell him to buy a new PSU.

 

Any ideas?

[Guest]

You're doing your friend a disservice by finding the problem and then not moving forward with helping him fix it just because you've never seen this kind of problem with a PSU before.

 

The problem could very well be IC related.  Like a false positive on a protection or maybe the power good signal timing is unstable and it doesn't hold the signal long enough for the motherboard to warm boot.

 

I've even seen perfectly good PSUs not work with particular make/model motherboards because of timings.   Yet people will still assume there's something else wrong (defective PSU, defective mobo) because "they've never seen this kind of thing before" or they didn't even understand that a PSU is more than a box that provides power.  Kind of frustrating.

[Justapersononthenet]

It's a laptop or a desktop? does he have a ups? If yes try removing it, also change the plug that you use maybe and if it's a laptop does it charge when off and plugged in?

[Spotty]

Out of curiosity, what model of PSU is it? 

 

13 minutes ago, Kilobytez95 said:

Used another known good PSU and it works fine.

If you've narrowed the problem down to the PSU and know that it works with other PSUs then I would go straight to the step of replacing the unit.

 

If the PSU is still under warranty then contact the manufacturer to have it replaced. If it's no longer under warranty then it was probably due to be replaced anyway.

[orangecat]

13 minutes ago, jonnyGURU said:

You do know that a PSU is more than a box of capacitors and a transformer that converts AC to DC.  

 

You're doing your friend a disservice by finding the problem and then not moving forward with helping him fix it just because you've never seen this kind of problem with a PSU before.

 

The problem could very well be IC related.  Like a false positive on a protection or maybe the power good signal timing is unstable and it doesn't hold the signal long enough for the motherboard to warm boot.

 

I've even seen perfectly good PSUs not work with particular make/model motherboards because of timings.   Yet people will still assume there's something else wrong because "they've never seen this kind of thing before" or they didn't even understand that a PSU is more than a box that provides power.  Kind of frustrating.

1. yes I know a PSU is a complicated device.

2. I'm not doing my friend a disservice by "finding the problem and not moving forward". He asked me to take a look at it for him and I merely asked on this forum for some advice so i can better diagnose the issue. i already know it's 99.9% likely a bad PSU. How I can "move forward" is beyond me as I'm not actually being tasked with a repair here but whatever.

3. yes I'm aware it's likely an IC issue hence why i asked if anyone can elaborate further on the issue and maybe shed some light into what exactly the issue may be. Surely there's someone out there way more knowledgeable than you or me.

4. What are you even talking about? How is this question frustrating to you? I'm pretty sure I know what the problem is I just asked for some extra insight. If you're so frustrated then go to a different post.

[orangecat]

5 minutes ago, Justapersononthenet said:

It's a laptop or a desktop? does he have a ups? If yes try removing it, also change the plug that you use maybe and if it's a laptop does it charge when off and plugged in?

It's a desktop. he has a Cooler Master V650 PSU.

[orangecat]

4 minutes ago, Spotty said:

Out of curiosity, what model of PSU is it? 

 

If you've narrowed the problem down to the PSU and know that it works with other PSUs then I would go straight to the step of replacing the unit.

 

If the PSU is still under warranty then contact the manufacturer to have it replaced. If it's no longer under warranty then it was probably due to be replaced anyway.

It's a Cooler Master V650 unit. It was supposed to be a rather high quality unit but I've had a V700 fail on me personally so I kinda question the quality of Cooler Master's PSU's. I don't think he has warranty anymore as the PSU is well over 5 years old at this point. Most likely I'll tell him to pick up a new PSU.

[Spotty]

7 minutes ago, Kilobytez95 said:

It's a Cooler Master V650 unit. It was supposed to be a rather high quality unit but I've had a V700 fail on me personally so I kinda question the quality of Cooler Master's PSU's. I don't think he has warranty anymore as the PSU is well over 5 years old at this point. Most likely I'll tell him to pick up a new PSU.

I believe those came with a 5 year warranty? I'm trying to think when the v650 first came out and I don't think it was much more than 5 years ago?

Might be worthwhile having your friend check for the receipt/order details to see when they bought it, just in case it might still be covered by warranty. 

[orangecat]

Just now, Spotty said:

I believe those came with a 5 year warranty? I'm trying to think when the v650 first came out and I don't think it was much more than 5 years ago?

Might be worthwhile having your friend check for the receipt/order details to see when they bought it, just in case it might still be covered by warranty. 

Yea I told him to check into that when he first told me he was having problems but I think he said it's out of warranty. I can double check.

[YoFavRussian]

I'm curious, if you power on the PSU by connecting the PS_ON# to a COM or Ground, Out of the system by the way. Does it power up when you do this? Remove the jumper and try again, does the PSU still power on?

[orangecat]

1 minute ago, YoFavRussian said:

I'm curious, if you power on the PSU by connecting the PS_ON# to a COM or Ground, Out of the system by the way. Does it power up when you do this? Remove the jumper and try again, does the PSU still power on?

I tried that to see if it was an issue with the PSU being loaded down and it worked but it's hit or miss. Sometimes it powers up and stays on and other times it powers off. I assume when it powers off there's some kind of a no load check or something but to my knowledge no ATX PSU does that.

[YoFavRussian]

2 minutes ago, Kilobytez95 said:

I tried that to see if it was an issue with the PSU being loaded down and it worked but it's hit or miss. Sometimes it powers up and stays on and other times it powers off. I assume when it powers off there's some kind of a no load check or something but to my knowledge no ATX PSU does that.

