Concerning issue with PSU?

[Spotty]

1 minute ago, jonnyGURU said:

Do you know what you're talking about or maybe you're just trolling me....

They said power supplies should have protections to stop them from destroying hardware when they blow up.

Someone replied saying "it doesn't matter if you get a RMx or hi [sic] power supply"

 

They mistook "RMx and hi" as being a form of protection (like ocp or uvp etc).

The other person was just suggesting with good quality power supplies (like the Corsair RMx/HXi models) you don't need to worry about it blowing up so it's not a concern.

[Guest]

7 minutes ago, Spotty said:

They said power supplies should have protections to stop them from destroying hardware when they blow up.

Someone replied saying "it doesn't matter if you get a RMx or hi [sic] power supply"

 

They mistook "RMx and hi" as being a form of protection (like ocp or uvp etc).

The other person was just suggesting with good quality power supplies (like the Corsair RMx/HXi models) you don't need to worry about it blowing up so it's not a concern.

Oh... Ok.

 

That makes sense.  Because from the direction the conversation was taking, it seemed like someone was wanting more specifications on paper.... but without anyone actually forced to implement it correctly.  Which I have to admit is a big problem in the industry.

 

Aris is doing great things trying to hold vendors at task with his Cybenetics program.  He's caught some corner cases for Corsair that we overlooked.  But his program (the beta testing one he does... Not the forward facing noise and efficiency one) is voluntary.  Not required.  

 

We can "define" all the technology and specifications we want.  But until we "enforce" implementation, then most of the PSUs on the market are going to be garbage.

 

And with implementation being a cost adder and most people seeing only "it's 500W" instead of "is it a GOOD 500W", most companies aren't going to make the investment so they can keep the extra money in their pockets.

 

[Bombastinator]

6 minutes ago, jonnyGURU said:

So you want there to be MORE protections available, but you don't care if they're actually implemented or not?

 

There are only so many protections you can have.  It's whether or not protections are implemented or not is the question.

 

Nothing actually REQUIRES a PSU to have OCP, OTP, OPP, UVP, OVP, etc.  You can buy a PSU today that has none of those... or has them implemented very poorly (i.e. UVP doesn't trip until +12V is down around 6V or OPP doesn't trip until the PSU is 200% overloaded).   A good PSU will have all of the necessary protections AND have them implemented correctly.

 

Do you know what you're talking about or maybe you're just trolling me....

I can state quite conclusively that I don’t know much about modern PSUs.

What little I do know is extremely old. Pentium III old.

 I can also state quite conclusively that I’m not trolling.  I don’t troll.  At least to my understanding of what trolling is which is to ask deliberately wrong questions or make deliberately inflammatory remarks in an attempt to provoke conversation.

 

This may be an issue where I know even less than you assume I do.

 

What it seems like at the moment is what I was interested in is the various 3 letter acronyms you mentioned.  I can decode UVP undervolt protection, and OVP as overvolt protection.  

39 minutes ago, jonnyGURU said:

Yes.  Once upon a time all PSUs were garbage.  For the same reasons a lot of SI's still use garbage PSUs today.

 

1.  There's no "bullet point" for PSU quality you can put on your spec sheet.  CPU is this fast, GPU is this fast, this much RAM, this much storage... .PSU?  Oh..... 500W something or another.

2. There used to be no requirement for safety for PSUs installed by end users.  Though now I feel we have too many (why do we need BSMI to sell into Taiwan when we already have CE to sell into Europe and CCC to sell into China and KCC to sell into Korea?)  Since the PSU was installed by an end user, the liability of failure is put on the end user doing something wrong rather than the PSU being garbage.

3.  PFC wasn't even required in the EU until the 2010 or so.

 

4. Warranties have gotten longer.  Used to be a PSU would only have a 1 year warranty.  Many only have a 2 year warranty because of local consumer laws.

5.  Competent PSU reviewers are few and far between.  I can tell you a CPU or GPU is fast by running benchmark software.  How did early reviewers tell you your PSU was Ok?  Oh yeah.... "The PC turned on."

I could go on....

This post I missed entirely which may be the root of the problem.

 

 

[Turbof1]

5 hours ago, jonnyGURU said:

80  PLUS certificaion is a badge of efficiency. Not quality.

 

Too bad power supplies are so much more than a box of capacitors.

