WHY is Everyone Buying This Power Supply??

[AlexTheGreatish]

The Thermaltake Smart 600W is the most popular power supply on Amazon... but that doesn't mean it is good.

 

Buy an EVGA 450 BR PSU: https://geni.us/qAuje2

Buy a Seasonic Prime TX 650 PSU: https://geni.us/yq4g

Buy a Thermaltake Smart 600W PSU: https://geni.us/ODPQMM

Buy a Corsair SF450 PSU: https://geni.us/ompP

Purchases made through some store links may provide some compensation to Linus Media Group.

 

 

 

[AndreiArgeanu]

Because it's cheap. End of video.

Also don't manufacturers list protections? Like the Thermaltake smart probably doesn't have under voltage protection (UVP).

I'd like to see more PSU testing, especially starting on the low end since those are most likely to fail and it's important to tell a good PSU apart from one that will fail at costs that low.

I'd also be interesting to import power supplies from other countries since there's quite a few brands that are only available in certain markets like nJoy, AQIRYS, Aerocool which completely lack a presence in the US but are fairly popular in other parts of the world.

[GodAtum]

I didn't even notice LTT stopped testing power supplies

[Failure 101]

20 minutes ago, AlexTheGreatish said:

The Thermaltake Smart 600W is the most popular power supply on Amazon... but that doesn't mean it is good.

 

Buy an EVGA 450 BR PSU: https://geni.us/qAuje2

Buy a Seasonic Prime TX 650 PSU: https://geni.us/yq4g

Buy a Thermaltake Smart 600W PSU: https://geni.us/ODPQMM

Buy a Corsair SF450 PSU: https://geni.us/ompP

Purchases made through some store links may provide some compensation to Linus Media Group.

 

 

 

Because Thermaltake is semi well known brand

 

But if doesn't mean it's good value compared the to the smart gx2 or the Bitfenix formula bronze

[da na]

I have the 430w Smart PSU in a small system of mine, it's served me rather well I will say. Going on 5 years.

[1300ny]

I think we might need a little bit of the data you collected to give you guys some crowdsourced feedback.

(like you said in the video: https://youtu.be/-n8N62DeNDU?t=215 )

eg: I have no clue if you did some 220V tests as well. I mean I guess you did but I don't know.

[Somerandomtechyboi]

Welp 0:46 - 0:59 already highlighting why tierlists are kinda dumb hence why i only treat them as a vague reference

 

He even debunks the always go with a reputable brand bullcrap with the evga 400w

 

9/10 an actually good video ngl

 

 

I do suggest testing extreme out of spec ripple and out of spec voltage on actual pc components cause i keep hearing voltage ripple being thrown around like a throwaway joke and why you should pay a shit ton for a pricey unit even though there doesnt really seem like theres any substantial data that suggests your pc components will age faster

 

heck i own a bunch of used <20$ units, half of em are <3$ oem trash, and pretty massive diff in terms of volt regulation, now i simply cannot tell what the benifits of ponying up an extra 30$ on a solid unit like the andyson 1200w px (50$ cause broken used market in indo) instead of using my 550w dsa ii bronze or 700w tundra 85+ (12v rails = 504w and 648w respectively), so do some sort of debunking video or something see if theres really any diff between cheap and expensive units if they are both visibly working just fine

[8tg]

Because the PSU tier list is pornography for losers with too much free time and most people will never look at it, and then will buy whatever works and is cheap.

The thermaltake smart series are cheap and work.

 

Its arguable that the only important statistic for a power supply is failure rate and the smart series has a good reputation and a low failure rate, therefore theyre good power supplies.

[Senzelian]

36 minutes ago, Somerandomtechyboi said:

Welp 0:46 - 0:59 already highlighting why tierlists are kinda dumb hence why i only treat them as a vague reference

 

4 minutes ago, 8tg said:

Because the PSU tier list is pornography for losers with too much free time

 

When I said this openly a few years ago in reference to the PSU and motherboard tier list I was laughed at.

