PSU Tier List [OLD]

[Guest]

1 hour ago, Stefan Payne said:

But you do.

I have an EVGA GS550 here that BLOW UP! 

Because of Overload -> DC-DC Converter blew. When that happens there is a great chance that you have +12V on minor rails, wich can destroy everything on those rails.

Oh look someone somewhere got a psu that blew up. 

[STRMfrmXMN]

3 minutes ago, deXxterlab97 said:

Oh look someone somewhere got a psu that blew up. 

He actually has (had?) the very same PSU I have

[79wjd]

4 minutes ago, STRMfrmXMN said:

He actually has (had?) the very same PSU I have

Why doesnt the lack of OCP push it down a tier? Isn't that a bad thing?

[STRMfrmXMN]

5 minutes ago, djdwosk97 said:

Why doesnt the lack of OCP push it down a tier? Isn't that a bad thing?

I'm fairly certain that's one of the main reasons the higher wattage GSs are at tier 1 and the others are tier 2.

[Stefan Payne]

1 hour ago, djdwosk97 said:

Why doesnt the lack of OCP push it down a tier? Isn't that a bad thing?

Exactly!

It is a very bad thing because the PSU _CAN_ damage your rig due to bad/out of spec voltage regulation...

And I have some Test for what happens when you don't have working OCP/OPP:

https://www.hardwareluxx.de/index.php/artikel/hardware/netzteile/29916-enermax-triathlor-eco-650w-und-lepa-maxbron-700w-im-test.html

 

On the Enermax Unit: They were able to push the voltage down to about 10V @ 83Amps. Ripple and Noise was way out of spec.

But it gets worse. They were able to push the Lepa to 100A @ 6V. And it still didn't switch off.

That this can kill your components should be obviouis. And a lower voltage is worse for DC-DC converters because of higher amperage!!

So with 10A or less, it's possible to kill the voltage regulators on your Mainboard or Graphics card.

 

 

 

And the EVGA GS550 I've got, blew due to lack of protection while overloading the minor rails!

That's the point of having Protections in this example OCP: To protect at least the PSU from damaging itself!

 

Why don't you want to understand it?! 

I'm writing someting and you don't even read what I've said...

 

And what I don't understand:

I'm telling you that this List is useless and how to make a good/helpful PSU Recomendations List and you attack me?!

 

WHY?! just Why?!

 

 

Oh and here you can take a look on some Reviews I did of some PSUs:

http://www.tweakpc.de/forum/netzteile/

You should at least take a look at that and then see if I know my stuff or not. 

Of course, it's not in english. I don't do english reviews right now for various reasons...

[vanished]

10 minutes ago, Stefan Payne said:

[snip]

It seems you're making two points that should be separate in my opinion.  Correct me if I've misinterpreted this though:

• PSU lists in general are inherently wrong / a bad idea
• There's some problems with the ranking of a few PSUs in this one

As for the latter, I think we would be all smart to listen to any corrections or new information that you or anyone can bring to the table since the whole point of this list is to make good recommendations, and if it's not doing that, it's doing more harm than good.

 

As for the former, I have to disagree.  If you clearly indicate the factors that you do and do not consider when making one, it's simple enough to sort a collection of them on overall quality/safety, for example, and is a great way for people to easily see how they all stack up against each other.  That is, assuming they've been sorted correctly.

[vanished]

23 minutes ago, Stefan Payne said:

And the EVGA GS550 I've got, blew due to lack of protection while overloading the minor rails!

That's the point of having Protections in this example OCP: To protect at least the PSU from damaging itself!

This is of particular an immediate concern, especially considering the review here that a lot of people will base their decision on:

Quote

Performance (40% of the final score) - and so we come to scoring again. With performance much the same as the larger 650W unit, this one didn't offer any surprises. Voltage regulation was strictly average on this unit, which is two notches down from excellent. That'll be a full one point deduction. To this, I will pull another half point for the 12V ripple. Meantime, efficiency was a pass for Gold on both times, so no deductions there. Since I didn't see anything else to worry about, this category gets an 8.5. Same as big brother.

(http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story6&reid=438)

 

He makes no mention of this problem you allege it has.  Therefore, one of you has to be wrong

Do you have links to reviews where they did discover this shortcoming?  That would be very interesting (And troubling )

[quan289]

11 hours ago, Aniallation said:

you really not nothing better to do lol

When a product that may be misrepresented resulting in potential lost in sales, it's to be expected that somebody will eventually comment negatively on this type of list, if they play a direct role within the PSU industry like Stefan had at one point in time (not sure if that's the case now, but he is in contact with some actual engineer in the industry). I'm pretty sure this isn't the first time those who is part (or had been part) of the industry who took some of their time commented on such list. Although Stefan's manner of speaking can be a bit off putting (some of it may have to due with language barrier)...

11 hours ago, Aniallation said:

Really though, stop being so anal about every tiny detail. It's not like we are suggesting stuff that will blow up.

This is not a simple White List to suggest stuff that won't blow up. The main purpose of this list to condense various of aspects of a PSU into a single ranking number in attempt to show how it stack up against each other. In which, a particular aspect which you considered to be a "tiny detail" may be needed to accurately evaluate a product and truly determine what "tier" it belongs. You can't just simply ignore it.