 

99% of the time, a failed PSU is not a failed PSU because a cap popped.

 

Also, old caps don't necessarily bulge.  What causes bulging caps is a chemical reaction that builds up pressure in the cap.  If the plug at the bottom dries out due to age, they'll leak at the bottom and not bulge at the top.  If the water in the electrolyte evaporates, you may not see signs of that until it's charged and pops.  Sometimes the oxide layer can break up, but there's no visual signs of this, though if you know if a capacitor is only suffering from lack of oxide layer, those capacitors can be "reformed" by slowly charging them to their specified capacity.

 

 

It's not going to get issues soon. The 80plus however is important for reliability. It's not just an efficiency thing; a psu operation near its lower or upper limits is a stressed psu, and will be quicker to fail in the long run. 80plus ensures it has a good stable power draw on the low and high limits.

 

Caps usually are an indicator. Can anything else go wrong? Of course. Is a cheap psu more likely to break down of something else? Sure. Is it good practice to actually rip off the warranty seal to remove dust and inspect the components and ask question you doubt yourself? Very much so (unless you want to risk your pc components to avoid breaking the warranty).

 

There is always risk involved. Let's not get tangled up in fearmongering about cheap psu's anybody basically knows. And should it not end well, we'll atleast have a nice campfire around my pc.

[seon123]

Just now, Turbof1 said:

 

It's not going to get issues soon. The 80plus however is important for reliability. It's not just an efficiency thing; a psu operation near its lower or upper limits is a stressed psu, and will be quicker to fail in the long run. 80plus ensures it has a good stable power draw on the low and high limits.

 

Caps usually are an indicator. Can anything else go wrong? Of course. Is a cheap psu more likely to break down of something else? Sure. Is it good practice to actually rip off the warranty seal to remove dust and inspect the components and ask question you doubt yourself? Very much so (unless you want to risk your pc components to avoid breaking the warranty).

 

There is always risk involved. It's a

You are trying to correct someone working with PSU RnD at Corsair for a living, and who is the founder of one of the most famous PSU review sites. I recommend you do some basic research, and accept that you are very wrong.

[LogicWeasel]

4 minutes ago, Turbof1 said:

It's not going to get issues soon. The 80plus however is important for reliability. It's not just an efficiency thing; a psu operation near its lower or upper limits is a stressed psu, and will be quicker to fail in the long run. 80plus ensures it has a good stable power draw on the low and high limits.

Everything seon123 said +1

Also you should read this so you have more of an idea of what you're (trying to) talk about:
https://linustechtips.com/main/topic/423141-80-plus-efficiency-and-what-it-really-means/

[Turbof1]

7 minutes ago, seon123 said:

You are trying to correct someone working with PSU RnD at Corsair for a living, and who is the founder of one of the most famous PSU review sites. I recommend you do some basic research, and accept that you are very wrong.

I'm not trying to correct him. He'll most likely know this himself and might have not seen this important himself, but the 80plus does contribute to reliability. To how much it actually does, that I'll leave myself uneducated on that.

 

I do want to stress I only came here to ask about the resin. Somebody also friendly asked about if there was corossion on the board (which there was none, luckily). I did not ask about a lecture on cheap psu's or old psu's. I know there are certain risks involved. That's on me. It's not like I will come with pointing fingers if the pc burst in flames and takes my house with it.

 

But lets say we turn this into a discussion about 80plus reliability, then if you want point where for instance this article is wrong:

https://www.acnodes.com/content/80plus-power-supply/

 

Genuinely, if the notion 80plus contributing to reliability is wrong, then there's a wrong general understanding. For which I will apologize if that is the case.

[LogicWeasel]

Just now, Turbof1 said:

but the 80plus does contribute to reliability

[Turbof1]

14 minutes ago, LogicWeasel said:

here's the link you missed:

https://www.acnodes.com/content/80plus-power-supply/

 

Also:

On 8/4/2015 at 8:03 PM, STRMfrmXMN said:

- "Higher 80 PLUS efficiency keeps the PSU cooler." Not to any serious degree, but this is technically true. A less efficient PSU will waste more electricity and wasted electricity is turned into heat. This is not likely to have an appreciable impact on the temperature of your room or system however as your system doesn't really draw that much power, thus it's better to optimize your system's airflow before throwing an AX1500i in your system to minimize heat created by the power supply. Since PSUs exhaust heat anyways the temperature of your system's hardware will not be impacted to any noticeable degree. Different PSUs also handle cooling differently and 80 PLUS efficiency doesn't correlate to the size of the fan used or the heat-dissipation abilities of the unit.