Now I'm enjoying my small moment of sweet sweet revenge.

[8tg]

8 minutes ago, Senzelian said:

When I said this openly a few years ago in reference to the PSU and motherboard tier list I was laughed at.

Now I'm enjoying my small moment of sweet sweet revenge.

I say this to those nerds all the time, PSU fetishists are the absolute worst because they jump into any build with some immediate nonsense that nobody could ever care about about the choice of PSU.

"oh dont buy that housefire psu, [insert psu model here] is low tier on the list, you should get [basically spec equivalent PSU in every single way] psu instead beacuse it wont kill you!"

 

Concept title for a video documenting this forum:

"PSU loser tries not to replace the PSU in any pcpartpicker list posted on this site challenge (impossible)"

 

Edit: years ago a literally built a PC for the sole purpose of pissing those people off because i hate them, behold:

Spoiler

This is an i7 7700T and RX 560

I got this because the PSU i chose originally, a seasonic focus GX or something, wasnt good enough in the eyes of these losers and they spent an entire thread arguing about the PSU choice, so i bought a corsair RM1000x.

 

 

[Somerandomtechyboi]

14 minutes ago, Senzelian said:

When I said this openly a few years ago in reference to the PSU and motherboard tier list I was laughed at.

Now I'm enjoying my small moment of sweet sweet revenge.

Welp it is time for the intellectuals to rise aka ppl that can actually think for themselves

 

but yea tierlists only good as a vague reference, once i get some money ill problably do some testing for both researching and debunking, maybe make a list that includes actual hardware so ppl have a reference as to what these psus can handle, i mean what good is an a tier if it cant handle a 4090? And why should i upgrade my psu that is currently working just fine with no visible issues?

 

thing is you really have to think for yourself and use some intuition for stuff like psu choices and overclocking, dont be too much of an npc to the point that you swap out your perfectly working psu for another one with no real reason other than "ThIs PsU iS 5 yEaRs OlD aNd Im SuRe ThAt ItS gOnNa DiE sOoN" or "ThIs PsU iS bAd BeCaUsE a TiErLiSt SaYs So" and whatever other variations while theres literally visible evidence that goes against that (10yo+ psus still working just fine, psu itself has not had any reliability issues), of course ill take the visible evidence instead of non concrete "because i said so" evidence

 

as for overclocking again just needs intuition, say i can push 45nm to 1.6v long term if i had the cooling, that would also mean things like vtt can be pushed to 1.6v because its inside the cpu. Or maybe things like this particular ram ic scales with voltage, if it really scales that well with voltage then its problably got a pretty high tolerance, and whaddaya know ppl have pushed bdie and rev e to 2v+ and they survived even nutcases dailying 1.7v and still no issues.

 

You really cant just take random hearsay with no evidence to back it up otherwise you end up with x58 vtt max safe 1.35v while i casually run 1.45v+ and have encountered no issues (id totally go higher but 3600 uncore on 6x4 dual rank 1900 ddr3 needs >1.6v to be stable, seems kinda uneccesary when 3400 needs alot less) or vtt volt and dram volt 0.5 ∆ otherwise "DaMaGe" while i casually run my 1gbit f die at 2.46v and vtt at 1.48v then prime95 largeffts for 8 hours and no degradation. Obviously this has limits cause when theres proof that degradation or death can occur like with the recent ryzen 7000 vsoc shenanigans where gn literally killed a cpu on video thats solid evidence and should be followed cause it has been proven on video

[Somerandomtechyboi]

32 minutes ago, 8tg said:

Concept title for a video documenting this forum:

"PSU loser tries not to replace the PSU in any pcpartpicker list posted on this site challenge (impossible)"

Actually yk what i think ill make a meme video of this

Imma make a video ideas note and save this

 

Maybe extend it aswell to overspending on mobos ssds or just extend to overly relying on "reputable" brands even if it means absolutely ruining the specs of a build (ex overspending on a noctua cooler or an msi tomahawk board, things i see quite alot)

 

I got a title idea

"reputable brands psu fetishist dude tries not to replace everything in a pcpp list with overpriced reputable brands and overpriced  psus (impossible)

[AAVVIronAlex]

I use a Thermaltake Smart RGB 700W and it has never dissapointed, unless it was the dust clogged in the fan which was making a load of noise.