19 hours ago, djdwosk97 said:

There is more to the quality of a PSU than whether or not it's group regulated or not

Well, whether or not it is group-regulated does play a large role in the performance metric of the power supply (which performance is a factor in this tier list), as it can affect efficiency, load regulation, line regulation, crossload, and its response to transient loads to some degree. So seeing certain DC-DC / Independent Regulated PSUs being placed below a budget group-regulated design seem very off, especially considering how an unit such as the FSP Aurum Pro that was made to target the high-end segment that is often recommended along side with the Seasonic KM2 design at the time to be placed below a 5+ year old, budget platform.

 

Maybe considering doing sub-tiers to lessen the potential confliction between units? I don't know.

 

1 hour ago, Ryan_Vickers said:

This is of particular an immediate concern, especially considering the review here that a lot of people will base their decision on:

(http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story6&reid=438)

 

He makes no mention of this problem you allege it has.  Therefore, one of you has to be wrong

Do you have links to reviews where they did discover this shortcoming?  That would be very interesting (And troubling )

Oklahoma Wolf / Jeremy doesn't do overload test. He test it within the limits that manufacturer had advertised. Aris of TPU / THW had only just started to test some of the protections recently.

 

I can't find the review, but Philipus II (a retired reviewer that's also also working within this industry) had stated that his sample had blow up. http://www.jonnyguru.com/forums/showpost.php?p=117340&postcount=13

 

As for a Gold Seasonic unit blowing / burning out, there is a review of the Seasonic S12G 550w along with the G series and Antec TP 550-C that burned. 

http://www.hardwareinsights.com/wp/seasonic-s12g-550-w-review/6/#Load-testing

 

"I have just witnessed the usual way of diode rectifier terminal overheat and shorting. Once again the SBR10U45 rectifier in Seasonic’s semi-synchronous rectification burned. I have experimented with this subject in the Antec TP-550C I reviewed at the end of last year, which also burned (together with SSR-550RM). Using STPS15L60CB rectifiers instead of SBR10U45, the unit was finally able to turn off gracefully under multiple overload scenarios. The OPP is clearly set higher than the SBR10U45 can handle. Seasonic is working to address this issue, the solution is already set, however, I cannot comment more on this subject before it is launched in next batch of products."

 

Pavel's sample of a newer revision of the S12G 550w did show it no longer burning out and the OPP working now; however, it did exhibit high / out of specs ripples. http://www.hardwareinsights.com/wp/new-prototype-seasonic-s12g-550-w-ssr-550rt/

[Stefan Payne]

1 hour ago, Ryan_Vickers said:

It seems you're making two points that should be separate in my opinion.  Correct me if I've misinterpreted this though:

• PSU lists in general are inherently wrong / a bad idea
• There's some problems with the ranking of a few PSUs in this one

As for the latter, I think we would be all smart to listen to any corrections or new information that you or anyone can bring to the table since the whole point of this list is to make good recommendations, and if it's not doing that, it's doing more harm than good.

 

As for the former, I have to disagree.  If you clearly indicate the factors that you do and do not consider when making one, it's simple enough to sort a collection of them on overall quality/safety, for example, and is a great way for people to easily see how they all stack up against each other.  That is, assuming they've been sorted correctly.

 

Quote

 

PSU lists in general are inherently wrong / a bad idea

 

No, if you do it right and categorize them by price and Watt, it's fine, like it's done here:

https://www.hardwareluxx.de/community/f238/die-hardwareluxx-netzteil-kaufberatung-bitte-startpost-lesen-797841.html

http://extreme.pcgameshardware.de/netzteile-und-gehaeuse/390781-empfehlenswerte-netzteile-2016-liste-mit-daten-und-preisen-stand-dezember-2016-a.html

Especially the PCGH list is a great example of how you do it right! 

 

Quote

There's some problems with the ranking of a few PSUs in this one

No, everything is the problem. Beginning it being a copy of the THG List wich Jon Gerow (aka jonnyGURU) reacted to extremely harshely:

http://www.jonnyguru.com/forums/showthread.php?t=13912&highlight=bannable

 

 

The main fact is that you just can not ever put an 80plus Bronze unit witch still is rather on the budget side into the same category as a high end Titanium certified unit as the efficiency is a very high factor in the price of that unit. We have here a ~72€ unit with no cable managment or a rather basic partial modular one against a 190€ fully modular one. WIth all the difference it comes with...

 

And if I want to buy a new PC, how do I know what to buy? This List doesn't help and does more harm than good...

 

In the end, all it takes is:
a) categorize them by watt

b) remove _ALL_ Group regulated unit (except for the under 400W range, for 'Office' Rigs

c) remove _ALL_ undesirable units

 

And you have something decent. And I doubt that anyone ever cares about what to NOT buy!

That's not what they want to know. They want to know what to buy...

 

As for DC/DC or Independently Regulated units vs. Group Regulated:
With a DC/DC Design, you put one voltage inside the transformer and one voltage comes out.

With a Group Regulated design, you put one voltage in and a Gropu of voltages (In this Case +5V and +12V) comes out. Well, that's perfectly fine if you know what to expect but exactly THAT is the problem:
You can not know how your unit is used and how much current the rails have to deliver.

We have here some rather huge extremes: Almost nothing of the +12V and everythhing that consumes Power (like the CPU) draws Power from the +5V rail - like it was done in the past up until the Athlon XP boards (most use the +5V rail for CPU, some later ones the +12V one. Mostly the latest nForce 2 or KT600/880) ones. As well as most S478 Boards (there are some that use the +5V rail for the CPU)...

To todays +12V heavy loads with almost no load on +3,3V or +5V - and if no one would make Group Regulated units, the manufacturers can work towards gettin rid of the 3,3V rail...

