Silverstone SX500-LG 500W Gold SFX-L

is it possible with analog electronics to map the voltage to do something like
if voltage < 4 then voltage = 0
 
El-Dubya said:
Still trying to catch up on previous posts, but if changing the fan out with something like that turns out to be really simple to do I'll be kicking myself for returning.
Click to expand...

Changing the fan by user was always doable but would void the warranty.
The best alternative would be for Silverstone to find a good quality, nice behaving fan to be installed at the factory. I had brought up their own FW121 for that purpose...
 
Tony Ou said:
I think we have a part in this too. After High Power showed that prototype in 2013, no one wanted to buy it for a year (because no case was built specifically for it). When we approached them and confirmed our order and commitment to support it with our case products, that is when they were able to get Chieftec and others on board. We should have been the first to market with it had we not held back and worked on making it quieter.
Click to expand...
I apologize if it seemed as though I diminished your role in it. Certainly, without your declaration of commitment to support the form factor it's very unlikely SFX-L would've come to market - as a Chieftec or anything else. Silverstone has always been at the forefront of SFF development, and here is no exception.

I only meant to clarify our role (as NCASE) - for having the idea in the first place, and for reaching out to manufacturers (for which the credit mostly goes to w360). Likewise, it's worth recognizing the initiative High Power took to develop a prototype for what is a pretty niche PSU form factor.

Tony Ou said:
Yeah, we've tried our own FW121 too, unfortunately its was too efficient in that it couldn't stop spinning even when voltage was dropped below 3.5V (despite its starting 5V rating). So it wasn't a feasible option for semi-fanless operation that we wanted on the SX500-LG.
Click to expand...
Is it technically unfeasible to just shut the power to the fan off completely for fanless operation?
 
theGryphon said:
Don't get me wrong, but I'm not sure if anyone really cares about semi-fanless operation. I mean, who cares if the fan is spinning at 300rpm? What matters, and what almost everyone here seems to care about is not the idle performance and db, but rather the load performance and db. That and irregular noises (galloping, etc.)

Thanks for your response btw.
Click to expand...

I agree completely, but then again semi-fanless operation seems to be really important from a marketing point of view (outside of this forum anyway). It's one of the biggest recent trends.
 
flod said:
@fusionxr: nice find, yes mine looks exactly the same as the one in your pictures, other than the wires and stickers



if i spin the fan manually with my fingers, yes i can hear a brushing sound. it's definitely from the bearing.

interesting thing is that the sound becomes worse at certain angles. for my fan, if i orient it vertically so that, when looking at the side without the stickers, the wires come out the bottom of the left edge, the brushing sound disappears!!!

in that orientation the motor is actually really quiet when i spin it with my finger.
Click to expand...

This is excellent investigation flod!

It appears the fan has been manufactured with a different motor and bearing, with the same fan blade profile as ID cooling's. Could it be the bearing isn't aligned properly and when spinning at low rpm, it brushes against a surface creating that galloping sound? At higher rpms the bearing probably evens out and align itself. exp;aining why the sound is only apparent at low rpms.

flod, do you mind connecting it to the psu (and put the gril back on the fanl), while keeping it outside and in that angle to see if it still makes that sound at low rpm?

Interestingly, this fan is quite cheap and does have some beefy specs for a slim fan. Silverstone are using this "semi fanless" feature as a big selling point. I can see why they're so hung up on "intelligent fan operation".

I think silverstone should keep this fan, and fix the bearing - instead of using a different fan which might drive up costs.
 
I'm wondering whether the galloping noise is related to the thermistor control.

As the heat sink warms up the voltage to the fan reaches a point where it starts, this then immediately cools the heat sink down and it stops. One could reach a steady state where the fan is starting and stopping fairly rapidly.

Is there a timer that keeps the fan going for a minimum time?
 
I'm wondering whether the galloping noise is related to the thermistor control.

As the heat sink warms up the voltage to the fan reaches a point where it starts, this then immediately cools the heat sink down and it stops. One could reach a steady state where the fan is starting and stopping fairly rapidly.

Is there a timer that keeps the fan going for a minimum time?
Reply With Quote Multi-Quote This Message Quick reply to this message
Click to expand...
I would think if that was the case more units would do it. Also, since fans or any electronic are not all created equal there are some variances in resistance and friction which in turn will create more noise. It's an interesting issue.

The desision to use a sleeve bearing in a psu sorta baffles me, other than from a cost cutting/stay in budget point of view. From what I understood about sleeve bearings for longevity sake it should be used in an application that keeps the fan vertical and psu's can be mounted both horizontally and vertically. Also the hotter temperatures can cause sleeve bearings to dry out faster, and ultimately fail faster. Being in a semi fanless psu makes me wonder if the heat in the fanless mode could potentially cause addition wear and tear on the bearing, since the psu is allowed to get a hotter.
 
fusionxr said:
flod, do you mind connecting it to the psu (and put the gril back on the fanl), while keeping it outside and in that angle to see if it still makes that sound at low rpm?
Click to expand...

blah i just ripped out the wires from inside the case, as i was convinced that my ap-14 provided adequate cooling. i guess i can hook it up again. does the grill matter? i'll try it sometime when i'm not feeling lazy :p
 
Tony Ou said:
Sorry to hear that we still can't satisfy some of you guys here, we really tried. We got a lot of flak internally and from management for delaying SX500-LG three months to make it as quiet as possible (to be quieter than our own ST45SF, ST45SF-G, SX600-G, and High Power/Chieftec's standard fan).
Click to expand...
Tony, thank you for engaging in this forum. As something positive, perhaps you could use people in this thread next time as testers - if they are the most critical voices, they are going to give you the best feedback on your products.

