Silverstone SX700-LPT 700W Platinum SFX-L

Do you know if it's 2 pin or 3 pin? I mean, does it use the fan tach sensor? Or just send a particular voltage to the fan?

That I don't know. I'd love for it to be calibrated using RPM but it's probably just set to voltage for simplicity.
 
The SX500-LG only has a bad rep here though. Pretty much every reviewer, even SPCR, gave it the thumbs up.

Not to say that the SX500-LG couldn't be improved, but it's good enough for most people.

Uh...please show the SPCR link to their SX500-LG test. I've never seen it reviewed there. They tested the SX600 and made note of the fan clickiness.

They gave the 600 a cautious thumbs up - because it's SFX it gets a 'silent pass':
"In comparison with ATX PSUs of the same power rating, the small SX600-G runs louder at mid & higher power level, though its subjective volume is still moderate; this is the price you pay for for high power in a small package. "

SX600 SPCR Review Link
 
SPCR seems to have a rigorous review PSU schedule - one new review a year. Take that, frenzied Consumer Marketing! ;-)
 
Dangit Silverstone, you have your own perfectly good 120mm slim fan! Stop using that POS now you're not locked into a tiny voltage range!

If I recall correctly, I believe you are confusing two PSUs. The "tiny voltage range" issue is with the Enhance-built SX600-G. The SX500-LG and other HighPower-built SFX-L designs use the full voltage range, or at least further than 3-5V

See http://hardforum.com/showpost.php?p=1041441547&postcount=463

Anyway, which fan are you talking about particularly? Might be worth trying a fan swap and testing noise and temp levels. I could do that, since I did buy a temp gun and tach meter for testing last year.
 
If I recall correctly, I believe you are confusing two PSUs. The "tiny voltage range" issue is with the Enhance-built SX600-G. The SX500-LG and other HighPower-built SFX-L designs use the full voltage range, or at least further than 3-5V

See http://hardforum.com/showpost.php?p=1041441547&postcount=463

Anyway, which fan are you talking about particularly? Might be worth trying a fan swap and testing noise and temp levels. I could do that, since I did buy a temp gun and tach meter for testing last year.
IIRC, the fan itself works over the full votlage range, but Silverstone chose it specifically because it shut off under a certain voltage (that happened to correspond to an acceptable heatsink temperature for a chosen thermistor heatshrinked to it), and was sufficiently fast when that thermistor hit a higher temperature. The effective working range is from like 3v to 5.5/6v in the SX500-LG. It was chosen because it just happened to have the desired fan curve to match the thermistor that Sirfa installed in the PSU, making the fan a drop-in replacement. Silverstone's own slim 120mm was ruled out for having too low a starting voltage to work in semi-fanless mode with that thermistor.
 
Silverstone chose it specifically because it shut off under a certain voltage (that happened to correspond to an acceptable heatsink temperature for a chosen thermistor heatshrinked to it), and was sufficiently fast when that thermistor hit a higher temperature. The effective working range is from like 3v to 5.5/6v in the SX500-LG.

Yes, that's what happens when using a piece of shit passive "controller". It's great to hear that they're actually going to use a real fan controller this time. Hopefully it won't still suck. :D
 
The effective working range is from like 3v to 5.5/6v in the SX500-LG.

6V? That would be around 1100rpm max according to Aiboh's chart.

However, the chart on Silverstone's product page indicates the fan goes up to almost 1500rpm which would be more like 9V according to the chart. Since Aiboh measured with no resistance / unimpeded airflow, then the actual max would be a little higher.
 
6V? That would be around 1100rpm max according to Aiboh's chart.

However, the chart on Silverstone's product page indicates the fan goes up to almost 1500rpm which would be more like 9V according to the chart. Since Aiboh measured with no resistance / unimpeded airflow, then the actual max would be a little higher.
I guess it could be pushed higher with a higher ambient temperature. It's still far from using the full fan range, and still constrained by the thermistor.
 
What would potentially be the drawbacks of running this SF700 in a system that really only needs a 450w PSU?

