[EOL] PSU Tier List rev. 14.8

[Sapphiree]

On 2/28/2020 at 5:44 AM, jonnyGURU said:

That is absolute bullshit and the reviewer knows it.
 

He's overloading the PSUs and expecting one of two things to happen.  Either OPP trips before the ripple gets out of spec or is hoping the PSU knows the PSU ripple is out of spec and shuts down.
 

But the fact of the matter is, even a Seasonic Prime or Seasonic Prime based Corsair AX does the exact same thing (which Aris has proven if you want third party proof).  It's a stupid test and the results are stupid.

 

22 hours ago, jonnyGURU said:

He's expecting optimal performance while the PSU is out of spec.

Who in the real world is going to use a 550W PSU for a 600W load?  You wouldn't.  So you would never see these results.
 

It's like a car review driving a car stuck in 1st gear, keeping it at redline for half an hour and then saying the car is shit because the engine blows up.

 

22 hours ago, jonnyGURU said:

To show that I'm not being biased, I pointed out that even a high end Seasonic Prime PSU does the exact same thing when you overload it.  

 

17 hours ago, jonnyGURU said:

It's not uncommon for a PSU under loads beyond it's rated specification for either maximum output or temperature to have an increase in ripple.

Meanwhile, we see completely different story on this tier-list - various psu's drop to lower tiers because they have high OVERLOAD ripple (which is often a measurement error like 268mV 120% load Focus FX on pceva when another reviewer measured max ~30mV on 130% load), so i want to know who is right, JG or author of this list @LukeSavenije ?.

 

[Guest]

25 minutes ago, Kaygooo said:

 

 

 

Meanwhile, we see completely different story on this tier-list - various psu's drop to lower tiers because they have high OVERLOAD ripple (which is often a measurement error like 268mV 120% load Focus FX on pceva when another reviewer measured max ~30mV on 130% load), so i want to know who is right, JG or author of this list @LienusLateTips?.

 

Precisely.

 

And want to know WHY they don't trip OCP or OPP when the PSU gets overloaded?  Because transient load spikes like those caused by graphics cards like Vega 64 can spike a load as high as 800W for about 15ms and cause the PSU to shut down.  But you can't convince the user that they NEED an 800W PSU.  So you relax the OCP/OPP.  15ms spike isn't going to have any amount of damaging ripple.  But if you use an ATE to HOLD a load out of spec for an infinite amount of time, it's going to look bad on paper.

 

But that's the goal, right?  Make an otherwise good PSU look bad on paper so people talk about your review even after it's published.... and it's working. 

 

[Juular]

43 minutes ago, Kaygooo said:

Meanwhile, we see completely different story on this tier-list - various psu's drop to lower tiers because they have high OVERLOAD ripple

1. Not various, it's all the same PSU essentially, Seasonic Focus FX\FM platform.

2. There are difference in performance expectations between 50$ PSU and 100$ PSU.

3. High overload ripple aren't the only issue with Focus FX\FM, the main reason why it were detiered is because it has compatibility issues with high-transient GPUs, with a lot of other alternatives in the market there are no reason to put them on the same tier. This tier list is supposed to be noob friendly way to choose a PSU, guess what happens when noob sees Focus FX at the top tiers and grabs Vega with it ? Right, it doesn't work out well and they blame us for recommending bad PSU.

4. Seasonic acknowledges issues with Focus FX, discontinued it and now manufactures newer revision that supposedly doesn't have such issues. Want a Seasonic PSU ? Get Focus GX\GM.

5. The decision on Corsair CX isn't made yet as far i can tell even if it's really not that important for 50$ PSU to perform good outside of rated power levels.

[LukeSavenije]

on both Focus and CX i'm currently changing nothing until I've gathered some more research

 

with that I'd like to also say that discussion is more than allowed, but don't expect me to react to any of it.

 

I'll be coming back to it on a later notice

[Sapphiree]

@LukeSavenije Focus is just only one example, there are many more models affected by this situation, @jonnyGURU explained everything but if you disagree even with him, well. 

Also if you think that his words are not allowed here because he said truth its strange.

[SexualHealing]

7 hours ago, jonnyGURU said:

Also not true.

 

How it is not true, the temperature is insane and the 230v test makes no sense.The explosion test at 230v, all his tests are 115v, in Brazil  itis 110v with the exception of a non relevant part of the country and it is already presumed to be better at 230v.

 

He never tests a PSU twice unless the manufacturer/brand questions his tests which happens a lot, he even receives treats from PSU manufacturers. My conclusion is that he have read this topic and your comments and tested it again because you represent Corsair.

 

6 hours ago, Juular said:

i can tell even if it's really not that important for 50$ PSU to perform good outside of rated power levels.

 

In my opinion you have a wrong perspective, the people who tends to buy inexpensive PSUs are the ones not well informed or the well informed who is not wealthy and simply can`t pay any more.

