Best PSUs (Archive)

Guest · Mar 17, 2016

Here are the best power supplies for gaming PCs. These PSUs offer the best reliability, performance and protection for your components.

Best PSUs : Read more

kasialorenc · Mar 17, 2016

Archived comments are found here: http://www.tomshardware.com/forum/id-2970385/psus.html

Bartendalot · Mar 18, 2016

>When it comes to power supplies, the rule of thumb that "more is better" doesn't apply any more. Before deciding on your next PSU purchase, you should calculate how much power your system draws from the wall and look for a capacity point that satisfies your demands.

Why exacty? Does it hurt to gives yorself some extra headroom?

Aris_Mp · Mar 18, 2016

if you choose a Platinum or Titanium PSU (even a good Gold one) usually not so much. Higher capacity PSUs have lower efficiency at light loads compared to lower capacity ones.

SviatA · Mar 18, 2016

Thanks for the article! Silent PSU is a great PSU – this is one of the main points. When I was building my current computer, I ordered initially the Cooler Master ATX 450W GX with a 120mm – it is pretty quite. Unfortunately, at Hardware.eu they didn’t have this PSU at that time, so they offered a Chieftec 600W with a similar price instead. And I was foolish when agreed on the deal – it has the same 120mm, but it is very loud:-(. I am thinking about changing it to a fanless Seasonic or Enermax in some time. Or, I will buy a quality PSU with at least 130mm fan.

AdviserKulikov · Mar 18, 2016

As much as I love Tom'sHardware articles, I leave PSU reviews to JonnyGuru.

turkey3_scratch · Mar 18, 2016

Thank you for including the Hydro G 650W. It has become one of my most recommended units. I have no complaints with this list.

Aris_Mp · Mar 18, 2016

As much as I love Tom'sHardware articles, I leave PSU reviews to JonnyGuru.

this is up to you. Personally when I want to learn more on something I read 2-3 different reviews about it in order to gather more opinions.

turkey3_scratch · Mar 18, 2016

Bartendalot :

Define what you mean by "headroom". As per my typical analogy, a power supply is not a squirt gun with power inside. A power supply is charge pump; it uses voltage to create current, which is electric charge moving in a circuit. The main difference in a quality unit with a higher labelled wattage and a quality unit with a lower labelled wattage are the ratings of the internal components and how much current they can handle, the protection thresholds such as when OCP and OPP kick in to shut off the unit. There is no "extra stored wattage" in the higher wattage models because power is not a physical stuff, charge is a stuff.

What matters most is the current and voltage - how well the components inside handle the current, and how stable the voltage is. When those conditions are met, buying a higher-wattage model will not benefit much because the lower-wattage model can easily handle the current and have sufficient voltage stabilization.

turkey3_scratch · Mar 18, 2016

Aris_Mp :

Hey Aris while you're on here I have just one quick question. In the AC loss to PWR_OK tests, it tests how long after a loss of AC the PWR_OK signal is dropped. However, how come normal hold-up time is longer than that test? It may be hard to word what I'm saying, but when the PWR_OK signal is dropped, would not the PSU shut itself off right then? Or would not the motherboard do so? And is PWR_OK the same as the PWR_good signal? How can it continue to create DC after the PWR_OK is dropped?

Aris_Mp · Mar 18, 2016

PWR_OK is in essence the power good signal. The power good signal's hold-up time must be lower than the actual hold-up time, in order for the mainboard to get the signal fast and shut down the system before the rails go out of spec. The minimum allowed difference between pwr_ok's hold-up time and the actual hold-up time should be 1 ms according to the ATX spec.

So the ATX spec states that the power good signal's hold-up time should be >=16ms while the hold-up time should be 17ms at least.

turkey3_scratch · Mar 18, 2016

Ahh thank you makes much more sense!

norseman4 · Mar 18, 2016

>When it comes to power supplies, the rule of thumb that "more is better" doesn't apply any more. Before deciding on your next PSU purchase, you should calculate how much power your system draws from the wall and look for a capacity point that satisfies your demands.

Why exacty? Does it hurt to gives yorself some extra headroom?

All PSUs, no matter the rating have an efficiency curve that until recently had peak efficiency at the 50% mark. Now some PSUs have efficiency curves that trend higher. (Reportedly some have a peak efficiency around 80%)

If you know your maximum power draw, add 10% and go for that. If you need to rely upon a power estimator (such as P.C.P.P,) add more % then round up to the next 50. (At least those are my own rule.)

jtd871 · Mar 18, 2016

So any recommended/approved PSUs below 500W? (Assume I want to power a mainstream (non-K) Skylake, SSD and a GTX970 all on air. That's like 300W. I've even seen non-OC Haswell builds with a 970 running on 250W HDPlex PSUs.)

turkey3_scratch · Mar 18, 2016

jtd871 :

You're assuming, though, that labelled wattage has a meaning to it, when it does not. The ratings of the internal components of that PSU, the thresholds of the protections, and the voltage stabilization all factor in over labelled wattage.

