PSU Capacitor Exploded - Cause?

[danomicar]

I experienced a catastrophic failure with my PSU where it appears a capacitor literally exploded and blew apart into the PSU fan.

 

Obviously, this kills the crab. But I was wondering, what could cause a capacitor to explode like this? My immediate thought is that it was simply overloaded.

 

PSU is a Solytech (literally never heard of this brand before) 480W power supply supporting a 1070ti and i7 8700. Was also driving plenty of USB peripherals for my racing simulator, save for the servo motor which is plugged into the wall. 

 

Was I just asking too much of the poor thing? Or is this just a crappy PSU?

 

It has since been replaced with an EVGA 1000GT in preparation for a whole-system upgrade before the end of the year. 

[An0maly_76]

8 minutes ago, danomicar said:

I experienced a catastrophic failure with my PSU where it appears a capacitor literally exploded and blew apart into the PSU fan.

 

Was I just asking too much of the poor thing? Or is this just a crappy PSU?

 

It has since been replaced with an EVGA 1000GT in preparation for a whole-system upgrade before the end of the year.

I haven't heard of this brand either. Probably a crap PSU with cheapie components that was pushed beyond its limits. Newer, better-quality ones use 105C rated capacitors which are far better. More wattage is always better, but what really counts is the 80+ rating in efficiency and quality of the internal components.

Edited August 25, 2022 by An0maly_76
Revised, more info

[da na]

8 minutes ago, danomicar said:

I experienced a catastrophic failure with my PSU where it appears a capacitor literally exploded and blew apart into the PSU fan.

 

Obviously, this kills the crab. But I was wondering, what could cause a capacitor to explode like this? My immediate thought is that it was simply overloaded.

 

PSU is a Solytech (literally never heard of this brand before) 480W power supply supporting a 1070ti and i7 8700. Was also driving plenty of USB peripherals for my racing simulator, save for the servo motor which is plugged into the wall. 

 

Was I just asking too much of the poor thing? Or is this just a crappy PSU?

 

It has since been replaced with an EVGA 1000GT in preparation for a whole-system upgrade before the end of the year. 

That's def a cheap no name PSU that got pushed a little too hard. Lucky none of your components died...

[An0maly_76]

2 minutes ago, Mel0n. said:

That's def a cheap no name PSU that got pushed a little too hard. Lucky none of your components died...

Yes, very lucky. I had an old eBay special Dell Dimension 3000 with a 3.0Ghz P4 that went out in a blaze of glory in similar fashion. And unfortunately, it nuked the entire machine. Even one of my hard drives was playing "Knock, knock. Who's there?" with its read / write heads. Of course, that was the drive containing all my irreplaceable music demo data backups, but I digress.

[da na]

Just now, An0maly_76 said:

Yes, very lucky. I had an old eBay special Dell Dimension 3000 with a 3.0Ghz P4 that went out in a blaze of glory in similar fashion. And unfortunately, it nuked the entire machine. Even one of my hard drives was playing "Knock, knock. Who's there?" with its read / write heads. Of course, that was the drive containing all my irreplaceable music demo data backups, but I digress.

RIP, was it an HT model or just a higher clocked earlier P4? 

[danomicar]

1 minute ago, An0maly_76 said:

Yes, very lucky. I had an old eBay special Dell Dimension 3000 with a 3.0Ghz P4 that went out in a blaze of glory in similar fashion. And unfortunately, it nuked the entire machine. Even one of my hard drives was playing "Knock, knock. Who's there?" with its read / write heads. Of course, that was the drive containing all my irreplaceable music demo data backups, but I digress.

Believe me, I was holding my breath when I pressed the power button again. Backing up the most important documents as we speak.

 

I had been looking for a reason to invest in an upgrade, but not like this! lol

[Spotty]

Can you post a photo of the label on the power supply so we can see what model it is. Also a photo of the part that exploded and is currently lodged in the fan blades with the text on it visible. I might be wrong but i think that looks more like a NTC thermistor, would need to know what the text printed on it says. From that angle it looks like it might say [N]TC.. 15

 

Was the system under load when it failed? How long was it on for? Did you just turn it on? Any power surges like a lightning storm?

[An0maly_76]

Just now, Mel0n. said:

RIP, was it an HT model or just a higher clocked earlier P4? 

I actually goofed, mine was a 4700, the 3000 was one I previously bought for a PC-based CCTV setup. Both were great machines that served their intended purposes quite well. The 4700 was purchased around 2009 and served me well with only a 4GB RAM and 1TB HDD upgrade, dying just before Christmas 2018, which was when I was introduced to Ryzen power. I don't remember for certain, but I believe the 4700 only came with a 3GHz P4 / 915G chipset.

[An0maly_76]

14 minutes ago, danomicar said:

Believe me, I was holding my breath when I pressed the power button again. Backing up the most important documents as we speak.

 

I had been looking for a reason to invest in an upgrade, but not like this! lol

Like Samuel Adams beer states -- Always a good decision...

But if I drank beer, I'd prefer Samuel Jackson beer...

