[EOL] PSU Tier List rev. 14.8

[GoldenLag]

24 minutes ago, Mezoxin said:

the latest i have seen was from zardon who stated that the Riotoro enigma g2 shutdown at 107% overload ,also when he tested the focus plus he reported the same ( This shows you he has a consistent flaw when taking measurements when testing for shut in load) , now that is a reviewer who has a pretty expensive lab and decent experience , but not expertise in PSU testing

that seems fairly consistant with pre-fix focus units with their lower OPP. Reviewed during the time period where there were pre-fix focus units around. 

25 minutes ago, Mezoxin said:

and also inconsistent with the real values of the seasonic fx ( all models)  that has  its opp at 137%

thats with the post fix focus units. which raised the OPP, but left the PSU in a state where it rippled on overload. 

 

 

these specific untis are a bit difficult considering they had a fix that you cant tell wether you have the pre-fix or the post-fix model. where one has a 107% OPP and the other a 137%. one shutting down with higher power GPUs and the other having a ripple issue at overload. 

[Mezoxin]

Just now, GoldenLag said:

that seems fairly consistant with pre-fix focus units with their lower OPP. Reviewed during the time period where there were pre-fix focus units around. 

thats with the post fix focus units. which raised the OPP, but left the PSU in a state where it rippled on overload. 

 

 

these specific untis are a bit difficult considering they had a fix that you cant tell wether you have the pre-fix or the post-fix model. where one has a 107% OPP and the other a 137%. one shutting down with higher power GPUs and the other having a ripple issue at overload. 

Nope multiple reviews with the riotoro g2 have stated that it has opp much above that value , you can see the German one I linked and look for others , also both Aris and zardon reviews are within the same time 

you are just speculating that it was the one based on the Previous version of the fx

 

[Juular]

@Mezoxin

So your points are :

 

1) Trust only select reviewers ? Then this tier list would shrink twice because Aris can't possibly test everything. And you can't possibly extrapolate reviews across units with the same platform so with this logic even half of Seasonic Focus or say CWT GPU based units would end-up untiered because they weren't tested by Aris and may perform differently to the base design. You still may tier those units to temporary positions but how that's a better approach than introducing the data from other, not so big reviewers if Aris didn't test them yet ?

2) Disregard multirail ? I don't think it benefits anyone, this is the whole point of this tier list, to show that one unit are better than the other, and multi-rail are one just another protection essentially, you could say disregard OTP in the same way then because who even runs the PSU this hot to trip it on most units ? But i agree that the requirements for tier A+ multirail should be revised, my proposition is to simply require that multi-rail unit should be digital, as in OCP tripping point should be configurable, then tier A+ will consist only of Corsair units but why not ?

 

Frankly, i think Luke and co have already heard us and there's no point in arguing on this over and over again, it's up to them to decide. Until then i really don't see nothing wrong both in Seasonic Focus FX units being detiered because of single review, this matter really brings a lot of attention here for some reason and yet people fail to recognize that Focus FX were phased out by newer revision and no longer relevant anyway. Neither i see wrong with current multi-rail requirements for tier A+\S, if you disregard it then just buy smth from tier A, merge A and A+ tiers and add some select tier A units to tier S (like Seasonic Prime maybe) in your mind. This is newbie friendly tier, newbies know shit about all this and only need basic guidance, this tier list provides that regardless of whether multi-rail units are tiered one tier up from single-rail ones, they're all good.

[GoldenLag]

12 minutes ago, Mezoxin said:

Nope multiple reviews with the riotoro g2 have stated that it has opp much above that value , you can see the German one I linked and look for others

so that one died during OPP testing at 130%. below what is expected for the OPP of post fix focus and pre-fix focus. suspected faulty sample. 

 

and we dont know which focus as both were present during that period. 

16 minutes ago, Mezoxin said:

you are just speculating that it was the one based on the Previous version of the fx

the OPP limits kinda point to that being the case when it shuts down at the level the pre-fix version should. and the other one shuts down at where the post fix version should. 