I'm trying to work out a nice PSU for a Node 304, i7 6700k, 1xSSD, 1xHDD, 1x GTX 970. I was originally aiming at a quality ATX 450w-550w. I'd love to go smaller, but SFX PSUs are like the wild west of PC components, so I'm extremely wary about the current offerings (and what the hell is taking Corsair so long with their SF600..). I'll never need 700w, but if it's a quality unit and is efficient in my low-power system, I'm interested.
 
There really isn't any drawback. In theory the system won't be running at the PSU's sweet spot for efficiency but it'll be platinum rated so it'll still be fairly efficient even at low load.
 
Looking for a SFX psu for a Skylake 6700K, z170itx, AMD Nano build. At what powerdraw does the switching occur at normal room temperature? Would the system be "clicking" when doing desktop work?
 
Looking for a SFX psu for a Skylake 6700K, z170itx, AMD Nano build. At what powerdraw does the switching occur at normal room temperature? Would the system be "clicking" when doing desktop work?

This model hasn't been launched yet, and it won't have the same fan control as the previous models therefore experience from earlier models won't help.

The other previous models (500 and 600w) do however click when in desktop, atleast the ones i had, but that seems to differ and since it's temp. controlled it'll depend somewhat on your case and it's cooling.
 
There really isn't any drawback. In theory the system won't be running at the PSU's sweet spot for efficiency but it'll be platinum rated so it'll still be fairly efficient even at low load.

Ah ha the fabled efficiency sweet spot - I've never been convinced :)

If you purchase your PSU based on twice your rig's theoretical max load, then your PC spends all its time below 50% efficiency. However, PSUs are comparatively more efficient above 50% load than below it. Under 20% they are significantly less efficient compared to 80,90,100% load.

Ecova test them at 20%/50%/100%, so it's in the PSU manufacturers interest to make sure 50% is the most efficient. They can just change the designated maximum load. Like the ST45SF-G which could have been a 500-550W bronze unit*. Or the SX600-G which could have been a 650-700W silver unit**. 80Plus rating sells more than current delivery.

*) See max load ST45SF-G http://www.chiphell.com/article-4711-5.html
**) See max load SX600-G http://www.chiphell.com/article-12094-5.html

Check the curves with the 80Plus logo to see efficiency.

So if your PC is mostly not under load, you'd surely better choose a PSU that is more efficient at the load you are using it for? And match the max load of your PC with say 80% of the max rating?

And if you are using your PC mostly under load (e.g. 24/7 folding or mining), then will you actually save more money from lower electrical bill than you pay extra for a more powerful PSU?
 
Good insights, interesting. With the first series of the Ncase M1, everyone going SFF was limited to 450W and with an i7-4770K with a GTX 780 or R9 290X was hitting the 350-400W under load, along the 80% line of typical load you mention. With the 600W this is not the case and most will be hitting 300-350W thanks to the more power efficient components.

But the 700W in this topic is going to allow two GPUs with ease, opening up a whole can of whoopass for SFF mATX.
 
I've always had this idea that people mistake "most efficient at 50% load" as "least heat and noise output at 50% load, more heat and noise at lower or higher percentages than 50%".
 
For any 80plus rated PSU, the efficiency curve under differing loads is going to be pretty flat. e.g. Jonnyguru's testing fo the SX500-LG had a variance of only 5% for all loads (other than the 'let's load up 3.3V and 5V like nobody ever does' test).

Overspeccing a PSU loses you effectively nothing in terms of efficiency*, it just costs you more up front. If you intend to upgrade components in the future, then this may be a worthwhile investment in not having to buy a bigger PSU later.

*If you're choosing between a 80+ Gold 600W or an 80+ Platinum 700w, load-for-load the Platinum will probably still be more efficient

I've always had this idea that people mistake "most efficient at 50% load" as "least heat and noise output at 50% load, more heat and noise at lower or higher percentages than 50%".
I'd like to see graphs of waste heat vs. load (i.e. (1 - efficiency) * tested load) be more common. That gives a much better idea of the absolute heat output a given PSU is dealing with at a given load. Graphing relative percentage efficiency makes comparisons between PSUs of different absolute load ratings much harder.
 