 

 

[danomicar]

5 minutes ago, Spotty said:

Can you post a photo of the label on the power supply so we can see what model it is. Also a photo of the part that exploded and is currently lodged in the fan blades with the text on it visible. I might be wrong but i think that looks more like a NTC thermistor, would need to know what the text printed on it says. From that angle it looks like it might say NTC.. 15

 

Was the system under load when it failed? How long was it on for? Did you just turn it on? Any power surges like a lightning storm?

Thanks for your reply. 

 

Sure thing, photos are below. You are correct that it is an NTC component. I believe it says NTC 50-15.

 

System was under heavy load at the time, likely running near maximum. No lightning storms in the area at the time. Connected to a decent-quality GE surge protector.

 

 

[Fl0yd-]

my bet is the GFCI equivalent in the psu failed and it overloaded.

[Spotty]

2 minutes ago, danomicar said:

Sure thing, photos are below. You are correct that it is an NTC component. I believe it says NTC 50-15.

The NTC thermistor is used on the primary side (AC input) to limit inrush current spikes. I'm not entirely sure what would have caused it to blow open like that. The good news is at least since it is on the primary side that means it's unlikely the rest of the system was damaged.

 

Good lesson not to cheap out on power supplies I guess. 

[danomicar]

15 minutes ago, Spotty said:

The NTC thermistor is used on the primary side (AC input) to limit inrush current spikes. I'm not entirely sure what would have caused it to blow open like that. The good news is at least since it is on the primary side that means it's unlikely the rest of the system was damaged.

 

Good lesson not to cheap out on power supplies I guess. 

Thanks for the informative reply.

 

I agree that it's most likely just a crappy power supply in my pre-built iBUYPOWER system. It was under heavy load at the time but has certainly held up to 24+ hours of holding similar loads for the 3 years ive owned it. 

 

Definitely a nail in the coffin for any potential pre-built systems - would much prefer to know exactly what I'm running. 

[An0maly_76]

15 minutes ago, danomicar said:

Definitely a nail in the coffin for any potential pre-built systems - would much prefer to know exactly what I'm running. 

Precisely why I built what I built. Of course, some bad info led to some interesting issues the first few months, but I believe I have that cured now. My advice, if you're going to build a custom PC, you might want to give this a read before you make any final decisions on component choices.

 

https://linustechtips.com/topic/1451215-the-qvl-is-real-anyone-upgrading-their-pc-or-building-a-new-one-needs-to-read-this/

 

This is what my build is currently rocking...

 

https://pcpartpicker.com/user/An0maly1976/saved/9s2npg

Edited August 25, 2022 by An0maly_76
Revised, more info

[Elisis]

4 hours ago, An0maly_76 said:

but what really counts is the 80+ rating in efficiency and quality of the internal components.

There's a ton more in a PSU that matters than just these. I wouldn't consider the efficiency to be that important all that much either to be honest.

[An0maly_76]

6 minutes ago, Elisis said:

There's a ton more in a PSU that matters than just these. I wouldn't consider the efficiency to be that important all that much either to be honest.

So internal component quality shouldn't be a primary concern?

[Elisis]

16 minutes ago, An0maly_76 said:

So internal component quality shouldn't be a primary concern?

That is absolutely not what I said

[An0maly_76]

Here, I'll quote it for you...

  

5 hours ago, An0maly_76 said:

I haven't heard of this brand either. Probably a crap PSU with cheapie components that was pushed beyond its limits. Newer, better-quality ones use 105C rated capacitors which are far better. More wattage is always better, but what really counts is the 80+ rating in efficiency and quality of the internal components.

34 minutes ago, Elisis said:

There's a ton more in a PSU that matters than just these. I wouldn't consider the efficiency to be that important all that much either to be honest.

26 minutes ago, An0maly_76 said:

So internal component quality shouldn't be a primary concern?

10 minutes ago, Elisis said:

That is absolutely not what I said

 

I clearly said what counts are efficiency ratings and quality of the internals, to which you said there's a ton more that matters than just these. So when I ask you to clarify if you are saying the quality of the internals shouldn't a primary concern, which is more or less what was said, you now say that wasn't what you said? So what exactly are you saying? Because that very much appears to be what you're saying when you say there's a ton more in a PSU that matters than just internal component quality.

Edited August 25, 2022 by An0maly_76
Revised

[Elisis]

11 hours ago, An0maly_76 said:

there's a ton more in a PSU that matters than just internal component quality.

Yes, I'm saying that there other things that count than only efficiency or component quality.

[An0maly_76]

2 hours ago, Elisis said:

Yes, I'm saying that there other things that count than only efficiency or component quality.

Such as? Seems to me efficiency and component quality should top the list. Nobody wants a janky PSU that sucks too much power and doesn't last. But if there's something that bears more consideration, by all means, I'd like to know.

[seon123]

7 minutes ago, An0maly_76 said:

Such as? Seems to me efficiency and component quality should top the list. Nobody wants a janky PSU that sucks too much power and doesn't last.

Same way efficiency and silicon quality are totally the main priorities for a CPU, and not things like IPC, core count, frequency, cache... Clearly an ultra efficient 9W Intel i7 1250U is a better product than a 150W 12900KS for a desktop computer.