I don't even think HD-Plex makes SMPS units.

JackNaylorPE · Mar 20, 2016

The reason why peeps like to add a bit of headroom would include:

1. A PSU hits peak efficiency at 50% load. So a larger PSU is operating closer to its best efficiency point when under load which means less heat is being produced.

2. A cooling system designed for a larger load will usually have bigger heat sinks / fan and therefore the unit will run quieter. The fan might not even turn on under moderate loads.

3. Even w/ the best capacitors, cap aging is still real.... may not be 20% like the cheap models but still real.

4. Voltage stability and noise increase the closer you get to rated power. If you want that last 0.1 Ghz OC, best to stay away from there.

5. Not leaving room for the addition of a 2nd GFX card or perhaps even a PhysX card is disappointing it ya addition(s) push you up over or even to close to rated power.

leo2kp · Mar 21, 2016

norseman4 :

They trend much higher than 80% efficiency. The Platinum models are well in to the 90s. In fact the SuperNOVA 850 P2 maintains over 90% efficiency on 115v between 200 and 700 watts, peaking at near 92% at just over 300 watts. If you're buying a high-quality PSU today, there is no reason you can't over-provision in the power area for future upgrades or capacitor aging.

norseman4 · Mar 21, 2016

leo2kp :

Agreed, but the peek efficiency is trending higher than 50% now, though it is not required for the 80+ [whatever] ratings. 80+ Platinum requires 90% efficiency over 20% load, though 100% load is 'only' required to be 89%. 50% load requires 92% efficiency

That is a great PSU, btw.
http://media.bestofmicro.com/4/V/557455/gallery/CL_efficiency_w_600.jpg

JohnMD1022 · Mar 22, 2016

1. jonnyguru is a plus.

2. Corsair's flat cables are hard to deal with.

What would I like to see?

Universal modular connectors. This would open a lot of windows, as vendors could offer cables without customizing for each vendor.

I abhor multi-connector SLI power cables, as an example. I never connect more than one device to an SLI cable. But, I must put up with a sloppy cabling job, or construct my own cables.

Suppose I could purchase a fixed length SLI cable with a single connector for any vendor? I could specify 12", 13", 17" in red, blue, yellow, etc. Someone would be making it.

Main MB power connector, or the frequently too short 12 MB cable? Same thing.

Anyone agree?

Anders235 · Mar 22, 2016

These are all solid units but really when you get to Platinum level its more about geeky hair splitting and warranty than anything else.

EETechs · Mar 29, 2016

LOL!! People have no clue. I just looked over the equipment list that Tom's hardware uses and its completely OBLITERATES what Johnny Guru uses by specs, price, and quality. The stuff that Tom's Hardware is using is LAB quality stuff that you find in high end R&D houses that design new electronic products and test them. I think you guys need to learn about test equipment before you go off saying that Johnny Guru does better reviews...When you see: Fluke, Keysight, Rigol, Chroma, Keithley, and Flir, then you know they are not messing around at all!

Johnny Guru still uses a crappy Techtronix TDS2012B. Compare that to the Kesight DSOX3024A used by Tom's Hardware. NO CONTEST. You have to be a complete FOOL to think the that his old Techtronix scope is better! And he relies on the voltage reading from the scope! That Keithly DMM meter used by Tom's has 6.5 digit resolution and it has very low drift over time in order a few ppm with very good DC volts accuracy. In fact, it's is complete over kill for measuring voltage for computer power supplies!!! Oh and costs about $3000.

I'm sorry guys, but I think Tom's Hardware has wiped out the competition when comes to power supply reviews. They pretty much do it like it is done by the professionals who certify the power supplies before releasing them for sale to the public. I don't see other review sites having the funds to even afford all the test gear used here!!! I mean you talking at least 500 top of the line gaming PC's for what was paid for all the gear they use here for testing power supplies!!!

Best PSUs (Archive)

Guest

Guest

Reputable

Distinguished

Contributing Editor

Reputable

Honorable

Expert

Contributing Editor

Expert

Expert

Contributing Editor

Expert

Honorable

Distinguished

Expert

Titan

Distinguished

Honorable

Distinguished

Reputable

Commendable

Share this page