PSU tier list 2.0

[jankerson]

Like this guy:
http://www.tomshardware.com/forum/id-3583394/boosting-125-budget-setup.html#20563765

£20 case with included 500watt PSU
"What's so bad about the PSU anyway... isn't a500w PSU a 500w PSU..? "

Yeah. LOL 😀

It doesn't help that some people tell them not to look at the tier lists at all when we point them to them for a point of reference.

That really doesn't help.

 

[marksavio]

this is the best list everrrrrrrrrrrr on the internet. thanks for updating and im happy how both my PSUs are put on the right place. keep it up!

 

[turkey3_scratch]

I was about to say, it's not EVER in ANYBODY's self interest to cheap out on the one thing that can ruin everything else. Anybody who believes that is either befuddled or besotted.

We're not talking about getting a $20 PSU, more talking about getting a low-end okay PSU versus a high end one with better safety and protections.

For instance, buy the S12ii and you risk your computer by buying a cheaper power supply that lacks overcurrent protection on the minor rails and has no undervoltage protection. This puts your computer more at risk. Yet, people buy the S12ii over a more expensive PSU with actually all properly implemented protection circuits because they'd rather save the money and take the risks because they believe it's in their best interests. In this case, our job is to inform them about its lack of protections, and then leave it up to them to decide.

Likewise, a doctor should not be able to force a patient to get surgery. It should always be the patient's choice because the doctor never knows the patient's best interests, only the patient does. The doctor only knows the best health interests, but one's best interests are more of a mental thing.

 

[Karadjgne]

I do believe the S12-II has a form of over-current protection on the minor rails, just not a set trigger like dc-dc psus use. And you don't need undervoltage protection to conform to ATX standards. That's just one of those extra protections more expensive units include. Just because it's an old design doesn't make it suddenly dangerous to pc health, just makes it old.

 

[Someone Somewhere]

I've seen plenty of people saying 'but it's worked fine for years'.

The issue is risk: it's a low-ish chance of a high-impact event. Humans notoriously underestimate these.

If there's as much as a 1% chance of it blowing up or catching fire in the next year, it should be replaced ASAP. And yet, that means that 99/100 people that don't follow that advice will still say "everything's fine"

I 100% do not agree with this. Especially not when there are case after case after case of testing facilities like Corsair, for example, that could demonstrably show units (Not exactly bottom of the barrel either, just poor quality) not only NOT meeting their RATED capabilities, but burning up, repeatedly, test after test.

I do agree that it's not going to be in every case in the real world, but I think the risks are actually more real than your statements give credit to.

When pushed to the limits, yes. Many PCs don't break 200W, even gaming ones, unless stress tested. Most PSUs can handle that, even the bad ones.

The number of things I've looked at and said 'how the hell is that still working'...

I was about to say, it's not EVER in ANYBODY's self interest to cheap out on the one thing that can ruin everything else. Anybody who believes that is either befuddled or besotted.

Of course it can be. Higher price does not necessarily mean higher quality for starters; super-gold-plated stuff can be utter crap.

Add to that diminishing returns: Has your switchboard, in an old home, been retrofitted with arc fault and ground fault protection for every circuit, with the cable derated to worst possible installation conditions? Actually, strike that, feed every circuit from its own isolation transformer.

That's safer. Not doing it could burn your house down. But... it's not actually worth it. The money is better spent elsewhere, or not spent at all. There is a point at which reducing the risk further simply isn't useful.

 

[Darkbreeze]

Ok, you guys are off the rails now.

We're not talking about an S12-II being a "cheap" power supply. We're not even talking older CX units. Neither was seeingeyegod. He, and we, were talking about ACTUAL cheap power supplies. Like, 90% of Raidmax units, or iBall units, or Apevia, or the very, very common, sold with almost every bundle Tiger direct EVER sold Ultra units. Those are the cheap power supplies we're talking about. And based on units that could realistically fall under the "cheap" category, rather than just the "mediocre" category, I stand by what I've said.

However, as we well know there are also some better known names with units that are horrendously bad, like the old TR2 units, which, are still around in the form of NOS and being sold. Not the still crappy but not AS crappy newer revision of that unit. I think you guys knew what we were talking about.

 

[jankerson]

I do believe the S12-II has a form of over-current protection on the minor rails, just not a set trigger like dc-dc psus use. And you don't need undervoltage protection to conform to ATX standards. That's just one of those extra protections more expensive units include. Just because it's an old design doesn't make it suddenly dangerous to pc health, just makes it old.

I never got that, it's still the same PSU as it was, NOTHING HAS CHANGED as it's still the same unit.

