Discussion PSU recommendations and power supply discussion thread - Tom's hardware

[jankerson]

It's only in 450W and only available in select countries so who knows.

 

[Darkbreeze]

Still, even at 450w, if it's an Indy or DC-DC unit it would likely be recommendable on a variety of builds. We have many members from "select" countries.

 

[jankerson]

Still, even at 450w, if it's an Indy or DC-DC unit it would likely be recommendable on a variety of builds. We have many members from "select" countries.

More than likely.

 

[Darkbreeze]

Aris says he doesn't have any information on that unit, which is probably, maybe, just the beginning of a refresh of the S12II line, but that he'll try to dig up some information on it. He says Seasonic doesn't usually send samples out of their budget units, so we will likely to have to wait for somebody worth reading their review to buy one and review it. That might not even happen as the budget units are particularly flashy or glamorous so they tend to get ignored. Which is lame, since affordable units are what the vast majority of users are wanting to know about.

 

[Rexper]

I’d love to see a Protection chip upgrade or independent regulation. Last time I heard, Seasonic didn’t have to the resources to refresh the S12ii.

But we still don’t see OCP, OTP and only 420 12v out of 450w in specs.
Wait, I see a sleeve bearing fan... the old S12ii’s used FDBs.

 

[Darkbreeze]

Well, they must have found the resources, since it seems they've done so or at least, have begun to with one model anyhow. Kind of hard to believe that arguably the most respected PSU manufacturer wouldn't have the resources to do, pretty much anything they wanted.

 

[4745454b]

My guess is it's a 230V only PSU. Minor tweaks to the original platform. As noted by the warranty and fan change the tweaks aren't good or substantial. It will be just like the S12ii. An ok PSU to use, but nothing great or awesome.

 

[jankerson]

My guess is it's a 230V only PSU. Minor tweaks to the original platform. As noted by the warranty and fan change the tweaks aren't good or substantial. It will be just like the S12ii. An ok PSU to use, but nothing great or awesome.

Yeah, likely just trying to sell more units in those select Countries.

Since most of what is available is garbage.

 

[Darkbreeze]

Probably. But it would be nice to know for sure. I can't imagine that they'd improve a series with a better platform and then add a crappier fan and reduce the warranty. I don't know that for certain though. I've seen stranger things happen like Seasonic using those smaller gauge PCI/PEG wires for a while there, on fairly hefty units.

 

[jankerson]

Probably. But it would be nice to know for sure. I can't imagine that they'd improve a series with a better platform and then add a crappier fan and reduce the warranty. I don't know that for certain though. I've seen stranger things happen like Seasonic using those smaller gauge PCI/PEG wires for a while there, on fairly hefty units.

Likely due to keeping price down on the units.

Yeah I noticed that too, but I stick with their flagship units for the most part these days.

2x Primes and an X-Series currently running.

 

[Aeacus]

So, i contacted Seasonic via e-mail regarding S12II EVO series and Seasonic customer support representative said this:

Thank you for contacting Seasonic.

S12II Evo is a limited Series that won't be available in EU and North america but most likely in Rest of the World (ROW). So do not expect it to be around in Europe. Compare to S12II, it offers slighty better compatibility with the last generation of Intel and AMD processors.

As i suspected, the S12II EVO is for Asian/African and the like market. And with reduced manufacture costs, the S12II EVO should be able to compete with the no-name PSUs on those markets, without costing considerably more, while still offering better build quality.

 

[Darkbreeze]

I guess that's not terribly surprising.

 

[turkey3_scratch]

After thorough look of the S12II EVO specs page, i saw that the fan bearing is downgraded to sleeve bearing. Original S12II series uses fluid-dynamic bearing fan. It doesn't look good for S12II EVO series.

Aris in his recent reviews is looking at the fan bearing because many these fans that say FDB aren't a real FDB, it's more common for them to be a rifle or some modified sleeve variant (from what I've read only a couple of fan manufacturers will use a true FDB). Then again, you can't blame anyone because if it says FDB you'd expect it to be one, but I wish that was the case 😛 Regardless, rifle or modified sleeve bearing still beats out a regular sleeve bearing, so I'm in agreement with you all be the reason for the warranty decrease.

As i suspected, the S12II EVO is for Asian/African and the like market. And with reduced manufacture costs, the S12II EVO should be able to compete with the no-name PSUs on those markets, without costing considerably more, while still offering better build quality.