 

As for how it works:
You have 2 voltages coming out of the transformer. And how do you regulate it, when one of the rails is heavily loaded and the other one isn't really used?
You don't!
And that's the problem. WIth that the not loaded rail raises in voltage, while the heavily loaded drops dramatically - and this is where the ATX specs come ind and where you need the +/- 5% voltage regulation. 

Here you have an example where a unit just barely stays in spec:

http://www.tweakpc.de/hardware/tests/netzteile/bequiet_pure_power_9_cm/s05.php

 

With a DC-DC unit, you just have the +12V rail and the +3V3 and +5V rail is regulated by a buck converter, so it absolutely does NOT matter how you use/ load that unit...

 

1 hour ago, Ryan_Vickers said:

This is of particular an immediate concern, especially considering the review here that a lot of people will base their decision on:

(http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story6&reid=438)

 

He makes no mention of this problem you allege it has.  Therefore, one of you has to be wrong

Do you have links to reviews where they did discover this shortcoming?  That would be very interesting (And troubling )

Yes, because he doesn't test overload conditions because he's afraid of his Equipment.

Only Aris (Techpowerup and Tomshardware) does test it but only started doing it recently. As well as the guys from TweakPC.de.

Besides them No one does test Overload.

And protection are of most importance!

A worst case scenario looks like this:

http://www.overclock.net/t/944707/why-single-rail-is-not-better-than-multi-rail

 

That's a case where the PSU could not detect an error, due to the way it was constructed.

 

And here you have a review of the EVGA GS650:

https://www.techpowerup.com/reviews/EVGA/SuperNOVA_GS_650/4.html

The used HY-510N does NOT provide any OCP channels anywhere. And it's not implemented on the DC-DC modules either -> you don't see no shunt or coil.

[vanished]

6 minutes ago, Stefan Payne said:

Yes, because he doesn't test overload conditions because he's afraid of his Equipment.

Only Aris (Techpowerup and Tomshardware) does test it but only started doing it recently. As well as the guys from TweakPC.de.

Besides them No one does test Overload.

And protection are of most importance!

A worst case scenario looks like this:

http://www.overclock.net/t/944707/why-single-rail-is-not-better-than-multi-rail

 

That's a case where the PSU could not detect an error, due to the way it was constructed.

 

And here you have a review of the EVGA GS650:

https://www.techpowerup.com/reviews/EVGA/SuperNOVA_GS_650/4.html

The used HY-510N does NOT provide any OCP channels anywhere. And it's not implemented on the DC-DC modules either -> you don't see no shunt or coil.

Well that's significant information I would say.  We absolutely should be taking the protection features and other design elements that could affect safety in extreme conditions into consideration @STRMfrmXMN (and others), wouldn't you agree?  And, as an aside, I think review sites should be clearly mentioning what they don't test for obvious reasons.

 

11 minutes ago, Stefan Payne said:

No, if you do it right and categorize them by price and Watt, it's fine, like it's done here:

https://www.hardwareluxx.de/community/f238/die-hardwareluxx-netzteil-kaufberatung-bitte-startpost-lesen-797841.html

http://extreme.pcgameshardware.de/netzteile-und-gehaeuse/390781-empfehlenswerte-netzteile-2016-liste-mit-daten-und-preisen-stand-dezember-2016-a.html

Especially the PCGH list is a great example of how you do it right! 

 

No, everything is the problem. Beginning it being a copy of the THG List wich Jon Gerow (aka jonnyGURU) reacted to extremely harshely:

http://www.jonnyguru.com/forums/showthread.php?t=13912&highlight=bannable

I don't think this is a copy of that list.  It seems to be quite different from what I can tell.

11 minutes ago, Stefan Payne said:

The main fact is that you just can not ever put an 80plus Bronze unit witch still is rather on the budget side into the same category as a high end Titanium certified unit as the efficiency is a very high factor in the price of that unit. We have here a ~72€ unit with no cable managment or a rather basic partial modular one against a 190€ fully modular one. WIth all the difference it comes with...

 

And if I want to buy a new PC, how do I know what to buy? This List doesn't help and does more harm than good...

 

In the end, all it takes is:
a) categorize them by watt

b) remove _ALL_ Group regulated unit (except for the under 400W range, for 'Office' Rigs

c) remove _ALL_ undesirable units

 

And you have something decent. And I doubt that anyone ever cares about what to NOT buy!

That's not what they want to know. They want to know what to buy...

So you'd like to see efficiency, price, features like modularity and other things all factored in as well?  Well, I can see the value in such a list, but that's not what the goal is here.  It's just supposed to be a rating of quality alone, and so long as that is clearly identified, I don't see the issue with that.  As for not showing bad units, I think it's important to have them on here since if we have them in a low tier, people know they've been considered and that they're not very good.  If we don't include them at all, people won't know if they've been looked at or not.

[79wjd]

30 minutes ago, Stefan Payne said:

 

 

 

No, if you do it right and categorize them by price and Watt, it's fine, like it's done here:

 

No, everything is the problem. Beginning it being a copy of the THG List wich Jon Gerow (aka jonnyGURU) reacted to extremely harshely:

http://www.jonnyguru.com/forums/showthread.php?t=13912&highlight=bannable

 

 

The main fact is that you just can not ever put an 80plus Bronze unit witch still is rather on the budget side into the same category as a high end Titanium certified unit as the efficiency is a very high factor in the price of that unit.

 

And you have something decent. And I doubt that anyone ever cares about what to NOT buy!

That's not what they want to know. They want to know what to buy...