 

[Guest]

29 minutes ago, Turbof1 said:

 

 The 80plus however is important for reliability. It's not just an efficiency thing; a psu operation near its lower or upper limits is a stressed psu, and will be quicker to fail in the long run. 80plus ensures it has a good stable power draw on the low and high limits.

That's is a gross generalization and in most cases completely false.

 

You can make a garbage PSU that passes 80 PLUS testing.  Has nothing to do with reliability.

 

Too bad Kyle shut down HardOCP.  Paul wrote an excellent article on why 80 PLUS is absolute complete bullshit when it comes to PSU quality and reliability.

 

 

 

[Turbof1]

1 minute ago, jonnyGURU said:

That's is a gross generalization and in most cases completely false.

 

You can make a garbage PSU that passes 80 PLUS testing.  Has nothing to do with reliability.

 

Too bad Kyle shut down HardOCP.  Paul wrote an excellent article on why 80 PLUS is absolute complete bullshit when it comes to PSU quality and reliability.

 

 

 

So the improved efficiency is not reducing waste heat, which in turn can impact reliability of components?

 

Look, I fully take it at this point I initially gave too much credibility on the 80plus label as a reliability thing. And sure, I am neither denying a crappy psu can get up to 80plus certification. And I am sure you'll agree there's technically the matter of waste heat that although a fan is more than capable of able to handle, is nice to have to boost reliability by either 10%, 1% or 0.1% (I frankly don't know and neither will claim to know the average impact; they'll know at Corsair).

 

So if you want to point I'm naive with an old psu, then there's a good chance you are correct. I frankly don't care. It hasn't let me down up to this day; I try to take care of it as best as I can. I'm not going to replace it due it powering basically a 2010/2011 system. I'm not going to wake up in the night covered in sweat because of a risk literally nobody is going to able to quantify unless you take the psu to a specialised lab. All I and anybody else can judge it, is the age, its consistency and how certain components look. Good indicators? No. Are there better indicators? If you have and I can do this at home, then please share. If not, pray it will blow it up. Again; I can't care about a risk that is not quantifiable and would at worse take 60 euro's with it.

[Guest]

47 minutes ago, seon123 said:

You are trying to correct someone working with PSU RnD at Corsair for a living, and who is the founder of one of the most famous PSU review sites. I recommend you do some basic research, and accept that you are very wrong.

That's actually neither here nor there.  All someone has to do is just look at 80 Plus's testing methodology to understand that 80 PLUS is meaningless beyond efficiency.  Instead, some store that nobody's heard of that has misinformation is linked.  Pity. 

 

80 Plus merely looks at how much power is converted from AC to DC.  The tests are done at room temperature and not for extended periods of time.  There is no DMTBF, no tear down, no component analysis, no confirmation of what protections are implemented and of they implemented correctly.  There's no analysis of long term component life. No show of derate of performance at higher temperatures.  Nothing.  Nothing but telling you how much of the AC is converted to DC and what the power factor is.  That's it. 

[Guest]

5 minutes ago, Turbof1 said:

So the improved efficiency is not reducing waste heat, which in turn can impact reliability of components?

 

 

Improved efficiency does mean there is less heat generated.  This is true.  That doesn't mean the product is more reliable. 

[Turbof1]

13 minutes ago, jonnyGURU said:

Improved efficiency does mean there is less heat generated.  This is true.  That doesn't mean the product is more reliable. 

So basically we are kinda agreeing, but are having a different viewpoint at this moment. it is a matter of whether you take waste heat in isolation as a reliability differentiator, or look at the whole product for the overall reliability (because of all reliability differentiators, not just one, being looked at).

 

I do want to thank you emphasizing 80plus is not per se a "quality label"; I might have taken that a bit too much for granted in hindsight. For the record, the best any average consumer can do to determine a quality psu, is to buy from an established brand. Not really an objective or standardized indicator of quality, but nothing much else. Maybe that's also why people see 80plus too much as a quality label, due to the lack of an actually standard reliability label.

