LukeSavenije

On 8/28/2022 at 10:19 PM, n0pain0gain said:

Why this page isn't supported/updated yet?

for the same reason as Tomshardware being discontinued. We as a team chose to go to our own platform due to some limitations I won't discuss here

For us there's (sadly) no longer a reason to re-format the whole list to fit LTT's coding, though we do offer a creative commons license that would allow the list as is to be re-published here under the credit of the site

Just now, Nystemy said:

5 Volt Standby is indeed a bit weird to not have.

even worse... you're required to test it for ATX 3.0. So they can immediately forget about validating that unless they dedicate the 2x100w to it and use 400w ones for 5v and 3.3v

2 minutes ago, Nystemy said:

The setup isn't hardwired to have all the 12x 400 watt loads strictly for one thing, they are still just 12 individual loads. So nothing prevents them from doing more nuanced tests.

of course, but they can't do a heavy load on both 12v and minor rails on higher wattage units this way

3 minutes ago, Nystemy said:

To be fair, the real limit I see is the AC source only being 3 kVA

can't say you're wrong... and you would actually be able to trigger it with consumer PSUs, as you have to account for double the capacity. ATX 3.0 requires testing of transients up to 200% of the unit's rating

4 minutes ago, Nystemy said:

And in regards to test setup quality, it is debatable. And largely dependent on the tests one desires to do and what features one desires and not.

of course, but considering they're competing with Aris now... they're well behind him as they can't do ATX 3.0 validation, are using real hardware instead of a simulated load for transient testing (which gives you measurement issues of their own)

22 minutes ago, yolosnail said:

This was my concern, the fact that they seem to be basically just doing a Ctrl+C and Ctrl+V of Seasonic's testing means that Seasonic will presumably do very well in the tests, whereas another manufacturer, which tests things slightly differently, may not.

they didn't... they talked to someone at Seasonic to help them at the config... and seemingly both Kyle and them have overlooked things. Chroma is used by many other companies as well, both OEMs (Andyson, CWT, Greatwall and so on) and validation labs (Cybenetics, Intel and so on)

to touch on a few other things mentioned here

is this actually useful?

Yes, if done properly... There are very few PSU reviewers because of the high cost to get into the market. A lot relies on a single person (Aris) right now, so it's always nice to see more people entering the space.

Can it do 115v/230v/whatever?

yes, the AC source allows this to be modified according to their needs

issues I do see here

- where's 5VSB? 

- capacity for 5v and 3.3v is enough for "realistic" tests, but not enough to trip it... I'd dedicate a full 400w load to each rail if you want to check for an OCP limit

- 8x400w will make it hard to test ATX 3.0 compatibility beyond (give or take) 1200w rated units

- you're heavily relying on Chroma's ecosystem this way... which is a choice

it's definitely a better setup compared to say Gamersnexus' Sunmoon, but considering these are around 6k+import/shipping, that should be expected... However, this isn't a groundbreaking thing. We've already seen a better setup than this from Tomshardware/TechPowerUp/Hardwarebusters' Aris...

note: I'm not an electronic engineer by any means, I just have a few friends in the industry so to say

Let's revive this account for a moment again... There's a lot wrong here...

Your voltage regulation numbers aren't deviations from zero, so 3.3v tier E and all of tier F could easily be well out of spec in the right scenarios 

Speaking of... the numbers also seem oddly specific, almost like you're trying to squeeze in or squeeze out certain units

I wouldn't even have specified beyond 18 awg, anything below that is more often than not cheap crap, because they're trying to save a few cents on cables

all the way down, you're somehow still expecting a crap unit to meet ATX spec ripple... and "all situations" would also imply overload beyond 110%, something many units have shown to be unable to handle continued

even most online UPS units won't need 17 MS to react to a power outage... why make holdup time that tight of a requirement?