You're not understanding....price and wattage is irrelevant in determining whether or not a PSU is good or bad. The purpose of this list is to show how various PSUs compare against other PSUs. It is then your job to determine how much wattage you need and what caliber of PSU you are willing to pay for. 

 

There's also another inherent problem with building a tier list that factors in price -- and that is that prices are constantly fluctuating. One day you'll find a G2 550w selling for $100 and the next day you'll find it at $50. That's a big difference and leads to massive problems with any list that takes pricing into account....unless you're going to have some API constantly adjust tiers based on current pricing. But then you also have to consider that pricing various drastically between different regions. What's priced well in one region could be ridiculously overpriced in another. THE POINT OF THIS LIST IS TO TELL YOU WHAT IT GOOD AND WHAT IS BAD. It's by no means perfect as no one person has physically tested everything about every PSU. It is a work in progress that takes into account what various reviewers have discovered about various units. Of course things will be missed and PSUs will have to be shifted accordingly -- and it has happened before and it will continue to happen. 

 

 

 

There is more to a PSU than its efficiency and wattage. I agree, two identical PSUs -- one that is 80+ Titanium and one that is 80+ Bronze should NOT be in the same tier. However, if the 80+ bronze unit is electrically superior, then I'd be okay with it. I'm not saying that is the case here, nor am I saying there aren't things that need to be fixed about this list.

 

 

In addition to helping people figure out what PSUs are good enough quality for their needs, a lot of people use this list to determine if the PSU they currently have is shit or not. It's also important to include any/all PSUs as that way users will know what PSUs have been considered and which haven't been looked at yet. Let's say you have a mediocre PSU and you want to upgrade to a GTX1070. You need to know whether or not it's a fire hazard and needs to be replaced when you upgrade or whether it's good enough until you can afford to upgrade it. Let's assume the list was perfect, then if I saw my PSU is in tier 5 (and it has sufficient wattage on the 12v rail) then I would be okay with upgrading to the 1070 and then upgrading the PSU at some point in the future when I could afford it. On the other hand, if I saw that I have a tier 7 unit, then I would wait until I could upgrade both. It's important to know how all PSUs compare -- both good and bad ones. 

[Stefan Payne]

I think Compuiterbase.de also started doing overload test...

And it's quite interesting because this is the area where many units still fail (either burn or not working either because of bad implementation or none at all).

With voltage Regulation as well as Ripple & Noise you don't have that much problems these days, maybe because of the wide availabiliuty of low ESR wet electrolytic and ultra low ESR Polymer capacitors...

What's really interesting are things like transient response tests and how the voltage looks with modern (extremely demanding) components like AMD Polaris and nVidia Pascal Graphics cards, where the load goes from next to nothing to 400W a couple of times in one second...

And that's why I'd prefer more capacity over "better quality" capacitors...

 

 

Besides that, you can put it all in 3 Categories, I think:

 

-As cheap as possible, without sacrificing too much (like Xilence Performance A+, be quiet Pure Power 10)

- Best Value for the money, like Bitfenix Whisper M, 

- The Best of the Best, Sir: Like Dark Power Pro P11, Maybe Silverstone Strider Platinum.

 

With Units Like Cougar GX-S or Super Flower Platinum King (gone as fast as they came) I'm not sure where to put them, they are somewhere in between the first and second category. Very good value for money and still good performance.

[STRMfrmXMN]

6 minutes ago, Ryan_Vickers said:

Well that's significant information I would say.  We absolutely should be taking the protection features and other design elements that could affect safety in extreme conditions into consideration @STRMfrmXMN (and others), wouldn't you agree?  And, as an aside, I think review sites should be clearly mentioning what they don't test for obvious reasons.

 

I don't think this is a copy of that list.  It seems to be quite different from what I can tell.

So you'd like to see efficiency, price, features like modularity and other things all factored in as well?  Well, I can see the value in such a list, but that's not what the goal is here.  It's just supposed to be a rating of quality alone, and so long as that is clearly identified, I don't see the issue with that.  As for not showing bad units, I think it's important to have them on here since if we have them in a low tier, people know they've been considered and that they're not very good.  If we don't include them at all, people won't know if they've been looked at or not.

Yeah, a lack of OCP isn't an absolute dealbreaker but it will fail overload tests like discovered above. 

 

Sorry I'm kinda going back and forth between here and not responding - doing computer science right now. Nonetheless, the list I wasn't even aware of. I modified Aniallation's list which is a mod of the PCMR subreddit list. 

 

Modularity, price and efficiency simply won't make their way in here. It doesn't really directly correlate to performance and build quality (although efficiency is a little subjective here).

[Stefan Payne]

6 minutes ago, djdwosk97 said:

You're not understanding....price and wattage is irrelevant in determining whether or not a PSU is good or bad. The purpose of this list is to show how various PSUs compare against other PSUs. It is then your job to determine how much wattage you need and what caliber of PSU you are willing to pay for. [/quote]

I do PSU Reviews myself. ANd the Price is a huge factor as well. Because if a PSU isn't priced competative, it just won't sell.

And people have (way) higher expectations of that PSU in every regard. And also not everyone has unlimited funs. Also some just think about what that thing is worth to them..

So this List absolutely does NOT help if someone wants a dual GPU System but only wants to spend 75€. Or whatever. So you HAVE to do it by price, at least...

 

Because THAT'S what people want to know: What's thje best unit I can get for my money for my system!

They just don't care about some imatinary Tier Lists, they care about what they should buy. And that's what the lists I've linked to focus on: What people should buy.