 

PSU topology (a low end PSU that uses a double forward or ACRF topology will have coil whine with higher end GPUs, while PSUs with LLC resonant converters are far less likely), independent and tight voltage regulation (low end group regulated PSUs regulate 5V and 12V together, meaning e.g. 12V heavy loads will make the 5V high and 12V low), low ripple, good transient response, long hold up time, advanced fan control, protections, low noise, no in cable capacitors, just off the top of my mind. If you are completely clueless about PSUs, please have the decency to not try to defend your faulty logic and statements, and instead just accept that you're wrong and were called out for it.

[An0maly_76]

14 minutes ago, seon123 said:

Same way efficiency and silicon quality are totally the main priorities for a CPU, and not things like IPC, core count, frequency, cache... Clearly an ultra efficient 9W Intel i7 1250U is a better product than a 150W 12900KS for a desktop computer.

 

PSU topology (a low end PSU that uses a double forward or ACRF topology will have coil whine with higher end GPUs, while PSUs with LLC resonant converters are far less likely), independent and tight voltage regulation (low end group regulated PSUs regulate 5V and 12V together, meaning e.g. 12V heavy loads will make the 5V high and 12V low), low ripple, good transient response, long hold up time, advanced fan control, protections, low noise, no in cable capacitors, just off the top of my mind. If you are completely clueless about PSUs, please have the decency to not try to defend your faulty logic and statements, and instead just accept that you're wrong and were called out for it.

I'm certainly willing to admit I don't know everything. I never CLAIMED to know everything. Nobody does. I'm really not sure where that impression is coming from, other than possibly my asking questions is coming off as challenging others, which is not my intention and quite frankly, a bit of pot and kettle. But the things you've mentioned here would very much seem to fall under component quality, which is what I put emphasis on. So if it is apples and oranges somehow, I'd like to know why.

Edited August 25, 2022 by An0maly_76
Revised, more info

[seon123]

1 minute ago, An0maly_76 said:

I'm certainly willing to admit I don't know everything. I'm really not sure where that impression is coming from. But the things you've mentioned here would very much seem to fall under component quality, which is what I put emphasis on. So if it is apples and oranges somehow, I'd like to know why.

You can use high end, high quality components on a double forward PSU, and you can use low end, cheap components on an LLC PSU. In both of those cases, the component quality is clearly separate from the topology of the PSU. An extreme case would be using very high end, very high quality diodes, connecting them to some cables, and plugging them straight from the wall to the PC. This "PSU" would obviously burn out and also kill the PC immediately. The component quality is very good, but it's not a good PSU.

 

The other points were about the performance, not the component quality. You can use high end components, but end up with a PSU that performs poorly. The same way Intel can use the same silicon quality to produce low end Pentium CPUs and high end i9 CPUs. Same quality, different performance. Maybe we interpret "component quality" as different things, but to me, it refers to the quality of the individual parts that are being used, not how they are used or assembled.

[An0maly_76]

25 minutes ago, seon123 said:

You can use high end, high quality components on a double forward PSU, and you can use low end, cheap components on an LLC PSU. In both of those cases, the component quality is clearly separate from the topology of the PSU. An extreme case would be using very high end, very high quality diodes, connecting them to some cables, and plugging them straight from the wall to the PC. This "PSU" would obviously burn out and also kill the PC immediately. The component quality is very good, but it's not a good PSU.

 

The other points were about the performance, not the component quality. You can use high end components, but end up with a PSU that performs poorly. The same way Intel can use the same silicon quality to produce low end Pentium CPUs and high end i9 CPUs. Same quality, different performance. Maybe we interpret "component quality" as different things, but to me, it refers to the quality of the individual parts that are being used, not how they are used or assembled.

So what you're saying, and I presume Elisis was hinting at, was that component quality and design quality are, in fact, different things. Which suggests to me, that the worst components can last if the design utilizes them better, and the best components in the world won't last if the design is poor. And I can see where that makes sense. It might have been easier had Elisis just said that in the first place, or the second. Can't read minds, you know.

 

I do have a knack for rubbing people the wrong way without intending to, because to some, I come across differently than intended. Perception gap, I call it, and it goes both ways. I'm just finding out recently that I'm autistic, and not only is this common with autism, COVID fog seems to have exacerbated those traits. Not to mention perceptions can vary widely in an Internet post, much like one person passing a message through thirty people in the room, they'll all interpret it differently and cause the message to be different by the time it reaches the end of the line.

 

So I get the general idea that I might rub some people the wrong way. Unfortunately, it doesn't really register to me in the moment. So there are two solutions. Ask others for some understanding and patience, or just don't interact with anyone. The latter of which is not good for mental health even without autism. But I will say that perceiving someone as a certain way without knowing their true intentions is a bit prejudicial.

 

That said, thanks for the clarification. Hopefully we've both learned something here.

[An0maly_76]

@seon123

 

Also, and this goes back to the design thing. My understanding is that in tems of efficiency...

 

80+ Gold > 80+ Bronze

80+ Platinum > 80+ Gold

 

And so forth. Does a better design not necessarily equate to better efficiency?