[Imakuni]

The most important PSU test for me is "can you tank a shjty grid" test, which is to say hold up time for flickers and the ability to survive grid spikes when returing from a power loss.

[dogwitch]

19 minutes ago, Imakuni said:

The most important PSU test for me is "can you tank a shjty grid" test, which is to say hold up time for flickers and the ability to survive grid spikes when returing from a power loss.

the sea sonic and cosair i have can!!!

[sub68]

1 hour ago, 8tg said:

Because the PSU tier list is pornography for losers with too much free time and most people will never look at it, and then will buy whatever works and is cheap.

The nerds on this fourm go all out. There is a reason why I don't care too much of that tier list I just buy the most decent PSU or just dont care as I am running all OEM psu.

21 minutes ago, Imakuni said:

The most important PSU test for me is "can you tank a shjty grid" test, which is to say hold up time for flickers and the ability to survive grid spikes when returing from a power loss.

Thats what a UPS is meant for...

2 hours ago, Failure 101 said:

Because Thermaltake is semi well known brand

 

But if doesn't mean it's good value compared the to the smart gx2 or the Bitfenix formula bronze

I have a computer that was from ~2011 a thermal take lanbox lite a 450 watt thermaltake was dead so I ended up using a different PSU.

[ImorallySourcedElectrons]

5 minutes ago, sub68 said:

The nerds on this fourm go all out. There is a reason why I don't care too much of that tier list I just buy the most decent PSU or just dont care as I am running all OEM psu.

The sad part is that they're often very wrong about what makes a good and correct PSU, and a lot of the YouTubers who test these PSUs often focus on the wrong things. A good example that comes to mind is GamersNexus their focus on the capacitor brand changing between production lots, while in reality, the capacitor shortage on the component market got so bad that we were swapping out capacitors in a similar manner on €100k aerospace PCBs where you got to go through a few weeks of paperwork to document and explain that change. So yeah, I kind of see why Gigabyte did what they did, but they then completely missed the point of the other design flaws in that PSU, but those are far beyond what an average YouTube public will probably understand.

 

Another pet peeve of mine is when they get started on the topic of over current protection, they completely neglect factors like bandwidth, and solely focus on peak current or some other value that's conceptually easier to grasp. Or they come up with stupid rules of thumb that make little to no sense, like saying that you should calculate the total system power and multiply it times 1.5 if you're running an RTX 3080 Ti or 3090. While what you really want to be doing is looking closely at the specifications and seeing if it can handle the power draw on the correct rails. They also often recommend PSUs that are way over-specced for the system they go in, but drawing 300W effective on a 1200W rated PSU is a good way to burn a lot of power and get shitty performance. Because when we (as EEs) design a PSU we usually design it to achieve the desired performance around the expected nominal load, we can significantly shift the point where you get that desired performance. But for a product like this you'd typically want to put it somewhere around 70% I'd imagine.

 

But I'm happy that LTT is going more by the numbers than anyone else so far. That's the way to move forward, and it should also catch companies that are trying to cheap out (e.g., by only regulating a couple of the rails and letting the others float along with the regulated ones). It'll be interesting to see if they find any serious skeletons hanging out in the proverbial closet.

 

44 minutes ago, Imakuni said:

The most important PSU test for me is "can you tank a shjty grid" test, which is to say hold up time for flickers and the ability to survive grid spikes when returing from a power loss.

A shitty power grid usually translates into reduced PSU lifetime, a lot of them will tank something like 300+V for a short time without apparent damage.

 

Quote

"This nest of wires won't exist in the final setup" - Linus

Keep on dreaming

[Imakuni]

26 minutes ago, sub68 said:

Thats what a UPS is meant for...