A Tier List is only interesting for discussions... 

But woever wants to discuss this stuff, does their research themselves and already know what makes a good PSU and therefore don't need no Tier List of any Kind.

 

 

 

6 minutes ago, djdwosk97 said:

 

There's also another inherent problem with building a tier list that factors in price -- and that is that prices are constantly fluctuating. One day you'll find a G2 550w selling for $100 and the next day you'll find it at $50.[/quote]

I doubt that that will ever be the case...

 

Of course there are the ocational sales, but that depends on some internal things, we don't have any Idea bout what's going on.

Like they ordered a container of PSUs and they don't sell as well as it was expected. And the successor to that series is ready to go (for whatever reasons)...

If that happens, it can happen that you sell them cheaply. Or maybe you do it because you've got a bargain deal from the manufacturer but have to order double of what you want. In that case it can also make sense to do a (limited) sale...

 

All in all, the price of a unit stays the same it will be adjusted to external influences of course. Like RMA rate, Price of shipping (the conainer), currency fluctuations and stuff like that...

 

6 minutes ago, djdwosk97 said:

That's a big difference and leads to massive problems with any list that takes pricing into account....unless you're going to have some API constantly adjust tiers based on current pricing. But then you also have to consider that pricing various drastically between different regions. What's priced well in one region could be ridiculously overpriced in another. THE POINT OF THIS LIST IS TO TELL YOU WHAT IT GOOD AND WHAT IS BAD. It's by no means perfect as no one person has physically tested everything about every PSU. It is a work in progress that takes into account what various reviewers have discovered about various units. Of course things will be missed and PSUs will have to be shifted accordingly -- and it has happened before and it will continue to happen. [/quote]

 

No it doesn't.

Because, look at the Links I've linked.

 

You can do this tier stuff and the price stuff. Just the price has to come first.

Like: Units up to 100€, you do your 3 "Tiers": OKish, good value for money, the best of the best...

You just don't have to be that exact with the price, that can be handeled rather loosely. It's only problematic if the price of a unit is kinda on the edge when you do it.

Like you do it for up to 100€,  And said unit costs 95€ on the day you put it on the list but the price got raised, so now it costs 105€ or so. 

Aside from that, it just doesn't matter if a unit costs 75€ or 50€...

The price _HAS_ to be a factor, but not as much as you understood...

 

And also the price makes a huge difference if a unit really is good or not. Take for example the Cougar GX-S wich is about 70€ right now for the 550W version. For that price it's totally worth buyin and a good unit. But if the price got raised to 80€ (or more) not so much. Because for that money you can get a Bitfenix Whisper M, also 550W, wich is superior to the Cougar in almost every aspect - but that's to expect, is it not?

 

And so with Price, a unit gets to be from worth buyin to not worth at all. Even some things you can overlook if the price is right, if not you'd go for something else either way...

 

6 minutes ago, djdwosk97 said:

There is more to a PSU than its efficiency and wattage. I agree, two identical PSUs -- one that is 80+ Titanium and one that is 80+ Bronze should NOT be in the same tier. However, if the 80+ bronze unit is electrically superior, then I'd be okay with it. I'm not saying that is the case here, nor am I saying there aren't things that need to be fixed about this list.[/quote]

Have you taken a look at the Reviews I've done of a couple of PSUs?

Of course I know that. As mentioned earlier I've worked for a Manufacturer for about 1,5 Years - and learned a couple of things there. And also I do my own reviews...

 

There is just so much to PSU that you just can not ever put them into a 'Tier List', that's just not possible! 
Because every PSU has it's own focus. Do you want/need a PSU with that's only 140mm deep? 

Or do you want a Multi Rail PSU instead? Do you want something as silent as possible or do you want something that's not annoying?

Or do you just want the best of the best, don't care about no price and are willing and able to spend 500 bucks on a PSU?

 

As for Efficiency Levels in general: The gap between 80plus Bronze (and maybe silver) and 80plus Gold and beter is just enormous!

 

You just use completely different topologys, different cicuits for secondary rectification and so much more!

 

On many Bronze units they use just a diode or two for the secondary rectification. And also the old Double Forward things are very present here. As for 80plus Gold and better, you rarely see anything else than an LLC-Resonant mode converter with Synchronous rectification (=they use FETs instead of DIodes) and of course DC-DC things. And many other things that differentiate it.

 

Just looking at efficiency, voltage Regulation and Ripple & Noise is just making it way too easy...

 

Because there are things like Holdup Time. Is that according to spec. Or didn't they care that much about it? 
Like Seasonic Prime vs. Corsair Vengeance mentioned above. THe Seasonic Prime is on the Overkill side and has way longer Hold Up Time than mentioned in the specifications while the Corsair Veangance is on the lower side and probalby at around 12-14s at best. As you can see at the size of the primary capacitors. The 550W (or was it the 500W?) has just 220µF here, where around 330 or 390µF are more in the ballpark...

 

And there is also the stuff about Coil Whine, is a uinit prone to it or not that much...

 

So in the End you agree with me, that Tier Lists aren't a good idee ;-)

 

 

6 minutes ago, djdwosk97 said:

Iaddition to helping people figure out what PSUs are good enough quality for their needs, a lot of people use this list to determine if the PSU they currently have is shit or not. It's also important to include any/all PSUs as that way users will know what PSUs have been considered and which haven't been looked at yet. Let's say you have a mediocre PSU and you want to upgrade to a GTX1070. You need to know whether or not it's a fire hazard and needs to be replaced when you upgrade or whether it's good enough until you can afford to upgrade it. Let's assume the list was perfect, then if I saw my PSU is in tier 5 (and it has sufficient wattage on the 12v rail) then I would be okay with upgrading to the 1070 and then upgrading the PSU at some point in the future when I could afford it. On the other hand, if I saw that I have a tier 7 unit, then I would wait until I could upgrade both. It's important to know how all PSUs compare -- both good and bad ones. 