Thanks for finding out!

 

[Aeacus]

Aris in his recent reviews is looking at the fan bearing because many these fans that say FDB aren't a real FDB, it's more common for them to be a rifle or some modified sleeve variant (from what I've read only a couple of fan manufacturers will use a true FDB). Then again, you can't blame anyone because if it says FDB you'd expect it to be one, but I wish that was the case 😛 Regardless, rifle or modified sleeve bearing still beats out a regular sleeve bearing, so I'm in agreement with you all be the reason for the warranty decrease.

Here, i got to ask, which fan bearing for you is the "real/true FDB"? The ones patented by Matsushita (Panasonic)?

Since as far as i'm concerned, any kind of fan bearing design that uses the principle of hydrodynamics can be called:"fluid-dynamic". Those designs also include but aren't limited to: rifle bearing and hydro-dynamic bearing (HDB). Basically any bearing where lubricant flows between the ends of bearing. Sleeve bearing is excluded here since in sleeve bearing, lubricant doesn't flow.

(sorry for off-topic reply)

 

[Darkbreeze]

Fans are relevant, and part of the reason I created this thread was so we COULD talk about things that are related, but not necessarily specifically PSU related. So if you feel like it's worth discussing, it is, and it's cool with me. I think we can relax the off topic worries a bit in here, so long as we keep it somewhat related.

 

[turkey3_scratch]

Aris in his recent reviews is looking at the fan bearing because many these fans that say FDB aren't a real FDB, it's more common for them to be a rifle or some modified sleeve variant (from what I've read only a couple of fan manufacturers will use a true FDB). Then again, you can't blame anyone because if it says FDB you'd expect it to be one, but I wish that was the case 😛 Regardless, rifle or modified sleeve bearing still beats out a regular sleeve bearing, so I'm in agreement with you all be the reason for the warranty decrease.

Here, i got to ask, which fan bearing for you is the "real/true FDB"? The ones patented by Matsushita (Panasonic)?

Since as far as i'm concerned, any kind of fan bearing design that uses the principle of hydrodynamics can be called:"fluid-dynamic". Those designs also include but aren't limited to: rifle bearing and hydro-dynamic bearing (HDB). Basically any bearing where lubricant flows between the ends of bearing. Sleeve bearing is excluded here since in sleeve bearing, lubricant doesn't flow.

(sorry for off-topic reply)

I never looked at it in that way, that's an interesting perspective! I've always, yes, gone by the idea of the Matsuhita patented ones, because technically speaking those are going to be quieter and longer lasting than other bearings except double ball should last longer (but not be quieter). So essentially you are looking at it as fluid dynamic bearing = non-sleeve, and I'm looking at it as fluid dynamic bearing = Matsuhita bearing.

I guess if you capitalize it, Fluid Dynamic Bearing, that'd be different than fluid dynamic bearing??

The more important question I guess is whether people were using the phrase "fluid dynamic bearing" to describe fans before Matsuhita ever patented that. Or did other people start using it after Matsuhita patented that, thereby trying to create the false perception that their bearing is the same thing?? I'm curious about this.

 

[Aeacus]

I never looked at it in that way, that's an interesting perspective! I've always, yes, gone by the idea of the Matsuhita patented ones, because technically speaking those are going to be quieter and longer lasting than other bearings except double ball should last longer (but not be quieter). So essentially you are looking at it as fluid dynamic bearing = non-sleeve, and I'm looking at it as fluid dynamic bearing = Matsuhita bearing.

I guess if you capitalize it, Fluid Dynamic Bearing, that'd be different than fluid dynamic bearing??

The more important question I guess is whether people were using the phrase "fluid dynamic bearing" to describe fans before Matsuhita ever patented that. Or did other people start using it after Matsuhita patented that, thereby trying to create the false perception that their bearing is the same thing?? I'm curious about this.

Capitalizing letters doesn't make it different product, it's just a shorter way to refer to something. E.g CPU = Central Processing Unit. If i were to write it as: "central processing unit", would it be different product compared to the one where first letter in each word is capitalized?