I'd like to see graphs of waste heat vs. load (i.e. (1 - efficiency) * tested load) be more common. That gives a much better idea of the absolute heat output a given PSU is dealing with at a given load.

There you go, I drew a sample graph in ms paint.

 
I plugged the 110V efficiency measurements from the two Chiphell reviews into a spreadsheet and plotted efficiency and waste heat values against actual output:

ooJyHzH.png


The 450W unit has a slight advantage between ~50W and ~180W. The 600W unit starts pulling away after ~200W. At 360W the advantage is ~10W. 10W at 12 cents per kWh comes out to be just above $10 a year.
 
I plugged the 110V efficiency measurements from the two Chiphell reviews into a spreadsheet and plotted efficiency and waste heat values against actual output:

ooJyHzH.png


The 450W unit has a slight advantage between ~50W and ~180W. The 600W unit starts pulling away after ~200W. At 360W the advantage is ~10W. 10W at 12 cents per kWh comes out to be just above $10 a year.

That's assuming 360W usage for 24/7. So the 450W is actually producing less waste heat for normal usage and costs less.
 
That's assuming 360W usage for 24/7. So the 450W is actually producing less waste heat for normal usage and costs less.

Whatever that normal usage is, at no point is the difference below 250W significant.

My point is, even if you're running 360W 24/7/365, the cost of the extra waste heat is insignificant compared to the overall cost of keeping the system running, so actually, buy a unit based on its quality and acoustics.
 
... buy a unit based on its quality and acoustics.

True. Having tried the SX500-LG and SX600-G now, and since I still have a couple of those Noiseblocker PC-P fans spare, I'd buy another ST45SF-G before either of the former.

:D
 
My "Sirfa 500W SFX-L"-based Sharkoon SilentStorm Gold 500W runs admirably, nothing to complain. Just to let people know there are other brands besides Silverstone that make SFX PSUs ;)
 
Very interesting info on PSU efficiency. Keeping in mind the 80% load idea, would you recommend the 500w or 600w for i7-6700k & 980 Ti? I thought the 500w would be fine, but I saw this Toms HW article claiming it can use up to 428.38W (Under Gaming Loop > Maximum). But that makes no sense to me, there must be something I misunderstood.
 
'Maximum' means exactly what it says: the maximum load TH observed. If you had a cheap-as-hell PSU whose 'rated' load was really it's peak load (i.e. if you tried to run it at rated capacity for more than a few ms it would fail catastrophically) it might be a worry. But those Sirfa SFX PSUs are not cheap-as-hell, so they will output their rated load indefinitely (as long as you don't do something silly like block the intake). A 428W peak is WELL within what the SX500-LG can provide, and a few peaks to >400W would probably be fine with a 400W supply too.
 
I have a 980 running on a 300W SFX PSU. Peak load doesn't matter too much if you have a PSU that isn't terrible.
 
I have a 980 running on a 300W SFX PSU. Peak load doesn't matter too much if you have a PSU that isn't terrible.

That's not always true because good quality PSUs may shutdown when there is unusually high peak power draw if they have sensitive OCP (over current protection). In contrast, a lesser quality PSU could have slower responding protection that's unable to catch the millisecond peak draw and continue working.
 
If the PSU has that over-sensitive overcurrent protection, that would punt it out of the 'good PSU' category IMO, due to the requirements of a 50% transient load tolerance on the 12V rail in the ATX12V specs.
 
Any updates on when we might see this released?

My trusty ole 500 is holding back my Titan X benching :D
 
Any updates on when we might see this released?

My trusty ole 500 is holding back my Titan X benching :D

Sorry to break out the bad news, release schedule for this PSU has been delayed to the end of this year as more fine tuning is required.
 
I hope you have recruited some noise-sensitive people from this forum to work out fan noise this time for really really real this time :D
 
I'm very curious how they tested, or how PSUs are tested before release (aside from normal testing). Does any real world testing get done? Does anyone even do that with PSUs?

Like beta testing for hardware.
 
Back
Top