If that's your oppinion, why don't you post YOUR  suggestion  of how to do it? 

I've already done some examples - and they are way easier than this Tier List is for whoever's doin it! Because all you need to do is pick out the decent and good ones, list them by Watt - and mention if the wattage is recommended in that Series or not...

Like the Dark Power Pro P11, where I'd recomend the 850W over the 750W anytime, because of the changed plattform. THe 850W is the lowest one of the plattform, so you can expect better performance. While the 750W is the lagest of the smaller plattform. And it's pretty normal that there are different plattforms used in one PSU Line...

The Bitfenix Whisper M, where the same is used for 450W all the way to 850W is a rather rare example, it's more common that there is some kind of change.

Like on be quiet Straight Power 10 Series, where AFAIR up to 600W use one Plattform, 700W and 800W use a different one...

 

And if a PSU is a good one can change with the years. A great PSU from 2006 doesn't have to be a good one today. Even a decent one from 2009 can be something you don't want to use today. Especially since there is absolute no way to tell if the PSU is fine or not. All you do is some guesswork. A 10 Year old unit can be still OK today but it also can be defective and you just don't see it until it's too late...

 

And mentioning every PSU that ever existetd is just not a good Idea. 
Or to be blunt: If you want something like that, YOU do it!

It's just way too much work for a single person to do!

And most units aren't reviewed anyway, so you're stuck with guesswork...

And you have the Problem with Changing Plattforms within a series and different specs.

So a 550W P7 is rather useless these days while the 750W could be used.

The value of listing PSUs you can't buy isn't very high, quite the opposite!
It's rather frustrating if you find such a list and half of the units in this list are EOL and you're not able to get them at all (like Super Flower Platinum King wich is rather new (but still EOL)

 

So less is more!

If you want a 'PSU Database' List, that's just not gonna happen, it's best to keep it as compact as possible and not overcomplicate things.

[Stefan Payne]

54 minutes ago, STRMfrmXMN said:

Yeah, a lack of OCP isn't an absolute dealbreaker but it will fail overload tests like discovered above. 

What's gonna happen if the 'OCP' (=Over Current Protection) of your house is not in place and you need it due to some fault for whatever reason?

 

If you guess that your house would catch fire, you're absolutely correct

 

 OCP in PSUs is equal to a breaker in your house. It does the same thing, though the implemetation is a bit different...

[vanished]

56 minutes ago, STRMfrmXMN said:

Yeah, a lack of OCP isn't an absolute dealbreaker but it will fail overload tests like discovered above. 

Well idk about you but I would like to know that my PSU won't let me do anything that would cause harm to it, or, through it, indirectly harm other components, especially if it's in the top two tiers.  For that reason I think looking into the OCP and group regulation issues that have been brought up is definitely in order.

[STRMfrmXMN]

31 minutes ago, Ryan_Vickers said:

Well idk about you but I would like to know that my PSU won't let me do anything that would cause harm to it, or, through it, indirectly harm other components, especially if it's in the top two tiers.  For that reason I think looking into the OCP and group regulation issues that have been brought up is definitely in order.

Thing as, AFAIK, about group regulated units - we don't have any group regulated units in any points above tier 3 because being group regulated does not mean inherently bad, but it can mean that the unit is older, like the S12 and M12 which are soon to be replaced. Also, lack of OCP, although technically the fault of the manufacturer, should not be your biggest worry if your house is wired in such a way that you'd shoot a million volts through the unit, through a surge protector. This can also occur if you plug in a bunch of washers and dryers into a socket and overload it. I'm not suggesting it's a luxury feature because I probably wouldn't buy a super high-end system without those insurances, however. 

 

That said, I'm looking into reviews of the 550W GS and see it doesn't have OTP which is a bit concerning. This is the sort of thing where I wish I could have like a tier 2.5 type of thing. It's still a well-built Seasonic unit with a solid fan, decent electrical performance and a long-enough warranty and I'd hesitate to drop it a tier. 

43 minutes ago, Stefan Payne said:

 

So less is more!

If you want a 'PSU Database' List, that's just not gonna happen, it's best to keep it as compact as possible and not overcomplicate things.

The closest thing to that that we have on the internet is some combination of these two:

https://www.techpowerup.com/reviewdb/PSUs/

http://www.realhardtechx.com/index_archivos/PSUReviewDatabase.html

[vanished]

1 minute ago, STRMfrmXMN said:

Thing as, AFAIK, about group regulated units - we don't have any group regulated units in any points above tier 3 because being group regulated does not mean inherently bad, but it can mean that the unit is older, like the S12 and M12 which are soon to be replaced. Also, lack of OCP, although technically the fault of the manufacturer, should not be your biggest worry if your house is wired in such a way that you'd shoot a million volts through the unit, through a surge protector. This can also occur if you plug in a bunch of washers and dryers into a socket and overload it. I'm not suggesting it's a luxury feature because I probably wouldn't buy a super high-end system without those insurances, however. 

My understanding of what OCP saves you from is, for example, hooking two overclocked Fury X's up to a 500W PSU and running furmark.