17 minutes ago, Mezoxin said:

also both Aris and zardon reviews are within the same time 

doesnt mean they received the same product. especially during that period where both units were around. 

 

 

but since there is testing of both sticking around at that period, perhaps moving the G2 down wouldnt be a terrible idea. 

[Sapphiree]

2 hours ago, Juular said:

Can you contribute to discussion about that with smth other than 'it's extremely stupid' ? This isn't your first post here and all of them were the same, you're not helping. ROG Strix were moved (i assume) because there were some review with testing overload ripple which may mean that it's newer revisions are based on Focus GX or there were some modifications that allowed it to perform better than base Focus FX platform.

Everything was explained on previous pages about multi-rail, they don't have any advantage compared to good single-rail psu's and this tier list is confusing.
It's clear that Strix is FX based psu with some additional radiators and different asus fan, for proof strix was released before GX was even announced, you speak without even read it's tests where reviewers many times and clearly said its FX platform (kitguru, ithardware), you can even compare pcb and electronics of FX and Strix by yourselve (they are the same). Sorry, but i'd rather trust experts than random forum members like you. Stop being a hater of seasonic psu's because every time you are funnier.

[Juular]

Just now, Kaygooo said:

Everything was explained on previous pages about multi-rail, they don't have any advantage compared to good single-rail psu's.
It's clear that Strix is FX based psu with some additional radiators and different asus fan, for proof strix was released before GX was even announced, you speak without even read it's tests where reviewers many times and clearly said its FX platform (kitguru, ithardware), you can even compare pcb and electronics of FX and Strix by yourselve (they are the same). Stop being a hater of seasonic psu's because every time you are more funny.

I don't know why Luke has moved ROG Strix up so you're objecting to the wrong person here. And i don't see the point in restating my opinion on multi-rail again, you can see it just in my previous post here.

[Mezoxin]

@GoldenLag no one tested opp for the prefix version of the Seasonic focus plus , do you have a review of that , all we knew about it is that it had shutdown problems with high end graphics cards 

 

strangely enough when Seasonic issued a statement  a while back on its Chinese website it said these shutdowns where due to aggressive OCP , which is odd as normally single rails don’t have 12v ocp , and the new version doesn’t as well , even the prime doesn’t have 12v ocp except for the 1000w variant 

 

@Juular these are not my exact two points , Specially when it comes to multi rail 

[Sapphiree]

2 minutes ago, Juular said:

I don't know why Luke has moved ROG Strix up so you're objecting to the wrong person here. And i don't see the point in restating my opinion on multi-rail again, you can see it just in my previous post here.

So if you dont know why he moved it, stay out of it please, i asked him not you.

[Juular]

2 minutes ago, Mezoxin said:

strangely enough when Seasonic issued a statement  a while back on its Chinese website it said these shutdowns where due to aggressive OCP , which is odd as normally single rails don’t have 12v ocp , and the new version doesn’t as well , even the prime doesn’t have 12v ocp except for the 1000w variant

Well, i can see that they refer to OPP in fact as it's really the same thing apart from different name for modern single-rail PSUs if we're talking about 12V rail.

3 minutes ago, Mezoxin said:

 these are not my exact two points , Specially when it comes to multi rai

Then i may have problems following you, care to repeat your point in short ?

3 minutes ago, Kaygooo said:

So if you dont know why he moved it, stay out of it please, i asked him not you.

Okay Xi Jinping.

[Mezoxin]

 @Juular   In short , trust only credible reviewers not select , and there is no decent psu that isn’t reviewed by Aris btw, he has the biggest database of psu reviews , and I didn’t say disregard multirail , I said with how it’s currently implemented and the actual benefit it gives us , then it shouldn’t be a demand for top tier , That makes the tier rate a BitFenix formula higher than prime titanium  because that is just ridiculous 

[GoldenLag]

19 minutes ago, Mezoxin said:

@GoldenLag no one tested opp for the prefix version of the Seasonic focus plus , do you have a review of that , all we knew about it is that it had shutdown problems with high end graphics cards 

how do you know which is being tested?

