PSU tier list 2.0

[turkey3_scratch]

Okay good. Ahh, I'm going to sleep well tonight.

 

[Darkbreeze]

Me too. There's currently a storm raging here in Colorado, and it's supposed to go on all day tomorrow. That means, all work I had planned to do is cancelled and I may just sleep in and then play TWD tomorrow. Or put up more drywall in my basement. Dammit. Even when it storms I can't get a day off.

 

[Someone Somewhere]

Firstly, watts. A watt is a rate. 1 joule per second is 1 watt. A joule is a measurement of energy. Seconds is time. Yes, power is a rate. It is the rate at which energy is transferred or converted. So by saying "using half watts" or "using half power" what direct translation merits is "using half rate of energy transfer". It does not make sense at all...

I strongly disagree. Speed is also a rate, and we can certainly drive half as fast.

The ampere is also a rate. It is coulombs per second, meaning how much electric charge passes a crosssectional area of a wire per second.

Normally considered a 'point', not a cross-sectional area.

Since charge cannot run out, and how much energy is demanded and received by components is directly related to the speed of this charge, any and every power supply has the full capabilities of pumping the charge to any speed with voltage.

No, or we'd see cheap chinese PSUs without protection outshine stars occasionally. PSUs can only push a certain amount of power per cycle on the transformer, and given they run at a certain frequency that's a flat limit on the power output.

The power plant! So unless they run out of energy, your power supply and hardware will never run out of "power" or more preferably "energy".

Disagree. Switchmode power supplies can reach a point where they simply cannot supply more power. Not due to temperature or input power, but simply due to electrical characteristics. Not being able to push enough electrons per second

What can cause issues is unstable voltage. The voltage from your wall socket alternates (like alternating current, just alternating voltage) and a power supply's primary purpose is to regulate the voltage. Voltage regulation is important! Watts aren't.

Most switchmode supplies don't care about the input voltage or frequency in the slightest, as long as it's at least ~40Hz (or DC), 90-260V.

I'm not much of an expert on the whole process of regulating voltage, but if you think of it it really has to regulate resistance as well. If current increases, that means resistance must decrease to make voltage remain stable. Otherwise voltage would double, which would kill hardware.

Ignore resistance. It's completely useless 90% of the time. Almost nothing in a computer is an ohmic resistor, with the exception of cables, which are generally negligible.

Please quit with the "rates are fake". They're not.

 

[turkey3_scratch]

Firstly, watts. A watt is a rate. 1 joule per second is 1 watt. A joule is a measurement of energy. Seconds is time. Yes, power is a rate. It is the rate at which energy is transferred or converted. So by saying "using half watts" or "using half power" what direct translation merits is "using half rate of energy transfer". It does not make sense at all...

I strongly disagree. Speed is also a rate, and we can certainly drive half as fast.

We wouldn't say "using half speed". Speed would be the rate at which distance changes, so "using half speed" again translates into "using half rate at which distance changes". If the word "using" was cut out it would make sense.

The ampere is also a rate. It is coulombs per second, meaning how much electric charge passes a crosssectional area of a wire per second.

Normally considered a 'point', not a cross-sectional area.

Same thing. More accurate actually. A cross-sectional area refers to a 2D geometric plane, whereas a point is 1 dimensional. Since charge is not all in a straight uniform line, cross-section makes more sense.

The power plant! So unless they run out of energy, your power supply and hardware will never run out of "power" or more preferably "energy".

Disagree. Switchmode power supplies can reach a point where they simply cannot supply more power. Not due to temperature or input power, but simply due to electrical characteristics. Not being able to push enough electrons per second

What would be the purpose of having protection circuitry then? Certainly if power supplies could reach a point where they could run out of power, there would be no need for over power protection because they could simply impose such a limitation on the hardware itself rather than paying money for protection circuits.

Eh, you're probably right. I'm just a confused soul 😛

What can cause issues is unstable voltage. The voltage from your wall socket alternates (like alternating current, just alternating voltage) and a power supply's primary purpose is to regulate the voltage. Voltage regulation is important! Watts aren't.

Most switchmode supplies don't care about the input voltage or frequency in the slightest, as long as it's at least ~40Hz (or DC), 90-260V.