[STRMfrmXMN]

12 minutes ago, Ryan_Vickers said:

My understanding of what OCP saves you from is, for example, hooking two overclocked Fury X's up to a 500W PSU and running furmark.

That too, but your house does require OCPDs if it was made in recent years so if your house lacks it, that's what he was talking about before.

[vanished]

1 minute ago, STRMfrmXMN said:

That too, but your house does require OCPDs if it was made in recent years so if your house lacks it, that's what he was talking about before.

Well I'm not concerned about a power spike being fed in, I'm thinking more along the lines of asking too much from it

[79wjd]

@Stefan Payne

I agree that price is an extremely important factor. However, pricing is very inconsistent due to sales and different pricing in different regions. As such, I don't believe a PSU list should factor pricing into the mix. Again, this list isn't saying "Buy PSU X". It's saying, "PSU X is $Y much better than PSU Z", if the price difference is worth the higher quality to you and for your specific needs, then go for it. It's not trying to tell anyone how much wattage they need or whether or not they need a decent PSU or the creme de la creme. 

 

Again, you're assuming people look to this list to know exactly what to buy for their specific needs. That's not the case. The way it ends up working out is people provide their specs/use case and ask what they should buy. Then someone tells them how many watts they need and what caliber PSU they need for their intended build. Then that person looks here and looks at https://pcpartpicker.com/list/ and finds the cheapest PSU that suits their specific needs. THE POINT OF THIS THREAD IS NOT TO TELL PEOPLE WHAT TO BUY. 

 

Sales aren't an infrequent concern factor though. The PSU I would recommend to someone literally changes on a day-by-day basis as some PSU is basically always on sale thus my recommendations will vary constantly. So factoring in price would be impossible unless you plan to constantly keep it up to date (i.e. every day).

 

I'll say it again since you for some reason refuse to listen......THE POINT OF THIS THREAD IS NOT TO TELL PEOPLE WHAT TO BUY. IT IS TO TELL PEOPLE HOW VARIOUS PSUs COMPARE. That is why it doesn't need to factor in price, efficiency, or specific use cases. People will ask about their specific needs and then they can refer to this list to decide which PSU to buy after they've already narrowed it down or been told what PSUs suit their specific needs (e.x. <=140mm length).

 

You're implying that this list is static. It's not. It's constantly changing and it's constantly being updated. There are several people who are constantly suggesting changes and those changes get made once they have been verified to be accurate. 

 

I personally think this is the best way for any tier list for any product to be. I don't like the idea of basing it on pricing as prices fluctuate constantly. Tell me what is good, tell me what is bad, and then I will look at the market around me and decide how much I'm willing to spend at various levels. 

[STRMfrmXMN]

1 minute ago, djdwosk97 said:

@Stefan Payne

I agree that price is an extremely important factor. However, pricing is very inconsistent due to sales and different pricing in different regions. As such, I don't believe a PSU list should factor pricing into the mix. Again, this list isn't saying "Buy PSU X". It's saying, "PSU X is $Y much better than PSU Z", if the price difference is worth the higher quality to you and for your specific needs, then go for it. It's not trying to tell anyone how much wattage they need or whether or not they need a decent PSU or the creme de la creme. 

 

Again, you're assuming people look to this list to know exactly what to buy for their specific needs. That's not the case. The way it ends up working out is people provide their specs/use case and ask what they should buy. Then someone tells them how many watts they need and what caliber PSU they need for their intended build. Then that person looks here and looks at https://pcpartpicker.com/list/ and finds the cheapest PSU that suits their specific needs. THE POINT OF THIS THREAD IS NOT TO TELL PEOPLE WHAT TO BUY. 

 

Sales aren't an infrequent concern factor though. The PSU I would recommend to someone literally changes on a day-by-day basis as some PSU is basically always on sale thus my recommendations will vary constantly. So factoring in price would be impossible unless you plan to constantly keep it up to date (i.e. every day).

 

I'll say it again since you for some reason refuse to listen......THE POINT OF THIS THREAD IS NOT TO TELL PEOPLE WHAT TO BUY. IT IS TO TELL PEOPLE HOW VARIOUS PSUs COMPARE. That is why it doesn't need to factor in price, efficiency, or specific use cases. People will ask about their specific needs and then they can refer to this list to decide which PSU to buy after they've already narrowed it down or been told what PSUs suit their specific needs (e.x. <=140mm length).

Too add to this - usually when someone asks "What PSU should I buy?" I immediately expect to see a specsheet and a location and perhaps a budget, although it can vary depending on the specs I see whether I ask for one or not. If they have, say, an RX480 and i5 6500 I'll say "Well I'd think a tier 3 unit or above would be a solid fit for you, but a tier 4 shouldn't have any struggles with a power-sipping RX480." So I'm not saying "obviously you need a Seasonic Snow Silent" or "I'd recommend this exact PSU" but I'm giving a range that the OP can look through to see if anything peaks their interest.

[quan289]

1 hour ago, Ryan_Vickers said:

My understanding of what OCP saves you from is, for example, hooking two overclocked Fury X's up to a 500W PSU and running furmark.

Well, a OCP can surely prevent people that misjudge their power requirements or plugging too much off of a single cable. It also can help in an event of a short that the SCP did not detect - causing the PSU keep supplying more and more power to the failed component until the component itself or the cables burns / melts.