So in effect it makes no difference at all if it's used in an i7 870 machine or an i7 8700, it will power both machines the same. The only real difference is the sleep states and that's not really that big a deal.

Yeah there are newer units out there, and better units from Seasonic, but that doesn't make the S12ii a BAD unit.

 

[jankerson]

Ok, you guys are off the rails now.

We're not talking about an S12-II being a "cheap" power supply. We're not even talking older CX units. Neither was seeingeyegod. He, and we, were talking about ACTUAL cheap power supplies. Like, 90% of Raidmax units, or iBall units, or Apevia, or the very, very common, sold with almost every bundle Tiger direct EVER sold Ultra units. Those are the cheap power supplies we're talking about. And based on units that could realistically fall under the "cheap" category, rather than just the "mediocre" category, I stand by what I've said.

However, as we well know there are also some better known names with units that are horrendously bad, like the old TR2 units, which, are still around in the form of NOS and being sold. Not the still crappy but not AS crappy newer revision of that unit. I think you guys knew what we were talking about.

Yeah it's those Tier 5 POS burn down your house PSU shaped objects that we are talking about.

 

[turkey3_scratch]

I do believe the S12-II has a form of over-current protection on the minor rails, just not a set trigger like dc-dc psus use. And you don't need undervoltage protection to conform to ATX standards. That's just one of those extra protections more expensive units include. Just because it's an old design doesn't make it suddenly dangerous to pc health, just makes it old.

It does not have overcurrent protection on the minor rails, it has been tested before in a review which I don't want to dig out (also, the 8-pin HY510N protection IC can only do so much).

You're missing the point, though. People keep saying it's idiotic to buy a cheap power supply because it is risky, but technically people are also taking risks by buying a power supply like the S12ii versus something like an RMx. I use the S12ii as an example because it's the perfect example - a PSU a lot of people buy but which doesn't have all the protections (and this is a fact, not some whimsical myth of my own).

How safe are we supposed to be? When you go out and buy a car, is your priority always safety? In fact, prioritizing safety of a car makes more sense than safety of a computer since it guards your life rather than a bunch of silicon. But a lot of people would be willing to buy a less safe car if they can save money because they calculated their own risks.

I don't see how we can tell people to never buy something like the crappy EVGA 400W power supply based on the sole premise that it is too risky, because that is within our own definitions of what "too risky" is and not the purchaser. If somebody is convinced to buy the more expensive power supply because they're told they need to and it's absolutely moronic to get the cheaper one, but then they end up unhappy because they acted against their best interests and the $20 extra they spent they wish they still had, are we really helping them out, or are we helping who we want them to be out?

 

[USAFRet]

I do believe the S12-II has a form of over-current protection on the minor rails, just not a set trigger like dc-dc psus use. And you don't need undervoltage protection to conform to ATX standards. That's just one of those extra protections more expensive units include. Just because it's an old design doesn't make it suddenly dangerous to pc health, just makes it old.

It does not have overcurrent protection on the minor rails, it has been tested before in a review which I don't want to dig out (also, the 8-pin HY510N protection IC can only do so much).

You're missing the point, though. People keep saying it's idiotic to buy a cheap power supply because it is risky, but technically people are also taking risks by buying a power supply like the S12ii versus something like an RMx. I use the S12ii as an example because it's the perfect example - a PSU a lot of people buy but which doesn't have all the protections (and this is a fact, not some whimsical myth of my own).

How safe are we supposed to be? When you go out and buy a car, is your priority always safety? In fact, prioritizing safety of a car makes more sense than safety of a computer since it guards your life rather than a bunch of silicon. But a lot of people would be willing to buy a less safe car if they can save money because they calculated their own risks.

I don't see how we can tell people to never buy something like the crappy EVGA 400W power supply based on the sole premise that it is too risky, because that is within our own definitions of what "too risky" is and not the purchaser. If somebody is convinced to buy the more expensive power supply because they're told they need to and it's absolutely moronic to get the cheaper one, but then they end up unhappy because they acted against their best interests and the $20 extra they spent they wish they still had, are we really helping them out, or are we helping who we want them to be out?

Point us to a recent thread where someone was called "moronic" for buying a S12ii.

Now....change "S12ii" for "RAIDMAX", and you'll see plenty.

 

[turkey3_scratch]

Point us to a recent thread where someone was called "moronic" for buying a S12ii.

Now....change "S12ii" for "RAIDMAX", and you'll see plenty.

That sentence was not talking about the S12ii.

 

[USAFRet]

Point us to a recent thread where someone was called "moronic" for buying a S12ii.

Now....change "S12ii" for "RAIDMAX", and you'll see plenty.