I was referring more to the voltage outputs. They are very important. Whether the voltage is 40Hz or 90-260V, the power supply has to do a lot of modifications to it all. If the GPU was running on alternating 260V, it would be fried.

I'm not much of an expert on the whole process of regulating voltage, but if you think of it it really has to regulate resistance as well. If current increases, that means resistance must decrease to make voltage remain stable. Otherwise voltage would double, which would kill hardware.

Ignore resistance. It's completely useless 90% of the time. Almost nothing in a computer is an ohmic resistor, with the exception of cables, which are generally negligible.

Good to know... but now I'm confused over ohm's law.

Please quit with the "rates are fake". They're not.

I didn't say they were fake (I even ctrl-f to make sure). They are just misused.

 

[bIink]

Hey! i'm planning to build a PC soonish~~ and after looking for some great quality PSU's i have met with a doubt..

Would you prefer a super flower 550w platinum or a 650w gold ? (since they have the same price 115 €)

 

[Darkbreeze]

Depends on what it's for and what hardware you're running. Personally, I think I'd rather have the 650w gold unit. The difference between Platinum and Gold efficiency is minimal, but the 100w difference between the units could be enough to make a difference in whether or not the unit is suitable for use with some hardware.

 

[stavrosmast]

What you just said is like saying: "Someone offered me 20$ and 120$ should I take the 20 or 120? 😀

 

[Darkbreeze]

Not necessarily. If 550w is more than sufficient, based on the current and potentially any future hardware requirements, then going with the Platinum model might be a justifiable decision.

 

[Someone Somewhere]

Firstly, watts. A watt is a rate. 1 joule per second is 1 watt. A joule is a measurement of energy. Seconds is time. Yes, power is a rate. It is the rate at which energy is transferred or converted. So by saying "using half watts" or "using half power" what direct translation merits is "using half rate of energy transfer". It does not make sense at all...

I strongly disagree. Speed is also a rate, and we can certainly drive half as fast.

We wouldn't say "using half speed". Speed would be the rate at which distance changes, so "using half speed" again translates into "using half rate at which distance changes". If the word "using" was cut out it would make sense.

We don't say 'using half distance', either.

The ampere is also a rate. It is coulombs per second, meaning how much electric charge passes a crosssectional area of a wire per second.

Normally considered a 'point', not a cross-sectional area.

Same thing. More accurate actually. A cross-sectional area refers to a 2D geometric plane, whereas a point is 1 dimensional. Since charge is not all in a straight uniform line, cross-section makes more sense.

CSA is more commonly used to talk about wire area (inverse of gauge, basically), for calculating resistance of conductors and sizing cable. Using it in this manner is likely to cause confusion, just as much as if you tried to measure fuel consumption in square milimeters.

The power plant! So unless they run out of energy, your power supply and hardware will never run out of "power" or more preferably "energy".

Disagree. Switchmode power supplies can reach a point where they simply cannot supply more power. Not due to temperature or input power, but simply due to electrical characteristics. Not being able to push enough electrons per second

What would be the purpose of having protection circuitry then? Certainly if power supplies could reach a point where they could run out of power, there would be no need for over power protection because they could simply impose such a limitation on the hardware itself rather than paying money for protection circuits.

Eh, you're probably right. I'm just a confused soul 😛

It's the near-instantaneous limit. Like how there's a limit on how much current a power socket can supply that's a lot more than the faceplate markings or the MCB rating. Usually in the hundreds or thousands of amps; look up fault current.

The difference between this limit and the safe long-term limit is why we have protection devices.

I'm not much of an expert on the whole process of regulating voltage, but if you think of it it really has to regulate resistance as well. If current increases, that means resistance must decrease to make voltage remain stable. Otherwise voltage would double, which would kill hardware.

Ignore resistance. It's completely useless 90% of the time. Almost nothing in a computer is an ohmic resistor, with the exception of cables, which are generally negligible.

Good to know... but now I'm confused over ohm's law.

Ohms law only applies to ohmic resistors. Things like actual physical resistors, lengths of cable etc.

If you try to apply it to say a diode or a MOV, you'll find the calculated resistance changes massively depending on voltage. This applies to normal resistors too, but on a much lower scale.

Please quit with the "rates are fake". They're not.

I didn't say they were fake (I even ctrl-f to make sure). They are just misused.