[LukeSavenije]

1 minute ago, SexualHealing said:

he even receives treats from PSU manufacturers.

i highly doubt that one... at best they'll ask him to show them the numbers first, maybe not to release it

 

2 minutes ago, SexualHealing said:

The reason to not explode is obvious, he always showed to be grateful to Corsair and even yourself.

a psu isn't grateful to a company... it explodes or it doesn't, and it clearly didn't here

[Juular]

10 hours ago, SexualHealing said:

In my opinion you have a wrong perspective, the people who tends to buy inexpensive PSUs are the ones not well informed or the well informed who is not wealthy and simply can`t pay any more.

Uninformed people are buying Seasonic S12II and EVGA W1\N1\BQ\BR\BT because the brand can't do wrong, right? In this case Corsair CX are better than any of them, don't expect them to stand 125% overload test well either. You barely can find anything at the level of Corsair CX for it's price, maybe CM MWE Gold and bq! SP9\PP10\11.

If you're (people in general) buying Corsair CX or any of aforementioned PSUs and load it past 100% you're doing it wrong. Not that you're doing it right if you overload Seasonic Focus either, but in that case it's at least 100$ supposedly high-end PSU, it's somewhat expected for it to perform perfect under any conditions, a good amount of high-end PSUs do (at least at 110% overload which a lot of reviewers test at, 120-125% are hard to come by).

10 hours ago, SexualHealing said:

How it is not true, the temperature is insane

That's Bronze rated PSU, don't expect it to be cool and quiet, also, if you look at other PSUs, it's typical to have OTP set to 100-170°C (now as far as i'm aware only Aris tests for OTP tho). 120°C you're seeing on CX under 100% load are normal if so says the man that makes PSUs.

[SexualHealing]

22 minutes ago, LukeSavenije said:

i highly doubt that one... at best they'll ask him to show them the numbers first, maybe not to release it

 

a psu isn't grateful to a company... it explodes or it doesn't, and it clearly didn't here

He does receive, i have no time to find it now. But i may do it later.

 

At 230v it does not explode, for sure.

[SexualHealing]

27 minutes ago, Juular said:

Uninformed people are buying Seasonic S12II and EVGA W1\N1\BQ\BR\BT because the brand can't do wrong, right? In this case Corsair CX are better than any of them, don't expect them to stand 125% overload test well either. You barely can find anything at the level of Corsair CX for it's price, maybe CM MWE Gold and bq! SP9\PP9.

If you're (people in general) buying Corsair CX or any of aforementioned PSUs and load it past 100% you're doing it wrong. Not that you're doing it right if you overload Seasonic Focus either, but in that case it's at least 100$ supposedly high-end PSU, it's somewhat expected for it to perform perfect under any conditions, a good amount of high-end PSUs do (at least at 110% overload which a lot of reviewers test at, 120-125% are hard to come by).

That's Bronze rated PSU, don't expect it to be cool and quiet, also, if you look at other PSUs, it's typical to have OTP set to 100-170°C (now as far as i'm aware only Aris tests for OTP tho). 120°C you're seeing on CX under 100% load are normal if so says the man that makes PSUs.

 

I am not familiar with the issues of the PSU models you mentioned, but EVGA is one he caught doing completely different batches to the same model,. 

 

CX is a great PSU except for ripple and temperature. Temperature gets bad enough at 100% load, so it is not good at all. You easily can find better  cheaper, Super Flower King i already mentioned is one example and it is cheaper.

 

[Firewrath9]

12 hours ago, SexualHealing said:

-snip-

If you picked the wrong size PSU, thats on you. Thats like saying a B450 Tomahawk is bad because if you run a 3950X on it with a ton of voltage (200-250W), the VRMs *might* fail.

As for spikes, short spikes won't affect your hardware at all, like @jonnyGURU said before, high OCP is needed because high spikes are common. Like I stated before, high OCP isn't an issue, and neither are bad readings outside of spec.

HDDs can also draw a lot of power. 15 HDDs had a startup spike of 420W, or ~30W per drive. While that may not seem like much, if you have an unreasonable amount of HDDs, if it tripped OCP every time you started up the PC that'd suck.

Spoiler

 

I also don't agree with this list either. High OCP or should not be a reason for bumping a unit down.

[LukeSavenije]

5 minutes ago, Firewrath9 said:

High OCP or should not be a reason for bumping a unit down.

i do see it as a problem (together with the high otp) on g3 for example, where the unit actually failed during testing

 

OCP can be set to 120%, 140%, 500%, i wouldn't care. It's just when it leaves specification i start to worry

[SexualHealing]

That is how a good PSU should behave overloaded, and this one is a cheap Corsair not in production anymore. It turns itself off with 26mv and just a bit higher than 40 degrees. cx430

 

 

5 years ago. This one has nothing to do with current CX.

 

Quote

I'm not saying his RESULTS are bullshit.  I'm saying that his expectations for these outlier tests are unrealistic and should not be a focal point.

 

How is it unrealistic? It used to be realist for the CX itself.