That sentence was not talking about the S12ii.

I see the word "idiotic" in the same sentence as "S12ii".
In any case...there is often little actual price difference between a good one and a crappy one.

But too many people just look at the advertised wattage, and assume that is the whole story. You've seen them, I've seen them.

But if you want to recommend someone saving $20 on the PSU so he can buy another fan or more blinken lights...go for it.

 

[jankerson]

I do believe the S12-II has a form of over-current protection on the minor rails, just not a set trigger like dc-dc psus use. And you don't need undervoltage protection to conform to ATX standards. That's just one of those extra protections more expensive units include. Just because it's an old design doesn't make it suddenly dangerous to pc health, just makes it old.

It does not have overcurrent protection on the minor rails, it has been tested before in a review which I don't want to dig out (also, the 8-pin HY510N protection IC can only do so much).

You're missing the point, though. People keep saying it's idiotic to buy a cheap power supply because it is risky, but technically people are also taking risks by buying a power supply like the S12ii versus something like an RMx. I use the S12ii as an example because it's the perfect example - a PSU a lot of people buy but which doesn't have all the protections (and this is a fact, not some whimsical myth of my own).

How safe are we supposed to be? When you go out and buy a car, is your priority always safety? In fact, prioritizing safety of a car makes more sense than safety of a computer since it guards your life rather than a bunch of silicon. But a lot of people would be willing to buy a less safe car if they can save money because they calculated their own risks.

I don't see how we can tell people to never buy something like the crappy EVGA 400W power supply based on the sole premise that it is too risky, because that is within our own definitions of what "too risky" is and not the purchaser. If somebody is convinced to buy the more expensive power supply because they're told they need to and it's absolutely moronic to get the cheaper one, but then they end up unhappy because they acted against their best interests and the $20 extra they spent they wish they still had, are we really helping them out, or are we helping who we want them to be out?

The S12ii/M12ii are proven solid platforms with a lot of years behind them so one can do a lot worse believe me.

Yes it's an older platform, but that doesn't change the fact that they are solid and proven.

Now are the G-Series and Focus better? Sure they are, but that doesn't make the S12ii and M12ii a bad choice.

 

[Darkbreeze]

Plus, the number of people who will run into problems or even NEED protections, because the power supply cannot even remotely supply what the label SAYS it can supply, are going to be practically non-existent on a unit like the S12II. Something is going to have to go terribly wrong in order for that to be a factor on one of those units, like using it where a higher capacity unit should have been used.

OTHER, really cheap units however, will see this problem almost universally, because they cannot even do a portion of what their label of advertised specifications SAY they can do/are rated for. You fully know this. I don't even know why you are arguing the issue. Seems like you're regressing rather than assuming the face of a member with enough wisdom by now to know better, as I'd have thought you'd be.

I know you wouldn't intentionally TELL somebody to buy a unit like that, but even arguing the fact as if it were perfectly acceptable to do something like that, here, where Joe Blow member from the nether regions of the web that isn't even a member is going to come read that statement seems risky to me.

 

[Someone Somewhere]

Ok, you guys are off the rails now.

We're not talking about an S12-II being a "cheap" power supply. We're not even talking older CX units. Neither was seeingeyegod. He, and we, were talking about ACTUAL cheap power supplies. Like, 90% of Raidmax units, or iBall units, or Apevia, or the very, very common, sold with almost every bundle Tiger direct EVER sold Ultra units. Those are the cheap power supplies we're talking about. And based on units that could realistically fall under the "cheap" category, rather than just the "mediocre" category, I stand by what I've said.

However, as we well know there are also some better known names with units that are horrendously bad, like the old TR2 units, which, are still around in the form of NOS and being sold. Not the still crappy but not AS crappy newer revision of that unit. I think you guys knew what we were talking about.

I am talking about those units. Even those units typically deliver 150W without catching fire, and you seem to be acting under the IMHO false assumption (or intentionally spreading false rumours) that the majority of units sold burst into flames. They don't.

Yes, they will fail when loaded up to 'specification' - which tends to be about triple the maximum real-world use, because the people who buy these PSUs are not the same people who run GPU+CPU stress tests simultaneously.

They even fail in a nasty way under certain circumstances, which causes the steady trickle of OMG MY PSU CAUGHT FIRE articles/posts. How many sales do you think there are compared to the number of posts? Hint: Lots.

Yes, there are legitimate concerns about using them. They're missing important safety features, may have design flaws, and may fail dangerously. As long as the buyer does *know* about this, that's generally their problem.

Proper government regulation would get rid of the low end, as most of the rest of the world has done. Until then, spreading falsehoods on forums doesn't help your cause.