budget in general is a really weird thing to even consider into the list

Because of all of these requirements and random footnotes that aren't even mentioned in the methodology, as well as having similar, if not the same platform and parts all over the place because of these weird requirements and minor differences in samples... I'm already not a fan

your only activity here is this, with the pfp of Jon Gerow, your team I don't see in here at all

outside of Jon and Aris... you've also talked to me about some units mentioned here, as well as our choice to include ACRF into tier B on Discord

I'm obviously biased towards my own work here... but I can't consider this worthy of much as is... 

ps. I guess my daily SX1000 is low-end now

3 hours ago, Booterbotter said:

@LukeSavenijeany updated psu tier list? 

we moved to cultists for that

1 hour ago, BlazeWingbreaker said:

Does this mean that inline caps are useless or is there a use case scenario where they will help?

not useless per say, they do filter out some of the ripple... it's just that we're talking millivolts, the majority is done internally

try slightly undervolting the GPU and leave it to run a stress test

some Seasonic PSUs have indeed shown issues with 3080, but these feel more like you pulling the power cord than a freeze of the system

3 hours ago, Spotty said:

They mean the new RMx.

https://www.corsair.com/us/en/Categories/Products/Power-Supply-Units/RMx-Series/p/CP-9020199-NA

https://www.tomshardware.com/reviews/corsair-rm750x-2021-power-supply-review

They released a new RM along with the 7000d/x launch which uses the same housing design as RMx 2021, just with the internals of RM 2019

6 hours ago, IIIIIIIIII said:

What is the difference between pre/post 2020 Leadex III?

early revisions had some minor issues with voltage regulation

19 minutes ago, Quackers101 said:

Banned my dog, now he is in heaven... anime heaven.

banned because it's been so long since I've banned someone

On 7/1/2021 at 12:08 AM, jonnyGURU said:

Just as an FYI:  I've tested both the EVGA GT and GA.  GT is "not bad".  It's Gold, has about 5% voltage regulation.  Not as good as RM, but passable.  The GA.... is just terrible.  LOL!  Failed efficiency and load regulation tests.  It sure looks pretty, though. 

 

EDIT:  I will say that I ran the tests at both room temperature and 50°C and it didn't blow up.  So there's that.

 

I wouldn't mind getting the full numbers/report if you have them noted somewhere

2 hours ago, jonnyGURU said:

RM or RMx?

 

RM, no.  It already had the PWM IC that supported burst and the supervisor IC that supported the timings required to support modern standby mode.  We had to do some updates to meet IEC 62368.  It was supposed to be simple, rolling change, but someone decided that if it was going through a dev cycle, update the ID to Corsair's current IDL.

 

RM, since that came out like... days ago

but thanks for confirming my suspicion

@jonnyGURUany changes beyond the housing and more EPS on the new RM refresh?

11 hours ago, Egz said:

I need some help for my build. I'm thinking about to buy

https://www.corsair.com/us/en/Categories/Products/Cases/Mid-Tower-ATX-Cases/iCUE-4000X-RGB-Tempered-Glass-Mid-Tower-ATX-Case/p/CC-9011204-WW

this and inside comes with Corsair CX750F.

Corsair CX750F RGB is a Tier B on that list. My question is what's actual diff between This psu and Tier A others?

Is it really worth to buy XPG Core reactor 750w separately? Thanks.

 

I have 

Intel 11700K 5GHZ

Corsair h110i

Asus Z590 Plus Mobo

RTX 3080

1 M2 SSD

2 HDD

I'd like to refer to the post right above you from Jonnyguru

Just now, jonnyGURU said:

So reviewers don't even come to use that one anymore?  Or they use them, but they haven't been calibrated?

they don't use them anymore, sadly... otherwise I would've asked them if I can borrow it

1 minute ago, jonnyGURU said:

Are you sure about that?  That seems pretty irresponsible.

I'm talking about the ones in Eindhoven they don't actively use anymore, not the ones they actually do use actively in Taiwan

1 minute ago, jonnyGURU said:

Dude.. I see it so often I see the insides switched out, it actually surprised me that it WAS an Andyson when I opened it up.