So the two signal wires are the PS_ON# and Power_OK as far as I know those are the only two signal wires on the PSU. However the Power_OK plays a part in checking the PSU during the POST process. And as far as I know, this works on any PSU because it's the signaling that says, "Hey, power on".

When troubleshooting a system that will crash at times under load, I will test the PSU without a load at first and record the results. Then I will continue and put the PSU in the system and find an open line and monitor it and record those results as well.

[orangecat]

Just now, YoFavRussian said:

So the two signal wires are the PS_ON# and Power_OK as far as I know those are the only two signal wires on the PSU. However the Power_OK plays a part in checking the PSU during the POST process. And as far as I know, this works on any PSU because it's the signaling that says, "Hey, power on".

When troubleshooting a system that will crash at times under load, I will test the PSU without a load at first and record the results. Then I will continue and put the PSU in the system and find an open line and monitor it and record those results as well.

From what I've been able to gather the power output is good. I checked PS_ON and Power_OK and they both seem to be working fine. Might be internal to the PSU's protection circuit.

[YoFavRussian]

2 minutes ago, Kilobytez95 said:

From what I've been able to gather the power output is good. I checked PS_ON and Power_OK and they both seem to be working fine. Might be internal to the PSU's protection circuit.

Protection or some sort of regulation, either way the PSU is no good and if possible should be returned or put under RMA.

[Guest]

3 hours ago, Kilobytez95 said:

1. yes I know a PSU is a complicated device.

2. I'm not doing my friend a disservice by "finding the problem and not moving forward". He asked me to take a look at it for him and I merely asked on this forum for some advice so i can better diagnose the issue. i already know it's 99.9% likely a bad PSU. How I can "move forward" is beyond me as I'm not actually being tasked with a repair here but whatever.

3. yes I'm aware it's likely an IC issue hence why i asked if anyone can elaborate further on the issue and maybe shed some light into what exactly the issue may be. Surely there's someone out there way more knowledgeable than you or me.

4. What are you even talking about? How is this question frustrating to you? I'm pretty sure I know what the problem is I just asked for some extra insight. If you're so frustrated then go to a different post.

Sorry I came across so snarky.  I only meant to be a little snarky. 

 

2. Of course you're not going to repair the PSU, but if you know the problem is the PSU, just replace the PSU. 

 

3.  Last time I played the "do you know who I am" card it didn't go well.  I could find out what's wrong with the PSU if you send it to me, but that might be a waste of both of our times.  I've got enough disassembled PSUs on my desk. And I doubt Cooler Master would do a failure analysis for you either. 

 

4. No no no.. I'm not frustrated at you.  I said it's frustrating when you have a good PSU and a good mobo and they don't work together.  The customer assumes that there's a "bad batch" (God I hate that term. I think of cookies baked wrong) of PSU's or motherboards) and have to deal with end users (if your a tech) or an SI (if you're a manufacturer or supplier) chasing the problem down, explaining it to the customer and then trying to rectify the problem. 

 

All I'm saying is that unless you know your way inside the PSU (not just a PSU, but that particular PSU) you're chasing ghosts trying to figure out why it doesn't work. 

[Guest]

3 hours ago, Kilobytez95 said:

I tried that to see if it was an issue with the PSU being loaded down and it worked but it's hit or miss. Sometimes it powers up and stays on and other times it powers off. I assume when it powers off there's some kind of a no load check or something but to my knowledge no ATX PSU does that.

If the PSU is group regulated, a crossload could cause the output voltages to be out of spec enough to trip OVP or UVP, but that PSU has DC to DC off the +12V for the +3.3V and +5V so you can't crossload it and it should run with no load. 

3 hours ago, Kilobytez95 said:

From what I've been able to gather the power output is good. I checked PS_ON and Power_OK and they both seem to be working fine. Might be internal to the PSU's protection circuit.

The power good signal works on a delay, so even if you see voltage, the board may not respond if the delay of the signal is too short or long.  Usually, the delay is controlled by a resistor between the power good lead and the supervisor ICs, and, of course resistors can fail, if not completely, enough for their value to change. 

[orangecat]

On 8/26/2019 at 4:32 PM, jonnyGURU said:

Sorry I came across so snarky.  I only meant to be a little snarky. 

 

2. Of course you're not going to repair the PSU, but if you know the problem is the PSU, just replace the PSU. 

 

3.  Last time I played the "do you know who I am" card it didn't go well.  I could find out what's wrong with the PSU if you send it to me, but that might be a waste of both of our times.  I've got enough disassembled PSUs on my desk. And I doubt Cooler Master would do a failure analysis for you either. 

 

4. No no no.. I'm not frustrated at you.  I said it's frustrating when you have a good PSU and a good mobo and they don't work together.  The customer assumes that there's a "bad batch" (God I hate that term. I think of cookies baked wrong) of PSU's or motherboards) and have to deal with end users (if your a tech) or an SI (if you're a manufacturer or supplier) chasing the problem down, explaining it to the customer and then trying to rectify the problem. 

 

All I'm saying is that unless you know your way inside the PSU (not just a PSU, but that particular PSU) you're chasing ghosts trying to figure out why it doesn't work. 

In the end I just RMA'd the PSU. I know there's something wrong with it but there is no obvious sign.

[Guest]

45 minutes ago, Kilobytez95 said:

In the end I just RMA'd the PSU. I know there's something wrong with it but there is no obvious sign.

Cool.  Glad to hear it's still under warranty.