As far as fluid-dynamic bearings go, they were long before in use in HDDs before Matsushita patented their version of fluid-dynamic bearing design. Here's some further reading for you,
link: https://www.hardwaresecrets.com/the-truth-about-fluid-dynamic-bearing-fdb-fans/

Also, in 1966, Philips patented one version of fluid-dynamic bearing design, which is the oldest patent i could find about fluid-dynamic bearings without extensive search,
link: https://patents.google.com/patent/US3467449A/en

Matsushita (Panasonic) patented their fluid-dynamic bearing design in 1992,
link: https://patents.google.com/patent/US5357163A/en

 

[Darkbreeze]

These others ought to just call their versions, FLB, or fluid lubricated bearings. Or, dynamic fluid bearings. Or, just active lubricated bearings.

I waiting for fans with actual onboard miniature electric pump lubricated bearings that have tiny sumps to allow the lubricant to cool. You could probably get another 1000rpm out of a fan like that AND likely reduce the noise as well. Lubricants get hot and thin, too thin in some cases, when they get too hot. When lubricant gets too thin, bearings get louder. But it also takes more power to spin a bearing when the lubricant is cool and thicker. It's a fine line to find the sweet spot. Ok, I'm probably overthinking it now. LOL.

 

[jankerson]

Starting to see a pattern develop as some seem to be getting GTX 1080Ti's and having issues like system shutdowns.

PSU's are the typical suspects like the Corsair CX line ands some other budget line units.

 

[Darkbreeze]

Probably worth paying attention to, and asking about, whether or not the bios has been modified on those cards as well. I've seen a fair number of threads with users who've had some issues, and resolved them by changing back to a standard firmware image from the custom bios mod that was used on the card while it was in a mining rig. Which is where a lot of these cards are coming from.

Obviously, on a new card this won't be the case. I'd like to see a lot of these cards get re-tested to look at power consumption compared to when they were new. I'd guarantee that due to the factors of driver changes and aging many of them are using considerably more power than when they were new and on the original firmware or drivers, yet users always point to those early reviews to indicate that they should only use X amount of power when we know that over time that becomes less and less true.

 

[jankerson]

Probably worth paying attention to, and asking about, whether or not the bios has been modified on those cards as well. I've seen a fair number of threads with users who've had some issues, and resolved them by changing back to a standard firmware image from the custom bios mod that was used on the card while it was in a mining rig. Which is where a lot of these cards are coming from.

Obviously, on a new card this won't be the case. I'd like to see a lot of these cards get re-tested to look at power consumption compared to when they were new. I'd guarantee that due to the factors of driver changes and aging many of them are using considerably more power than when they were new and on the original firmware or drivers, yet users always point to those early reviews to indicate that they should only use X amount of power when we know that over time that becomes less and less true.

Yeah the power draw is usually way off as they tend to use the Founder Edition (Reference) as examples.


With the GTX 1080Ti it's normally been like 50W to 80W off as in too low.

With some of the high end cards like the EVGA GTX 1080Ti FTW3 people can switch (actual switch) to the other BIOS and crank up the voltage to 127% and OC them even higher.... That's some serious power draw.

 

[Darkbreeze]

I've seen 100w differences on the exact same card, a year later. And I don't mean same model. I mean SAME CARD using Kill-A-Watt readings and running the exact same benchmark. Only differences, time, sometimes firmware revision and usually driver version. That's without any overclocking or motherboard considerations thrown in.

 

[jankerson]

I've seen 100w differences on the exact same card, a year later. And I don't mean same model. I mean SAME CARD using Kill-A-Watt readings and running the exact same benchmark. Only differences, time, sometimes firmware revision and usually driver version. That's without any overclocking or motherboard considerations thrown in.

Interesting.

Could explain a lot.

 

[jankerson]

I've seen 100w differences on the exact same card, a year later. And I don't mean same model. I mean SAME CARD using Kill-A-Watt readings and running the exact same benchmark. Only differences, time, sometimes firmware revision and usually driver version. That's without any overclocking or motherboard considerations thrown in.

Starting to look like the CX units are doing what I expected them to do, more of them as time goes on.

Yet people keep recommending them.

 

[Onus]

I think all but the very best PSUs are not really meant to run flat out (e.g. 90%+) for extended periods of time. An occasional all-day gaming session, sure, which is why I like HardOCP's stress test; but not all the time. What you've shown about cards exceeding their expected power draw just makes this worse; people buy the minimum they need, which gets exceeded as their components need more power and capacitor aging cuts down what their PSUs can reliable produce.
For typical sub-100W business systems and/or sub-200W home PCs, a 450W Corsair CX should be fine. For mining rigs, overclocked and run hard 24/7? Hell no.