 

also checked result for each and the corresponding effiency at a given wattage were the same on both Aris and Zardon. 

 

which would mean that either the equipment of Zardon suddenly stops delivering the correct current at 910w. or they were not testing identical PSUs. 

[Juular]

27 minutes ago, Mezoxin said:

and I didn’t say disregard multirail , I said with how it’s currently implemented and the actual benefit it gives us , then it shouldn’t be a demand for top tier , That makes the tier rate a BitFenix formula higher than prime titanium  because that is just ridiculous 

Again, i agree that with current trend in how it's being implemented it's not really useful unless you can configure it yourself lower with digital units, and that because of that most multi-rail units should be on tier A. But with digital units it's still benefits you so to please both sides i think we need to get rid of tier A+ and make a requirements for tier S to be either digital multi-rail or stand-out in build quality or performance from tier A so both Corsair digital units and Seasonic high-end units (among others) would end-up there.

[Mezoxin]

21 minutes ago, GoldenLag said:

how do you know which is being tested?

 

also checked result for each and the corresponding effiency at a given wattage were the same on both Aris and Zardon. 

 

which would mean that either the equipment of Zardon suddenly stops delivering the correct current at 910w. or they were not testing identical PSUs. 

The methodology for testing shut in power with increased load per step is different than measuring efficiency, I don’t fault the equipment , I suspect he measures incorrectly 

you can see also what I mean by load per step here , and also see how OW got 115% with the 750w variant with no shutdowns back in July 2017 before the new version was ever released , the fix was supposed to be released in Jan 2018 btw after both Aris and zardon reviews 
https://www.jonnyguru.com/blog/2017/07/10/seasonic-focus-plus-750-gold-750w-power-supply/3/

 

 

 

[LukeSavenije]

i got too tired of these comments to do it earlier... so here goes a bit of a rant

 

@Spotty @Mezoxin @Kaygooo @grmaster @Beerus, since you made some comments earlier and wanted me to react directly

 

to list a couple things down

 

of course Aris wouldn't rate in multirail. because as exceptions of bitfenix, cm, be quiet, cooler master and some others, multirail on higher end isn't widely spread all the time. this honestly doesn't matter for anything up to 650, maybe even 750w since i personally set an estimated limit at 60a/rail. so yes, below that limit i wouldn't care less about it. on the other hand, look at something in the regions of 1000-1500w on a single rail. what do you think that can happen? we're talking about a rare case here, but in the case uvp would fail, you'd need a resistance of about under ~0,067 Ohm to be detected as a short. let's say i have a 1500w psu with a 17% overload. this would mean close to 150a going through one cable, possibly causing a fire with it. this has been demonstrated here

 

honestly, i'll get my method of explaining rails in general here. should i just remove the breakers in your house because the chance of something failing is slim anyways? it's still a protection. i'm talking about a lower current level with a psu here, but 50A will melt a 1,5mm², let alone if you'd do more than that, you are able to take a connector with it, killing some of the hardware that's connected to it.

 

the thing is, most people don't need more than 650w, as that's around where a mainstream single gpu system maxes out on (something like a 9900k 2080 ti system), so they'd be totally fine picking something from tier a as their psu. the thing is that tier a+ and tier s are oriented for people that need more, as clearly noted by the tiering, as i won't deny that tier a and a+ perform similarly on paper. tier s just stands out in build quality to me.

 

Quote

strangely enough when Seasonic issued a statement a while back on its Chinese website it said these shutdowns where due to aggressive OCP ,

That was ONE of the Issues, but their Statement, wich you can still get on some websites, included an ASUS GTX970, wich isn't high Power at all. On the Contrary. So they not only have Problems with too low OCP/OPP, they likely have Issues with Transients as well.

 

a psu can still shut down with vega by reserving just 40a for a vega, and this has been shown with a shutdown on the hxi and a stock vega.

