Power Supply Corsair SF600 - 600W SFX PSU

[jeshikat]

Welcome to this review of the highly anticipated Corsair SF600 SFX power supply! The SF600 is a 600W SFX form factor power supply, with Gold-rated efficiency, fully modular cables, all-Japanese capacitors, and a semi-passive 92mm fan. The SF600, along with the SF450, is part of the new SF Series of power supplies from Corsair, hopefully the first of many SFF-orientated units from the company.

Read more here.

 

[Pat-Roner]

Excellent review. It's interesting to read your remarks on the noise levels.

I can barely hear my M1 on full load(6700k, 980ti etc) and I find that my HDD is the noisiest part of my build.
I need to change from my Sharkoon 500w SFX-L and wonder if this or the 700wSFX-L from Silverstone will be best.
My main concern is noise.

Would you say that it was noisier that the strix at load?

Edit:

I wonder if the noise levels could have been improved if the fans where swapped with a Noctua A9x14.

 

[jtd871]

The indications are promising. I'll be adding the 450 to my short list for a likely build in the A4 (provided *that* campaign is successful).

 

[jeshikat]

I need to change from my Sharkoon 500w SFX-L and wonder if this or the 700wSFX-L from Silverstone will be best.

The 700W has a higher efficiency rating, roomier housing, higher wattage, 120mm fan, and is supposed to have a more sophisticated fan controller so in theory it's the better unit. I should be getting a review unit of it so I'll be able to compare them.

Would you say that it was noisier that the strix at load?

I'd say it's a bit louder than the Strix. That was an open-air setup though, I did a brief test with the 970 Strix, Noctua C14S and the SF600 in Cerberus and the PSU runs louder since the back of it is getting warmed by the exhaust from the CPU and GPU. Adding a 120mm fan to the front helped.

I wonder if the noise levels could have been improved if the fans where swapped with a Noctua A9x14.

The main problem is the voltage range, I didn't want to risk damaging the PSU so I'm not sure if the fan header will go all the way up to 12V so that something like the A9x14 would work. You could power it externally but I don't find the A9x14 that quiet at full speed either so I'm not sure if it'll be much of an improvement.

 

[EdZ]

Slightly disappointing to see the same voltage-divider-style fan 'control' as the SST-SX600-G and the SX500-LG. At least the fan doesn't have an unpleasant startup characteristics.

I should be getting a review unit of it so I'll be able to compare them.

!

 

[Dedaciai]

I am currently running a Silverstone SFX-600G in my M1 build (Noctua fans and EVGA 970 FTW) and the loudest thing in my case is the Silverstone PSU by far (occasional whine/click at start up and loud during gaming). Would you say the Corsair SF600 is quieter than the Silverstone?

Thanks in advance!

Dedaciai

 

[jeshikat]

Would you say the Corsair SF600 is quieter than the Silverstone?

At low loads absolutely, at high loads it's quieter but not quiet.

 

[drojasxlt]

Disappointing that they are not including the SFX-to-ATX adapter, as indicated in their initial marketing video. I'm hoping the SilverStone PP08 will do the trick.

 

[jeshikat]

Jon says "...and with the available SFX to ATX adapter plate..." so I guess they'll be selling one separately? It's not on the site yet if they are.

If it has to be purchased separately then the PP08 is the better choice anyway.

 

[Phuncz]

Excellent review ! It's good that there finally is some competition for SFX, hopefully leading to better products from Silverstone.

Why does the ATX 24-pin cable require 28 pins on the PSU's side ? Is it to load-balance and stabilize the CPU and PCIe power lines ?

 

[jeshikat]

Why does the ATX 24-pin cable require 28 pins on the PSU's side ? Is it to load-balance and stabilize the CPU and PCIe power lines ?

They're sense wires so the PSU can adjust the voltage as needed.

The 24-pin of the Type 4 cable set also has a unique feature not implemented in the standard Type 3 cables. There are a total of 28-pins where the cable plugs into the PSU's modular interface, instead of the usual 24-pins. The extra four pins are "sense wires". While a +3.3V sense is part of the ATX standard and is found on most PSUs, it is not often implemented. The RMi not only utilizes the +3.3V sense on the ATX connector, but also utilizes an additional +12V sense and +5V sense. These sense wires read the voltages at the load (the end of the cable) and can increase voltage automatically if voltages drop.

http://www.corsair.com/en-us/blog/2015/july/explanation-of-rmi-new-type-4-cables

 

[jeshikat]

I wonder if the noise levels could have been improved if the fans where swapped with a Noctua A9x14.