OK. Please stop calling them misused. They may be misunderstood, but that is very different.

Calling something 'coulombs per second' instead of 'amperes' is a waste of time, given that one is literally defined as the other. You may as well start using '8 billion binary digits' instead of 'GB', or " the duration of 9192631770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the caesium 133 atom" instead of 'second'. See which is easier to say, and less confusing?

 

[bIink]

Depends on what it's for and what hardware you're running. Personally, I think I'd rather have the 650w gold unit. The difference between Platinum and Gold efficiency is minimal, but the 100w difference between the units could be enough to make a difference in whether or not the unit is suitable for use with some hardware.

It would be a gaming pc and the hardware is probably gonna be an i5 6600k, some 350 €~~ gpu (waiting for next gen, pascal and stuff..), asus z170-a motherboard, some ram, hdd and maybe a ssd..


But yeah, what i was thinking is if 100 extra watts is worth or not over a higher efficiency level (from gold to platinum).

EDIT: Also, i don't know if i should worry about this but after reading some reviews of EVGA and Super Flower PSU's i saw that in order to increase efficiency and lower the production cost, they used lower capacity bulk caps that offered inadequate hold-up time, so.. is this really important or not?

 

[Darkbreeze]

What country are you in, and what is your budget?

 

[turkey3_scratch]

Calling something 'coulombs per second' instead of 'amperes' is a waste of time, given that one is literally defined as the other. You may as well start using '8 billion binary digits' instead of 'GB', or " the duration of 9192631770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the caesium 133 atom" instead of 'second'. See which is easier to say, and less confusing?

Misunderstanding is exactly the problem though. It's tough for me to understand stuff when nobody talks about charge and voltage fields. Basically, I don't understand anything unless it's told in terms of chemistry and physics, charge, electric, and magnetic fields. Any other explanation does not make any sense. I mean, I can be taught all the mathematics of circuitry, but that doesn't really teach me anything. This is part of why I'm obsessed with Beaty's writings. They reach out to me and I literally feel like I am exactly like he once was. For instance:

My solution as an engineering student had been typical: dive into mathematics, understand electricity in the form of interconnected equations, but without having any real, visual, gut-level "feel" for the concepts. School was turning my brain into a Spice program, a math simulation. But this didn't help me explain electricity to the public! I couldn't even explain it to myself. So... should I just tell everyone "first learn algebra and calculus, and a bit of Quantum Mechanics, then come back and ask me about Electricity"? No way!

As an experienced adult who was re-examining his childhood misconceptions, I found that it was fairly easy to root out the bad stuff and to construct a sensible view of the "electricity" world. Slowly I came to look at electrical physics and circuitry in a new light, seeing them not as abstractions or just some math models, but instead I learned to see them in a direct, visual, gut-level way. I'd never been able to do this before. I'd been blind for decades. Until finally I learned to "see" again, I didn't realize how poorly I understood this aspect of physics! Yet as a degreed electrical engineer, I was supposed to be an expert...

This is part of what I fear. I will be going to college next year, and I have fear that mu electrical engineering professors will teach me improper stuff. My mind has already been corrupted from my childhood textbooks.

Then during my science museum exhibit design work in 1988 I acquired a stack of elementary school textbooks. We were working on our new Electricity/Electronics exhibit, and I wanted to find out how to explain electricity to 6th-graders. But when reading the books I was totally stunned. The electricity chapters were wrong. Terribly terribly wrong, and it wasn't just simple factual errors. Also, they weren't wrong like saying "atoms are little solar systems" this was different. The authors clearly had no grasp of their subject. The books' electricity chapters were teaching bizarre things. If "electricity" is like a gas, then the books were doing the equivalent of teaching us that wind moves at the speed of light, or that sound waves are the same as air molecules. Or that nitrogen is a kind of invisible energy.

I'm also a student who has discovered great personal flaws, who has gone through a recent traumatic learning experience, has stumbled on some important keys to understanding. And now I want to benefit the other students by telling them what I learned.

Who are the nitpickers? Why, any author who wants to "get it right," who wants to avoid spreading misinformation far and wide. And any student who wants to "cut through all the BS" and clearly see how things really work. Also the entire communities of scientists and engineers, of course. Terrible pedantic nitpickers, they should be ashamed!