 

ocp is set before the PFC section, putting less stress on it and giving the psu potentially more lifespan, as it seems to be one of the first internal parts that dies in modern ATX PSUs from what i can find. 

 

I have contacts with Seasonic, have even done an overview for them with the new Focus GX and you're claiming i have something against this company? i have nothing and everything against every single company. i don't care who made it, who sells it or whatever. what i care about is how the model is, and if applicable what problems it has. a problem at seasonic would count to me as equal to one at CWT or whatever.

 

and you know... ask Aris himself, ask Jon, ask whoever you want. there are two downsides to integrating a proper multirail into a psu

• people being stupid enough to misconnect cables
• cost of the unit

what difference does 10mv vs 40mv make for the lifespan of a part? because i don't think that's going to be a long time... as amazing as some might look in reviews, it's not like it's a super important factor

 

and with only relying on certain reviewers, i really don't want to do that. not only would that pretty much cut the list in half if i were to work together with Aris, it would also be problematic if Aris himself does a misread (which has happened with the pp11 for example), and would make this basically a personal set cyben database, not an independent PSU tier list made by a community, which has so far been my idea to it. I trust Aris just as much as someone like Tweakpc. 

 

but of course I'm not an engineer, i don't have a PHD, i don't have lab equipment. I'm a simple enthousiast, doing his best to make this work. does that mean that there's some subjective parts to it? of course there is, I'm a human

[GoldenLag]

7 minutes ago, Mezoxin said:

you can see also what I mean by load per step here , and also see how OW got 115% with the 750w variant with no shutdowns back in July 2017 before the new version was ever released , the fix was supposed to be released in Jan 2018 btw after both Aris and zardon reviews 
https://www.jonnyguru.com/blog/2017/07/10/seasonic-focus-plus-750-gold-750w-power-supply/3/

he tested up to a load of 750w in that review. no overload. 

 

and OPP was never mention throughout the entire review. (checked with search)

9 minutes ago, Mezoxin said:

The methodology for testing shut in power with increased load per step is different than measuring efficiency, I don’t fault the equipment , I suspect he measures incorrectly 

i checked step by step efficiency for both Aris and Zardon. so unless his testing equipment malfunctions for no good reason, they are not testing the exact same units. 

 

also it was production fix back in 2017. didnt stop units from circulating on the market long after. 

[Mezoxin]

10 minutes ago, GoldenLag said:

he tested up to a load of 750w in that review. no overload. 

 

and OPP was never mention throughout the entire review. (checked with search)

i checked step by step efficiency for both Aris and Zardon. so unless his testing equipment malfunctions for no good reason, they are not testing the exact same units. 

 

also it was production fix back in 2017. didnt stop units from circulating on the market long after. 

He doesn’t test for opp but he tests efficiency till 870w you can see it in the results and doesn’t report shutdowns 

[GoldenLag]

2 minutes ago, Mezoxin said:

He doesn’t test for opp but he tests efficiency till 870w you can see it in the results and doesn’t report shutdowns 

 

he tests 866w input power to the PSU. which then the PSU outputs 755w. 

 

its 5w overload. thats not 115% which you get by deviding 866w by the rated wattage of 750w

[Mezoxin]

49 minutes ago, LukeSavenije said:

 

 i personally set an estimated limit at 60a/rail

 

 

again would you explain why 60 A ?

the same thing would happen ( but needs more time) when you introduce a sustained 30-35A to most pcie cable  with 12v OCP set at 40A , same thing would happen when you overload the single 12V  pin in peripheral ports in the back of the PSU over 8A  , it's just a matter of time to generate enough energy to melt the wire or the pins   j = I ² x R x T , all you need is more time . or do you suspect that 10ms transients  alone above 60A would be sufficient in causing a pcie cable to melt ? because they won't 

 

when  circuit breakers are maintained and  implemented properly in our houses it would never cause the wires to fail catastrophically under any overload scenario they also allow for transients , but sadly the way 12V OCP is currently implemented in PSU's doesn't do the same so we cant fully  make the analogy of not having  Multirail 12V OCP is like not having CB in our houses 

 

I am for the importance of Multirail 12v OCP , but not for the reasons you stated above. 