I asked Corsair about whether the fan controller goes up to 12V:

We do not recommend swapping out the fan for the simple fact that it’ll break the warranty and we have not set the profile for any other fan. The built in fan curve is designed for that designated fan, so attaching a different one in there will yield an undesired result.

He didn't say it does, but he didn't say it doesn't, so...

Also, the NR092L fan is a rifle bearing fan and was custom made to their spec.

Edit: What do you all think of the testing methodology?

 

[Phuncz]

What do you all think of the testing methodology?

I'm personally a fan of delta temperatures and I would suggest not going beyond the decimal point for temperatures. I'm a fan of keeping it obvious, if something is really different in experience and not within the margin of error.

About the review itself:
I don't mind the lack of a purely scientific approach, since you are lacking the expert equipment right now, but I like what you did that you could do with the gear on-hand. The comparison to the only other true competitor is also important. While most other reviewers would focus on it's non-perfection in regards to 200$ ATX PSUs, I like that you could put it in an SFF-builder's mind-set with most of those questions answered. Like the cable's flex, lack of ATX adapter or the characteristics of the semi-fanless with the Silverstone's weird noises in mind.

The photos are also very clear and I love that they aren't "barrel-distorted". Some are at a very high level of quality (lighting, sharpness, lack of distortion) and are top quality in review land.
Also the number of photos is spot-on in my opinion, without devoting to identifying every single component in there.

 

[drojasxlt]

I'm still new to a lot of the technical testing, but it's definitely fascinating to learn about and relatively easy to understand within the context of this article. I loved the tour of the internals, and the diagram showing cable lengths is going to be great for planning my current build in the Node 202 while I wait for Amazon to ship my SF450. Overall a great read!

I wonder if that internal fan is something Corsair would consider manufacturing as a standalone fan. It would be great to have a relatively high-quality 92mm fan that isn't brown and tan or some crazy neon color.

 

[jeshikat]

I'm personally a fan of delta temperatures and I would suggest not going beyond the decimal point for temperatures.

I'd like to do delta temps in the long run but the little AC window unit in the room where I currently test has trouble keeping up in the summer so if I review a PSU in a few months the ambient temperature will be quite a bit higher than this review.

The photos are also very clear and I love that they aren't "barrel-distorted".

That's the Canon 60mm macro for you. I don't even bother using the Adobe lens distortion correction tool on photos I take with it.

I wonder if that internal fan is something Corsair would consider manufacturing as a standalone fan.

It would have to be a derivative of it, you really don't want a 3900 RPM slim 92mm fan in your computer.

 

[EdZ]

I'd like to do delta temps in the long run but the little AC window unit in the room where I currently test has trouble keeping up in the summer so if I review a PSU in a few months the ambient temperature will be quite a bit higher than this review.

That's the problem that listing delta temperatures solves: rather than (or in addition to) listing the recorded absolute ambient and absolute outlet temperatures, you list the delta between ambient and outlet. That way, even if the ambient temperature you tested at is different review-to-review, the delta temperatures can still be compared to each other.

 

[jeshikat]

I could list the delta temp in addition but I don't like having only the delta temperatures because I find the ambient temp can make the difference on semi-fanless functionality.

 

[Phuncz]

You could surely list both the absolute and delta temperature readings, but with focus on the delta to sort the table in later comparisons.

 

[jeshikat]

Hmm, I may do delta in parentheses or something like that. I think having addition columns would make the table too cluttered.

I'm seriously considering saving up for proper equipment, but I can't afford to do both load testing and noise testing anytime soon. What aspects of power supplies would you all like to have priority for testing?

 

[Phuncz]

Personally I don't care how perfect or clean the line readouts are (oscilloscope and whatnot), just that they stay within spec at the extremes of the load and in between (0 - 50 - 100%). I don't care how awesome the brands of the components are, I'll value a 7 year warranty more than "All Japanese Caps", although this PSU has both. I feel the noise has been the main issue with most SFF users so this should receive a lot of focus.

I'd also like to know what was possible with max load for a sustained amount of time (100% load over 8 hours for instance), what possible power draw would allow a passive function and what components could allow that without fully loading them. But this seems more of the stuff when you somehow don't have anything better to do.

The difference is that most of this is possible without spending money but needs time and experience.