As for me - I'm happy with my (second*) ST30SF, where the fan never turns on ;)



* The first one made a lot of electrical noise and sparked inside once or twice
 
I think a lot of the issues with fan selection comes from the very, very simply method of direct thermistor voltage control. The use of an actual fan controller would dramatically broaden the number of fans that could be used, as well as massively increasing the range of configurations available (e.g. PWM fan control, configurable hysteresis, non-linear response, user-electable profiles, etc). It unfortunately has a minor increase in cost (the cost of a very basic microcontroller, plus the PCB space to mount it on, plus testing and validation of the uC program) which is likely why OEMs have not bothered to offer it.
 
Dear Silverstone, please use this hole in the market to develop a simple, yet effective fan controller for a PSU. If the user could choose by means of a switch at the rear between the following:
- always-on fan with a temp-based curve (minimally 5V and can ramp up to 12V)
- fanless mode with a temp-based curve (like it is now, but for the target demograph that seems to exclude most if not all SFF people here)
- static fan speed that would allow ~100% load with 40°C intake temperature

These can be made rather small on a slim piece of PCB, but it would require a creative electronics expert to make this. Not just some send it forward to the OEM or it will be botched up to have a lot of iffy side effects most likely. Case in point: Ncase M1 vs any Lian-Li mITX case. Maybe someone on this forum might be willing and able to get it done ? Why not have it crowd-sourced ? A company doesn't have to have magical powers like Apple does to make well-engineered products, it's about the the care, creativity and passion people put into their work.
 
I think warranty reasons are a factor, too. If you only have that one configuration, one speed curve and one fan, and tested a PSU running 100% load on that for 24h, you can safely guarantee that your PSU won't overheat.

If you allowed the user to swap and change that sort of stuff, you'd be fine until someone hurts themselves doing it. It's just safer to have one set configuration for the manufacturer.
 
There's nothing wrong with a simple thermistor. The only reason we would want an adjustable controller is because we want to change the fan. As IFreilicht says, it becomes a warranty and certification nightmare.
 
But again, I hear no one say the fan needs to turn slower or stay off longer. I personally want it to be on all the time, opposed to the more-prone-to-failure-and-issues semi-fanless mode.

What I proposed would also not change anything in regards to cooling capacity, as the fan curve can stay the same but just not drop below 300rpm (for example), the clicky-clacky semi-noiseless erm... semi fanless mode and an always-on non-adjustable speed setting for stuff like servers, mining rigs and whatnot (that will be noisy).
 
Assuming the thermistor and fan are wired in series, and that the thermistor's resistance decreases with increasing temperature, I think the simplest way to add adjustability without allowing the fan speed to drop further would be to add a variable resistor parallel to the thermistor.
 
vipz said:
Assuming the thermistor and fan are wired in series, and that the thermistor's resistance decreases with increasing temperature, I think the simplest way to add adjustability without allowing the fan speed to drop further would be to add a variable resistor parallel to the thermistor.
Click to expand...

I'm not an expert but the thermistor will be part of an electronic circuit controlling the current supplied to the fan.

The resistance of a thermistor is quite high and will only pass a very small current (mA).
 
Last edited:
I wouldn't be surpised if there's already a resistor in parallel to the thermistor. It could be replaced.
 
So I got a SilverStone FN123 to evaluate as a possible replacement:

aT8LBDg.png


Fan starts at 3.7V though like Tony mentioned once it started it would stay running down to about 3.1V.

Maybe it's just my unit but it makes an annoying rasping noise though. It's fairly faint and seems to lessen at higher RPMs but it's audible. So it looks like the hunt for the perfect fan is still on.
 
Personally, I'm quite happy with my SX500-LG, it's mounted sideways in a Lian-LI PC-Q25 using the ATX bracket with the fan facing outwards as to not bother the CPU's airflow if the fan spins.

I can't hear any weird noise from it when it does spin.
The fact that it's mounted vertically and not horizontally may be actually helping with this, as was mentioned earlier in the thread.
When I was testing with the side of the case opened, semi-fanless operation does work, and even if it does spin, I can't hear it.

I have an intel pwm fan on the Intel stock cpu cooler, a Noctua NF-A14 PWM in the front and a Noctua NF-S12A PWM at the top.

Essentially the loudest component in the entire case is the 6x 7200rpm HGST drives, which I can't do much about.

Overall, the SFX-L form factor made a huge difference with regards to space in that case and I'm quite happy with it.