 

as a side note  I applaud your efforts and IQ for making this algorithm of ranking PSU's , and I think someday you will surpass all of current reviewers 

 

@GoldenLag

thank you are correct there didn't read OW review carefully 

 

 

[LukeSavenije]

4 minutes ago, Mezoxin said:

again would you explain why 60 A ?

because i don't have exact numbers and I have to draw the line somewhere, and chose to do it at 60a here. 

 

19 minutes ago, Mezoxin said:

the same thing would happen ( but needs more time) when you introduce a sustained 30-35A to most pcie cable  with 12v OCP set at 40A , same thing would happen when you overload the single 12V PSU pin in peripheral ports in the back of the PSU over 8A  , it's just a matter of time to generate enough energy to melt the wire or the pins   j = I ² x R x T , all you need is more time . or do you suspect that 10ms transients  alone above 60A would be sufficient in causing a pcie cable to melt ? because they won't 

you're totally right on that, but from what i know it's hard to do anything about that in the first place, but i could be wrong here

 

20 minutes ago, Mezoxin said:

as a side note  I applaud your efforts and IQ for making this algorithm of ranking PSU's , and I think someday you will surpass all of current reviewers 

sarcasm?

[Mezoxin]

8 minutes ago, LukeSavenije said:

 

sarcasm?

Not at all 

[LukeSavenije]

Just now, Mezoxin said:

Not at all 

then thank you

[Visometric]

Hi everybody,

 

Will a XFX TS 750 Gold be enough to safely power the following build? 

 

i9 10920X

Asus TUF Mark 2

RTX 2080 ti

128 GB DDR4 3000 Mhz

 

The PC will be used in a Archviz design studio, so it will get pretty abused over time.

 

I am a little bit concerned regarding the position of the XFX TS Gold series in the new PSU Tier List: Tier B Recommended for the entry level gaming systems with 1 PCIe power connection?? I have purchased this unit a couple of years ago, based on some pretty stellar reviews:

 

https://pcper.com/2016/11/xfx-ts-series-750w-power-supply-review/6/#ftoc-heading-7

https://www.jonnyguru.com/blog/2015/06/10/xfx-ts-650-power-supply/

 

Has this unit become too old for modern components? I know that the TS Gold series does not support Intel's new Modern Standby features but is that such a deal breaker?

[LukeSavenije]

11 minutes ago, Visometric said:

I am a little bit concerned regarding the position of the XFX TS Gold series in the new PSU Tier List: Tier B

it's in the same tier as the platform it's based off, known as Seasonic's s12g

 

12 minutes ago, Visometric said:

I know that the TS Gold series does not support Intel's new Modern Standby features but is that such a deal breaker?

you mean the 2.52's? not currently taken into ranking. it does meet c6/c7 sleep states, if that's what you meant

 

it's not a bad psu, it's just rather outdated. if you're spending this much money and the psu is already 5 years old or so, you might as well, with this kinds of budget

[GoldenLag]

19 minutes ago, Visometric said:

Will a XFX TS 750 Gold be enough to safely power the following build? 

is this a new build? irrelevant to the PSU question, but id probably reconsider the hardware unless you are an avid user of avx 512. 

[GreY2231]

Hello again I have another question about my Xperience perf X 550 watt PSU. For 2 days now whenever I'm playing a game the computer will randomly shut off completely and restart. Even on my failing 750w sharkoon it never did that. I was recommended 550 Watts for my setup is that maybe still too low?

I am running an I5 4690k overclocked to 4.2ghz, Asrock H87 Performance Fatal1ty, 16gb ddr3 1600, GTX 980ti, 3 HDDs, 1 ssd, 3 case fans and a CM 212 hyper led. Half the usb ports are used.