[Guest]

11 minutes ago, Turbof1 said:

So basically we are kinda agreeing, but are having a different viewpoint at this moment. it is a matter of whether you take waste heat in isolation as a reliability differentiator, or look at the whole product for the overall reliability (because of all reliability differentiators, not just one, being looked at).

 

I do want to thank you emphasizing 80plus is not per se a "quality label"; I might have taken that a bit too much for granted in hindsight. For the record, the best any average consumer can do to determine a quality psu, is to buy from an established brand. Not really an objective or standardized indicator of quality, but nothing much else. Maybe that's also why people see 80plus too much as a quality label, due to the lack of an actually standard reliability label.

There's plenty of proper PSU reviews done by reviewers that know what they're doing.   Sure, they're not videos from guys like Linus or JayZtwocents, which requires a person to actually read.  But they're out there.  So you don't have to actually "guess" or rely on a brand that is "typically good" (most brand names offer a full gamut of product from garbage to great).

 

A good place to start is here:  http://www.realhardtechx.com/index_archivos/PSUReviewDatabase.html

 

Another good place to go is here:  https://www.cybenetics.com/

 

[LogicWeasel]

1 hour ago, Turbof1 said:

it is a matter of whether you take waste heat in isolation as a reliability differentiator, or look at the whole product for the overall reliability (because of all reliability differentiators, not just one, being looked at)

You are still making the leap that just because one PSU creates something like 10 watts more heat while operating that somehow damages it's long term reliability because it's at 27 C and not 25 C and it's really not acknowledging the idea that when you can have caps rated for 40 C constant that makes no difference on long term use or derating of power delivery.  Slight extra waste heat alone I do not ever consider to be even a single differentiator

[Spotty]

2 hours ago, jonnyGURU said:

There's plenty of proper PSU reviews done by reviewers that know what they're doing.   Sure, they're not videos from guys like Linus or JayZtwocents, which requires a person to actually read. 

Do you know if gamers Nexus are still planning on doing PSU reviews? They had the video with yourself in it then the Great Wall unit that was in the Walmart PC but after that I don't think I've seen them touch any power supplies.

 

41 minutes ago, LogicWeasel said:

when you can have caps rated for 40 C constant

40°C? I'm guessing you're talking about the power supply rated temp or just ambient room temp...?

Capacitors used in power supplies are typically rated between 85-105°C.

There's a formula where you can work out the life expectancy of capacitors based on the load and temperature. I can't remember the formula but Jon might share it with us. Basically any decent power supply manufacturer should be doing that calculation to ensure that the capacitors will last at least through the warranty period.

[Guest]

17 minutes ago, Spotty said:

40°C? I'm guessing you're talking about the power supply rated temp or just ambient room temp...?

 

I knew what he meant.  You would never want to run 85°C caps at 85°C and you would never want to run 105°C caps at 105°C.  If you do the cap life math, you find they don't last at all. 

 

[Turbof1]

5 hours ago, LogicWeasel said:

Slight extra waste heat alone I do not ever consider to be even a single differentiator

Than let us just agree to disagree. Frankly at this point I don't care anymore

 I could give a whole explanation on statistics, normal distribution and how that 2°C helps pushing margins, but I honestly think this discussion has run its course.

[Mezoxin]

20 minutes ago, Turbof1 said:

Than let us just agree to disagree. Frankly at this point I don't care anymore

 I could give a whole explanation on statistics, normal distribution and how that 2°C helps pushing margins, but I honestly think this discussion has run its course.

heat is a factor in reliability ofcourse  no one is disagreeing here and  it is has an operating envelope for any component  , if we  talk about heat only  that badge only certefies the produced heat , but has nothing to do with the product ability to dissipate that heat (quality of heatsinks and ventilation design) , has nothing to do with the internal components thermal ratings ( as some capacitors are rated at 85c and other at 105C) so you can see how misleading can this certification when you are trying to judge how heat  can affect  the PSU operation

[Turbof1]

22 minutes ago, Mezoxin said:

So you can see how misleading can this certification when you are trying to judge how heat  can affect  the PSU operation

Yes. I did already agreed with that. My point was taking all other things equal, heat is a factor. Of course, there are a lot of other factors, all probably even more significant, like materials used, the cooling, the ventilation, even air humidity. Again it is no so much a full disagreement, rather a different viewpoint.