 

boy that's a concern on it's own... but EVGA is a mess with over 40 models registered at this point

10 hours ago, jonnyGURU said:

So I want some pictures to make a comparison to see if there's bait and switch going on.

wouldn't be the first time... cough B3 and GD

3 hours ago, electropical said:

Would you say it's just as quiet as the Focus (same wattage?), quieter, louder?
I thought you were going to test it with a SunMoon or something (lol), but I guess you don't have access to lab equipment (which is understandable).

Btw, is the whine coming from the fan, or is it coil whine?

as much as I would wish to have a load tester... I don't have 20-120k spare for a good one right now, and CoolerMaster hasn't calibrated their Chroma's for years now, which is the only test lab (relatively) close to me

couple things I should probably note about both

Focus - Rather agressive once you reach higher loads, some batches have motor noise, but my personal sample doesn't. Personally kept it on always spinning because I prefer it

RMx - Similar fancurve to the old RMx, which is imo better at higher loads, but very similar at lower loads. RM850x specifically had it's fan curve updated after Aris' review, so it's hard for me to reference the exact wattage here. 

The whine seems to come from the PSU itself from what I could hear, but it was already the case on RM and I expect it's for the same reason... the burst mode

Then again, for the same price I'd pick up an RMx, if there's a price difference I'd personally just pick up the cheapest. I've used both daily, I've been happy with both daily and both come out pretty well out of review.

oh, and since I didn't say it in my last message, @jonnyGURUdon't be worried about your RMx, it's all safe in my own system and I'm not getting rid of it any time soon 

10 hours ago, electropical said:

Has Luke tested it yet?

10 hours ago, Juular said:

@LukeSavenije

I've had it in my system for a bit now... it's far from loud, it's just that I'm able to make it whine at some idle loads... but I was also able to do that with focus at the time

(for the notice, I have my fans running at a very low RPM at idle... with a normal fan curve I doubt you'd even hear the PSU in the first place)

2 minutes ago, jonnyGURU said:

I don't know anything about it, but I doubt it's Seasonic.

 

How likely would you say something like highpower? Or is there anyone more likely to make something like this?

45 minutes ago, OfficialTechSpace said:

When and where?

48 minutes ago, OfficialTechSpace said:

How? This is based off of testing, not assumption.

48 minutes ago, OfficialTechSpace said:

This is simply wrong. It's NVIDIA's recommendation, not their spec.

I was never talking about Nvidia's own in the first place... I was talking about PCIe

15 hours ago, OfficialTechSpace said:

Corsair for one, doesn't have this issue

Even the head of PSU R&D at Corsair completely disagrees with you

You're completely wrong about wattage, fans and your complete ignorance by not even showing the right fan for the revo df, as it's using an Enermax ED142512M-FA-1 makes you only look less credible. If you're helping someone, at least look at 5 seconds of googling, where users have reported issues, where people like Aris and Igorslab have done testing and so on... It's not poor rail design, it's Nvidia going completely out of spec

7 minutes ago, jonnyGURU said:

Do we know for a fact that it actually has multiple +12V rails?  Remember back in the day when  Seasonic was selling PSUs with "multiple +12V rails" but when you opened up the PSU, all of the +12V wires were soldered to the same place on the PCB?

well, I sadly can't just ask CWT for OEM samples...

On 5/17/2021 at 2:06 AM, jonnyGURU said:

Yes.  Multiple +12V rails don't work with 3090 if the OCP is 40A or lower (and I believe most are).

 

curiousity thing... didn't HX work with it?

20 hours ago, jonnyGURU said:

GPU is not a platform that tends to have multiple +12V rails.

then why does it have multiple +12v rails on CWT's site for both GPU-G and GPU-P? Is this just because Bitfenix originally made the platform that way and CWT just started selling it while giving you an earlier or different variation?