I'm just afraid that I'll never be able to learn this stuff properly. In college they're either going to throw a bunch of math which at me doesn't really make one understand the concepts, or they'll teach those concepts improperly. And then, even if they do try to explain those concepts, they'll explain them in a manner in which it assumes you already know the information. Then how will I ever learn? Forums only go so far... So does putting together circuits. Even if I had the full knowledge on how to build a power supply, soldering everything to the PCB, understanding the mathematics, I'll still be unsatisfied if I don't have an understand of exactly what is going on with the charge, voltage fields, and magnetic fields.

Note: The stuff that you say I said is wrong is not stuff Beaty said. The stuff he says is all right though. I think all my struggle results from me being one of not very many who thinks about this stuff. It bugs me. Anyway, I have to thank you for being patient with me.

 

[Darkbreeze]

"All" is highly subjective. "All" of any group of people, even exceptionally adept ones, can't ever agree on everything. Certainly Beaty has some theories that are discounted or at least dubious in the eyes of his peers. Not even Bell, Edison, Tesla or Einstein always got it right.

 

[turkey3_scratch]

True. I'm biased toward him. I shouldn't have said "all" though. Some of his writings are his own theories, but others are facts.

 

[exkhill]

Is silver power sp-ss500 500w 80+ white efficiency any good? In which tier would it go?

 

[Nuckles_56]

My guess is that it would be a tier 3 at best depending on what internal components are in it, most likely a tier 4 unit. It is also a very old unit now, as there is a very short review from this site from 2008, with very little detail on it. I would honestly spend more money and get a newer unit which will perform better and has better components

 

[exkhill]

Thanks. I'll go with one of the EVGA B series that I found, probably.

 

[bIink]

What country are you in, and what is your budget?

Right now i live in Spain and my budget is around 1100 €. So far i'm planning to get:

CPU : Intel i5 6600k - @ 226€

Mobo: Asus Z170-A @ 142 €

GPU : @ 350 € - either an Asus r9 strix 390 or a MSI gtx 970 gaming at this point, but i'll wait for Pascal and Arctic Islands cards since they should be coming soon

HDD : Seagate barracuda 2TB @ 68 €

PSU : Super Flower Leadex Gold 650w @ 115 €

Case : Fractal Design R5 white @ 112 €

Cooler : Cooler Master 212 evo @ 30 €

RAM : Kingston Savage 8GB DDR4 2400MHz @ 40 €

I've never build a pc, but after searching and searching reviews and reading websites it doesn't look that bad. What do you think and would you recommend me anything else?

 

[Someone Somewhere]

If you're looking for more general advice, you should probably post a thread in Systems.

 

[bIink]

All right, thanks!

 

[deramatic351]

Please add HuntKey to tier 5. That psu is really low quality and runs less than it's rated wattage. Sadly I myself have one. 🙁

 

[turkey3_scratch]

I'm hoping eventually we move away from putting a whole company into Tier 5. It'd be better if we just have individual series/units from those companies. A lot of then do have some less-sucky lineups.

 

[Onus]

Huntkey sells a range of PSUs. Yes, Gabriel Torres at HardwareSecrets blew quite a few of them in testing. The new ones, however, passed and were decent.

 

[turkey3_scratch]

Hey I have a question and I hope somebody has an answer (surprise!). So looking at the Corsair SF450 review on Tomshardware, on the hold-up time page, it reads "PWR_OK to inactive DC loss delay". And it says higher is better.

It doesn't make sense to me. The PWR_OK signal should be dropped as soon as a voltage goes out of the ATX spec range, so once the PWR_OK signal is dropped, wouldn't you want your PSU to shut off as fast as possible? Why would you want there to be a longer delay while your components are operating on out-of-spec voltages?

 

[Someone Somewhere]

PWR_OK should drop before the PSU goes out of spec. It's the PSU's way of telling the motherboard that "I'm no longer able to deliver power properly; you should shut down as cleanly as possible". A larger value means more time between PWR_OK dropping, and the voltages going out of spec.

More time from PWR_OK dropping to losing power is good, though some PSUs only drop PWR_OK after going out of spec - so this value would be negative. This is bad, and if you look at the slides, you can see that a couple do this.