 

[turkey3_scratch]

OTHER, really cheap units however, will see this problem almost universally, because they cannot even do a portion of what their label of advertised specifications SAY they can do/are rated for. You fully know this. I don't even know why you are arguing the issue. Seems like you're regressing rather than assuming the face of a member with enough wisdom by now to know better, as I'd have thought you'd be.

We're just talking about completely different things, as that is not an issue I'm arguing against.

 

[jankerson]

Ok, you guys are off the rails now.

We're not talking about an S12-II being a "cheap" power supply. We're not even talking older CX units. Neither was seeingeyegod. He, and we, were talking about ACTUAL cheap power supplies. Like, 90% of Raidmax units, or iBall units, or Apevia, or the very, very common, sold with almost every bundle Tiger direct EVER sold Ultra units. Those are the cheap power supplies we're talking about. And based on units that could realistically fall under the "cheap" category, rather than just the "mediocre" category, I stand by what I've said.

However, as we well know there are also some better known names with units that are horrendously bad, like the old TR2 units, which, are still around in the form of NOS and being sold. Not the still crappy but not AS crappy newer revision of that unit. I think you guys knew what we were talking about.

I am talking about those units. Even those units typically deliver 150W without catching fire, and you seem to be acting under the IMHO false assumption (or intentionally spreading false rumours) that the majority of units sold burst into flames. They don't.

Yes, they will fail when loaded up to 'specification' - which tends to be about triple the maximum real-world use, because the people who buy these PSUs are not the same people who run GPU+CPU stress tests simultaneously.

They even fail in a nasty way under certain circumstances, which causes the steady trickle of OMG MY PSU CAUGHT FIRE articles/posts. How many sales do you think there are compared to the number of posts? Hint: Lots.

Yes, there are legitimate concerns about using them. They're missing important safety features, may have design flaws, and may fail dangerously. As long as the buyer does *know* about this, that's generally their problem.

Proper government regulation would get rid of the low end, as most of the rest of the world has done. Until then, spreading falsehoods on forums doesn't help your cause.

I know it's not directed at me... But.

Don't you think that it would be in the best interest to warn people about them?

That and others that might be reading the threads in the future?

I steer people away from them, seen too many GPU's etc cooked due to garbage units.

We have however seen more than a few CX and CXM's burn here lately, that's actual smoke rolling... I know that's not the norm, but it does happen. Budget units and they are a known issue.

 

[Karadjgne]

Ask Ford Motor Company. As soon as news was released that the Pinto was likely to explode on a rear end collision, just how many Pintos were then sold. Just knowing that, you couldn't give me a Pinto, never mind sell me one. Doesn't even matter that it only happened in 10% of every rear end collision, it's a risk factor that's not worth the price. Knowing that a cheapo psu stands a better than average % chance of taking out other, more expensive components such as the mobo or gpu, is not worth the risk either. When you have older designs, that have sold millions of units worldwide like the S12-II, with a @1% failure rate vrs units like the TR2 with a 20% failure rate, and compare that to damages due to failure, the 0.1% chance of a unit like the S12-II burning up a system vrs the 10% chance of a unit like the TR2 doing the same is an acceptable risk. Any psu is suspect, no matter how well built or designed, chit happens, ask Murphy. To what degree and with what frequency is a whole different story.

 

[Someone Somewhere]

Don't you think that it would be in the best interest to warn people about them?

That and others that might be reading the threads in the future?

Absolutely. Telling them it is guaranteed to fail in a fireball and expressing disbelief when it doesn't, on the other hand, simply makes you look ignorant. Don't lie while trying to explain the dangers.

When you have older designs, that have sold millions of units worldwide like the S12-II, with a @1% failure rate vrs units like the TR2 with a 20% failure rate, and compare that to damages due to failure, the 0.1% chance of a unit like the S12-II burning up a system vrs the 10% chance of a unit like the TR2 doing the same is an acceptable risk. Any psu is suspect, no matter how well built or designed, chit happens, ask Murphy. To what degree and with what frequency is a whole different story.

Do you have a source on those failure rates, and what the conditions were (e.g. failed within warranty vs failed at any point)? I'd be most interested.

I expect only a very small proportion of those 20% caused damage to other components.

 

[jankerson]

Don't you think that it would be in the best interest to warn people about them?

That and others that might be reading the threads in the future?

Absolutely. Telling them it is guaranteed to fail in a fireball and expressing disbelief when it doesn't, on the other hand, simply makes you look ignorant. Don't lie while trying to explain the dangers.