 

It is funny that for all the arguments everybody is making, nobody mentions external factors . The way you treat your electronics is equally as important as the build quality.

[Guest]

8 hours ago, Turbof1 said:

Yes. I did already agreed with that. My point was taking all other things equal, heat is a factor.

That's the problem.  All other things are not equal.  That would make things so much easier, but that's not the case.

 

The devil is in the details.  Weekly, we're changing parts in BOMs (replacing UBIQ MOSFETs with APEC MOSFETs; replacing APEC MOSFETs with Infineon MOSFETs) after failure analysis due to unexpected failures, not because of heat, but because of a plethora of other things.

By the way... did you notice how none of those unexpected failures I listed weren't capacitors? 

 

[Turbof1]

4 hours ago, jonnyGURU said:

That's the problem.  All other things are not equal.  That would make things so much easier, but that's not the case.

 

The devil is in the details.  Weekly, we're changing parts in BOMs (replacing UBIQ MOSFETs with APEC MOSFETs; replacing APEC MOSFETs with Infineon MOSFETs) after failure analysis due to unexpected failures, not because of heat, but because of a plethora of other things.

By the way... did you notice how none of those unexpected failures I listed weren't capacitors? 

 

I did. I also never claimed that a psu will only ever fail due to capacitors . I lack the knowledge and the tools to do an extensive inspection of an old psu, I think most of us, with you of course being the big exception, will not be able to do anything of maintenance outside of blowing out the dust, checking the capacitors and otherwise checking for anomalies (and let us not mince words, most people do not even do that).

 

The only thing any normal person can do outside that, is replacing the psu that up to now has always worked and never showed anything out of the ordinary. Is it recommended to do so? Between it never have failed or showing any visible signs it will fail in the near future, the high chance that if it fails it will not take the whole system with it (because you know, there are safeties build in) and even if it should take the whole system with it, it's a loss of 60 euro's in hardware. If it was 5-7 year old hardware, I'd probably be thinking about replacing the psu.

 

Let me infact give you a heart attack. My other pc has much newer hardware, around 3-4 years old on average. The psu in that one is from 2005. Currently. I am looking to replace it, but I am currently stuck with it and try to leave that pc off.

[Bombastinator]

9 minutes ago, Turbof1 said:

I did. I also never claimed that a psu will only ever fail due to capacitors . I lack the knowledge and the tools to do an extensive inspection of an old psu, I think most of us, with you of course being the big exception, will not be able to do anything of maintenance outside of blowing out the dust, checking the capacitors and otherwise checking for anomalies (and let us not mince words, most people do not even do that).

 

The only thing any normal person can do outside that, is replacing the psu that up to now has always worked and never showed anything out of the ordinary. Is it recommended to do so? Between it never have failed or showing any visible signs it will fail in the near future, the high chance that if it fails it will not take the whole system with it (because you know, there are safeties build in) and even if it should take the whole system with it, it's a loss of 60 euro's in hardware. If it was 5-7 year old hardware, I'd probably be thinking about replacing the psu.

 

Let me infact give you a heart attack. My other pc has much newer hardware, around 3-4 years old on average. The psu in that one is from 2005. Currently. I am looking to replace it, but I am currently stuck with it and try to leave that pc off.

Ah! Dirt ignorance is useful in this case.  Technical knowledge I lack but dirt ignorance I have mountains of.  

 

The old old school opinion which was when overvolt protection existed but under volt protection was new (I think.  I may have this backwards or just wrong) was any good well made PSU will last 5 years with a nearly zero percent chance of failure, but that number went up kind of steeply after that so the move was to buy a good PSU and replace it every five years whether it needed it or not just to make sure you never even got close to having a problem, because problems tended to be very bad.  My understanding is that there are more and better safety features now  (the particulars of which I am not clear on) and when they are done correctly (which is apparently much less common than it should be) they can increase this 5 year time.  I do not know what it increases to though.

[Turbof1]

3 minutes ago, Bombastinator said:

I do not know what it increases to though.

I'm not sure on the average, but a 15 year old psu still working (and just to be clear, don't do this. I am an idiot for doing this!) should give an indication of what can be reached.

 

Sometimes you need a jack*ss to get these numbers