 

You may already know that the Corsair AXi has the ability to switch between single rail and multi-rail (no OCP vs multiple of OCP). Originally, it was set to single rail mode by default; however, there was reports of a short caused by a pitched wire during installation or a failed component. So Corsair took the initiative to set it to multi-rail mode by default. There's a documented example of this happening at Legit Reviews. A while ago, there was a line of faulty Asus X99 motherboards due to the VRM's implementation where it had went up into smoke. The motherboard would have failed no matter what; however, the fact that it was in single rail mode did exacerbated the issue. You could read about it here: http://www.legitreviews.com/intel-x99-motherboard-goes-up-in-smoke-for-reasons-unknown_150008

 

Another example is these guys at PCGH simulating a short to see how two single rail units and a single multi-rail unit would react. The result was the cables had fried on the two single rail units, while the multi-rail had shut down before the load get too high. http://www.pcgameshardware.de/Panorama-Thema-233992/Videos/PCGH-in-Gefahr-Rauchexplosion-bei-Netzteilkurzschluss-1057938/

[Stefan Payne]

1 hour ago, Ryan_Vickers said:

My understanding of what OCP saves you from is, for example, hooking two overclocked Fury X's up to a 500W PSU and running furmark.

 That's one thing it can do. And that's what got Multi Rail Units the bad name, at least in the english speaking (Forum) world. In Ther German world not so much. Well, I'm partly to blame for that, I think.

 

Anyway: 
There are two parts in what OCP is good for.

One Part is to keep harm from the PSU -> damaging the PSU with a normal load is not possible.

The second part is to keep harm in the computer to a bare minimum. 

 

A very good example of not working Protection is seen here

The one with that problem had a 1600W PSU (Ultra I think). The first problem is the shier powerr of that thing. Even if we expect that unit to have a rather close set up OPP, we are talking about ~2000W power or ~167Amps - that's the area of really good Stick Welder!

And for 170A some cable calculator seems to have calculated something like 35mm² cables...

Anyway: THe Problem here was some kind of short that wasn't short enough so that the PSU could have detected it as a short. WIth at least ~170A and Ohms Law (U= R * I, for this we need U/I = R): 12V/170A = 0,07 Ohms...

SO PSU with this power are in absolute need of multiple +12V Rails to prevent things from burnin...

 

The other thing is that with a [url=https://en.wikipedia.org/wiki/Buck_converter]buck converter[/url],  you have two parts: the high side parts that switches +12V and the low side part. 

If the High Side fails open - nothing bad will happen.

But if the high side FET fails short, you are in a world of hurt. Because everything connected to this rail is dead due to overvoltage...

So a working OCP for the buck converter is an absolute must! Because if you destroy it with a high side short, every component connected to this rail is dead...

[Stefan Payne]

1 hour ago, STRMfrmXMN said:

because being group regulated does not mean inherently bad, but it can mean that the unit is older, like the S12 and M12 which are soon to be replaced.

There are still new group regulated units coming out. Like the Cooler Master Master Power Lite series. 

And yes, it means it's bad because you must not have  0 load condition on any rail because you need a load on the other side for this thing to work. If you don't you have out of spec voltages...

 

 

1 hour ago, STRMfrmXMN said:

 

Also, lack of OCP, although technically the fault of the manufacturer, should not be your biggest worry if your house is wired in such a way that you'd shoot a million volts through the unit, through a surge protector.

Ähm, what are you talking about?! What you said sounds like some kind  of over voltage or surge protection. OCP is over current protection, totally different thing.

And it's the same as a fuse or a breaker in your breaker box. If current is too high, PSUI switches off, thus trying to prevent more damage than already done...

 

 

1 hour ago, STRMfrmXMN said:

 

That said, I'm looking into reviews of the 550W GS and see it doesn't have OTP which is a bit concerning. This is the sort of thing where I wish I could have like a tier 2.5 type of thing. It's still a well-built Seasonic unit with a solid fan, decent electrical performance and a long-enough warranty and I'd hesitate to drop it a tier. 

Why do you mention Seasonic? If it is a Seasonic or not, does it really matter? And that "well built" thing is something I don't really agree...

There are some things that make me wonder. Like the huge 3900µF Nippon Chemicon KZH (or was it an even better KZM??) cap right next to the +5VSB transformer. Or that there are so much SMD components used and next to none heatsinks - with this unit you have a backplate with some thick silicone thermal pads that should transfer the heat...

 

A high quality cap isn't always a good sign! It can also be quite the opposite....

 

Sorry, but the low end Seasonics aren't that great at all and don't really deserve to be mentioned!
Especially since there are a whole bunch of units out there that are better overall - have bigger heatsink, protection chips that use OCP , sometimes even OTP....

It gets funny when you find more than one OTP implementation in one PSU (no I'm NOT kidding!) and three NTCs on the heatsinks  - one for the Primary side OTP (=done via the primary controller like those Infineon CHips), one for the protection IC like the SiTi PS223 and one for the fan controller...

And here in Germany the EVGA GS550 cost ~85€ - the same as the Bitfenix Whisper M...

So it's not good valuie at all. If it would be cheaper, you can think about it...

Besides that it has absolutely no advantage over the Bitfenix.

And that's what I meant with the price. And it's a good example where the price ruins it all - if it would be dirt cheap, it could have been something to think about but it's not it's rather expansive for what you get!

 

1 hour ago, STRMfrmXMN said:

The closest thing to that that we have on the internet is some combination of these two:

https://www.techpowerup.com/reviewdb/PSUs/

http://www.realhardtechx.com/index_archivos/PSUReviewDatabase.html

Yes and there are so much PSUs that hjave no link at all and aren't reviewed for whatever reason...