 

Edit: It seems to be at 220v so i am not sure 

[Guest]

1 hour ago, LukeSavenije said:

i do see it as a problem (together with the high otp) on g3 for example, where the unit actually failed during testing

 

OCP can be set to 120%, 140%, 500%, i wouldn't care. It's just when it leaves specification i start to worry

Right.  You always want OCP and OPP to shut down the PSU BEFORE it blows up.   THAT is the point of it.

46 minutes ago, SexualHealing said:

That is how a good PSU should behave overloaded, and this one is a cheap Corsair not in production anymore. It turns itself off with 26mv and just a bit higher than 40 degrees. cx430

But it's not turning off BECAUSE the ripple is getting higher.  It turned off because it is either overheated or overloaded.  There is no such thing as "ORP" (over ripple protection).

 

But in the video, does Rolando state WHY it shut down?  No.  Just that it shut down.

[LukeSavenije]

1 hour ago, jonnyGURU said:

There is no such thing as "ORP" (over ripple protection).

not going to lie... this would be cool

 

but if possible, only with digital ones...

[Guest]

1 minute ago, LukeSavenije said:

not going to lie... this would be cool

 

but if possible, only with digital ones...

With digital, it's easy.

[Juular]

5 hours ago, jonnyGURU said:

With digital, it's easy.

Why it wasn't implemented yet then ? Seems like yet another box to tick for marketing team but it's actually useful. Not to mention that Corsair are practically the only manufacturer that makes digital PSUs to start with, which i can't wrap my head around ...

[Guest]

9 hours ago, Juular said:

Why it wasn't implemented yet then ? 

You're assuming it isn't.

Remember:  There is A LOT of information coming from the PSU that isn't communicated to the end user because the vast majority of end users don't know what to do with the information and cause panic with pitchfork and torches when a reviewer tells them something doesn't seem right.... sort of like what Rolando has done with SexualHealing here in this thread. 

That said, I'm not going to lie, we have considered a next gen digital with the ability to trade off different performance aspects based on real world needs/results.  Like increase efficiency in lieu of higher ripple (still within spec, of course), or increased voltage regulation (still within spec, of course) for lower efficiency.  But to output that information to the customer in a meaningful way is a huge undergoing for Corsair's software/firmware team.  Remember:  For the GUI side, it's the same team for ALL of iCUE.  It's not one team for PSUs, one team for coolers, etc.

[Juular]

4 minutes ago, jonnyGURU said:

You're assuming it isn't.

Well i'm talking about High Ripple Protection specifically, so to say, i.e a digital PSU sees say 100+ mV ripple on any rail, obviously smth have gone wrong (and other protections didn't catch that), it's shutdown time then.

[LukaTCE]

You should add Seasonic core to the list
Also what is better to pick Seasonic core 500w or be Quiet System Power 500W,600W or Corsair CX maybe VS (CX 450 is so close to hit the limit)
Think 500w will be enough for RX 580 or RX 590 or RX 5500 it's 130-180W and 2500K ?

[LukeSavenije]

37 minutes ago, LukaTCE said:

You should add Seasonic core to the list

if you got any reviews for me... last time all i had was internal shots, which still indicate it being a downgraded focus

[Guest]

2 hours ago, Juular said:

Well i'm talking about High Ripple Protection specifically, so to say, i.e a digital PSU sees say 100+ mV ripple on any rail, obviously smth have gone wrong (and other protections didn't catch that), it's shutdown time then.

Again.

 

You're assuming it doesn't. 

 

The MCUs in a digital PCU literally monitors everything.  What it does when it sees something unusual is the result of a couple lines of code. 

[SexualHealing]

On 2/29/2020 at 8:44 PM, jonnyGURU said:

But it's not turning off BECAUSE the ripple is getting higher.  It turned off because it is either overheated or overloaded.  There is no such thing as "ORP" (over ripple protection).

 

But in the video, does Rolando state WHY it shut down?  No.  Just that it shut down.

 

I got what you mean, but in relation to ripple it is shutting down when it should anyway. And it does not seem to be any hard to do? Many PSUs do that.

In that old interview you already say a PSU should work well only up to its rated power, some e-mails Ronaldo receives from manufactiurers states the same. They say he is being unrealistic. Given the fact that all conditions are possible, i don`t think anything is unrealistic unless you presume the end users are experts. I initially posted the first video `cause the maintainers of this list think differently. They think like Ronald, it has to shut down before anything goes wrong.

 

Intel processors are full of protections, unless you buy the K version at your own risk. Fortunately a CPU can only damage itself.

 

But let`s suppose everyone agrees about up to its rated power, the CX temprature is bad at 550w and there is something else wrong with it. And why you review 110% yourself?

 

 

 

 

 

 

 

 

 

[LukaTCE]

So what would be best to pick ? Yeah I Googled about core and seen it's downgraded Focus. But which Focus ? I seen in this topic some Focus were bad
Seasonic core
Corsair CXM or TXM
Be Quiet System power 9 or pure power is a bit expensive here except if order from Germany which would take a week or 2
Will be 500w enough for RX 580 and 2500K ? Or do I need 600w ?
 

[Tim Serious]

@LukeSavenije 

 

Any update on Antec Neo Eco Zen ?