When you have older designs, that have sold millions of units worldwide like the S12-II, with a @1% failure rate vrs units like the TR2 with a 20% failure rate, and compare that to damages due to failure, the 0.1% chance of a unit like the S12-II burning up a system vrs the 10% chance of a unit like the TR2 doing the same is an acceptable risk. Any psu is suspect, no matter how well built or designed, chit happens, ask Murphy. To what degree and with what frequency is a whole different story.

Do you have a source on those failure rates, and what the conditions were (e.g. failed within warranty vs failed at any point)? I'd be most interested.

I expect only a very small proportion of those 20% caused damage to other components.

I don't do the fireball thing, I am more cooking parts type and that is what we normally see. 😀

I just tell them there is a good chance that their parts could be damaged by said PSU.

 

[Onus]

I agree with DB. Whether or not caused by misunderstanding, this topic has somewhat wandered away from the reason we have a tier list. There are machines in which I'd use a tier-4 PSU and not worry about it, such as those in which integrated graphics were being used, and I'd advise using a surge suppressor or UPS; many basic business boxes that will draw 90W on a hard day qualify. There is NO PC, for any purpose, for which I'd buy or recommend a tier-5 unit.

 

[Darkbreeze]

I am talking about those units. Even those units typically deliver 150W without catching fire, and you seem to be acting under the IMHO false assumption (or intentionally spreading false rumours) that the majority of units sold burst into flames. They don't.

Actually, I never said ANYTHING like that, so I don't know where you get off. What I SAID, was that SOME cheap units, in EXTREME circumstances, COULD catch fire.

A cheap PSU fails, and has no protections, you lose other stuff too, maybe a fire in extreme cases.

I certainly didn't say anything about the majority of units sold, even cheap ones, bursting into flame. SO perhaps YOU should quit it with the accusations. I'm disliking this a lot. Accusations of spreading misinformation intentionally are not taken lightly, and I think perhaps you've forgotten yourself sir. I think I'll leave it at that before I too forget myself.

Also, to the extend that we DO sometimes wander away from the purpose of this tier list, I wonder if there are any of us who are interested in the idea that perhaps we should create a distinctly different thread strictly for non-tier list related power supply discussion? Certainly we have plenty of regular members that might be willing to participate in that and then we can avoid sidetracking this thread further.

http://www.tomshardware.com/forum/id-3612443/power-supply-discussion-thread.html

 

[Someone Somewhere]

Let me re-post this quote from you:

What is considerable is the fact that out of the thousands, if not tens of thousands of people, who through even just the last few years have come here and received similar advice concerning a poor quality power supply, you are the ONLY one I've seen come back and say "well, I ignored your advice and now one year later my system is still working fine."

THE.ONLY.ONE

You'd think out of all the thousands of people, if there was fault in the advice there'd surely be more than one. Why is that? Because you are the only person out of thousands who ignored the advice? Doubtful.

Because out of the thousands you are the only one who thought of coming back to refute the advice? Even more doubtful, because anybody who HAD thought they knew better and ignored it would have felt the same as you do and would have wanted to come back and gloat about it.

You've gotten lucky, and that's ALL it is.

No, it doesn't specifically call out fires. However, I read it as saying that you find it extremely rare that a cheap PSU could last one year, let alone four, and started accusing them of essentially cherry-picking data, or being an aberration.

99% of people will ignore your post, forget about it, and not come back to gloat or apologize whether it fails or not.

 

[turkey3_scratch]

@jankerson: I never disagreed with you on that we should warn people of the risks. We should definitely do that all the time. I'm just more in favor of presenting people with a lot of options and telling them what's right and wrong with each and then letting them decide for themselves.

@Karadjgne: I'm convinced you kind of just made up those numbers.

Also, I agree that it is rarely rarely in someone's best interest to ever buy a $20 PSU. Maybe only if they have a $300 office machine and can't afford anything better.

 

[jankerson]

@jankerson: I never disagreed with you on that we should warn people of the risks. We should definitely do that all the time. I'm just more in favor of presenting people with a lot of options and telling them what's right and wrong with each and then letting them decide for themselves.

@Karadjgne: I'm convinced you kind of just made up those numbers.

Also, I agree that it is rarely rarely in someone's best interest to ever buy a $20 PSU. Maybe only if they have a $300 office machine and can't afford anything better.

I think we have to tell them about the risks no matter what, yeah it hurts their feelings sometimes when we point out their PSU is a POS or the one they are looking at is total garbage.

I have a problem recommending budget units in gaming machines so I don't. I recommend known stable PSU's and those don't tend to be budget units.

Also the tier list needs to be updated and has for awhile so people can have a source to look at.