Ah, yes, paying almost twice minimun monthly wage for a UPS that actually protects my PSU (while also turning power efficiency to shjt thanks to the 127v downgrade) rather than just buying a good PSU that costs... what, a 5th of that to begin with?

 

No sir.

[Imakuni]

4 minutes ago, ImorallySourcedElectrons said:

A shitty power grid usually translates into reduced PSU lifetime, a lot of them will tank something like 300+V for a short time without apparent damage.

Well yes, but a good one will probably last longer.

 

And there's also holdup, as I mentioned. I'm always impressed how my EVGA Gold is able to keep the PC going on a second long blackout, whereas every other PC I have just shuts down.

[sub68]

4 minutes ago, Imakuni said:

Ah, yes, paying almost twice minimun monthly wage for a UPS that actually protects my PSU

You save up? I am being seriouse it took me a solid 4 years to get to the setup I want by saving, scavenging friends stuff.

[ImorallySourcedElectrons]

5 minutes ago, Imakuni said:

Well yes, but a good one will probably last longer.

Not necessarily, exceeding rated parts stress is pretty much a lottery. If you purchase a 10V rated resistor, in reality the manufacturer probably already derated it with a factor of 0.5 to account for quality drift and production tolerances, and odds are you'll be derating the manufacturer listed value by another 0.67 if you're a decent designer, so your resistor normally sees 5 - 6.7 V while in reality it could handle 20V. If you got lucky it'll handle that 20V and then some, if you're unlucky and you got a bad production lot it might already fail at 16V.

[FlyingPotato_is_taken]

Nice

 

 

Where can I apply to work for sugary daddy Linus?

 

For those who don't know:

The R&S is one of the hottest scopes on the market.

The R&S one the left was the hottest thing for a while and still an absolute dream to use.

That thing on the right is a Lecroy with 12 bits@10GS/s (1 GHz bandwidth)!

If there is one scope missing is the Tek MSO-5.

[ImorallySourcedElectrons]

9 minutes ago, FlyingPotato_is_taken said:

Where can I apply to work for sugary daddy Linus?

 

For those who don't know:

The R&S is one of the hottest scopes on the market.

The R&S one the left was the hottest thing for a while and still an absolute dream to use.

That thing on the right is a Lecroy with 12 bits@10GS/s (1 GHz bandwidth)!

If there is one scope missing is the Tek MSO-5.

I'm kind of missing something like a Keysight EXR or 4000X in the line-up though. The R&S scopes have a really nice UI and are pretty good bang for the buck, the Lecroys are nice on paper but they're really not so nice to use for reasons you can't quite put your finger on, and modern Tek just downright sucks. Ever since they got bought up by Danaher/Fortive their software started sucking, they have bad UX design, their instruments are unresponsive, the ethernet interface has trouble with anything except the simplest of networks, the programming APIs are just bad, and software support is non-existent. And then we haven't gotten into the fact that a significant portion of their product portfolio these days consists of rebadged products from Taiwanese brands like ITech, which sell the same gear at a fraction of the price and with the same or better support.

But with regards to the Lecroy, we had a couple of nice high-end ones in the lab at work, but everyone always ended up using the cheapo Keysight 2000 series scopes because they were so much nicer to use if you didn't need the memory depth and bandwidth capabilities of the Lecroy.

[Offron]

I just saw the video, and was wondering if the tests took the difference of 120V and 240V into account?

 

As far im aware it does make a diffrence for efficency if input is either 120V or 240V

If im wrong, please correct me and / or ignore this message :D

[Nafis]

If a PSU dies because of overloading, then the blame should partly go on the user who picked it for the system. If you overload your car and burn the clutch, then the fault is on the driver. Anybody building a PC should have the basic knowledge of matching a power supply rating with his intended load in the setup, and then he/she should just look for the OCP, UVP, SCP options and not solely rely on them to protect you on your bad decisions.