Only thing I'd comment about would be the included flat cables are flexible in 1 direction but very stiff in the other. Maybe a compromise of flat/round power cables would be something to look at.

Actually, speaking of the hunt for the perfect fan, I bet if there's anyone that can make a 120mm slim fan with excellent specs, it's Noctua.

Might be worth if Silverstone contacted them. With all their acoustic, bearing and airflow expertise, pretty sure they'd make a killer fan for the SX500-LG v2.0.
Even if the fan ends up having their signature beige/brown color scheme, who cares :) Although they have now made gray and black fans, so it doesn't have to.

Essentially, moving air through all the tightly spaced components in the SX500-LG is like moving air through a radiator, it's all about static pressure.
So a 15mm slim version of their "Focused Flow" system, combined with their excellent bearings and excellent acoustic properties and you have a clear winner.
 
Last edited:
While technically it could be an awesome combo of a custom Noctua fan + SX500, and I would buy one in a heartbeat, it's a marketing nightmare for Silverstone, who would be admitting that even tho they make some decent fans, can't actually put one in their own power supply.

Best to leave it as it is and just claim a tiny fraction (some 44%) of their customers are just too picky to be pleased by a poorly selected and built, clicky sleeve bearing fan. Or something.
 
They actually already confirmed that their own FW121 fan is not up to par, Tony Ou said it and the fact the SX500-LG's fan is a Globefan OEM fan also does.

Also, we (on the H-forum) aren't really representative of the average SFF demographic, since we engage in active discussion and try to go beyond a plug-and-play build. Most people don't, even people building SFF themselves.
 
I just noticed something interesting with my SX500-LG's fan. I am using this PSU in a NAS and at this moment the fan is turning on for ~10s, shuts off for ~5s, and then turns back on and repeats this cycle continuously. Should I be worried that this constant on/off may prematurely kill the fan? The load on the psu must be pretty low but just enough to kick the fan on.
 
flod said:
is it possible with analog electronics to map the voltage to do something like
if voltage < 4 then voltage = 0
Click to expand...

You certainly could with a mosfet. Spec the gate voltage such that Vgs -Vt = 0 at 4v. Or a comparator Opamp config.
 
Don't know if it's already been mentioned, but the unit isn't available in any stores in Sweden right now.
They're not out of stock, they've been removed from several (10+) retailers website.
Seems odd.. unless it's discontinued, maybe replaced by a new version?

EDIT: found it, might just been a name change which caused it.
 
Last edited:
henkzinho said:
Care to share the new name? I bought mine from inet and whilst contemplating if I should replace the unit or not I also noticed this. Not available at a single swedish retailer.
Click to expand...

https://www.komplett.se/silverstone-sst-sx500-lg-500w-psu/840272
Is the one i found, i think it was listed @ prisjakt without the sst in the name earlier.
Therefore it's still available, but it's weird that neither webhallen, inet or cdon list it.
But that could be of other reasons..
 
How bad do you think it would be run a Titan X on this thing given that there is 40A on the 12V rail for the SX500-LG?
 
Well the Titan X and R9 290X seem to have similar power and heat characteristics. It seems to work fine for people on the Silverstone 450 PSU with mild overclocking so I don't see a problem.
 
Aibohphobia said:
I really like Anandtech but I wouldn't rely on them for video card power consumption figures because they're still measuring total AC system draw.

TechPowerUp is measuring DC draw for just the card itself (along with PCPer and Tom's Hardware). According to them running FurMark the 980 draws 190W, Titan X 275W, and the 290X is 328W.

http://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_Titan_X/27.html
Click to expand...

The difference is striking. Could it have to do with 290X's power profiles? I never used one, but I know there is this "Uber" mode...
 
I received and installed my SFX-L yesterday and was quite surprised by how loud the galloping noise is, to the point where it made me want to reinstall my old 450w unit, which is clicking, but not as loud. Ultimately, I decided to try the sfx-l with the fan facing down (I have a SG05 and had it facing up, in order for it to have fresh air and, hopefully, keep the fan off). With the fan facing down, I can't hear the PSU at all and I'm quite happy about it.

I use a GTX 960 and noctua CPU and case fans, now my build is finally completely inaudible while doing some light work/gaming, even with the case on the desk.
 
Last edited:
Thanks for the feedbacks guys, please keep them coming. I also want to reiterate again to everyone that SX500-LG was designed to be quieter than all available 450W+ SFX PSU on the market. So any experience to the contrary would be our priority.

WiSK said:
Kolink, the Taiwanese company who make Noctua fans, are based across town from SilverStone. Perhaps Tony can pop by one afternoon and ask if they would make a slim 120mm fan.

https://www.google.com/maps/dir/Lan...5296a8!2m2!1d121.4890177!2d25.0002475!3e0!5i2
Click to expand...

Noctua practically owns Kolink so they won't make fans for anyone else. With that said though, they've not had experience making slim fans so don't automatically assume that if they do, they'll be better.
 
Aibohphobia, if you don't get enough interest to get the brackets made, could you put diagrams and dimensions up so we can get them 3D printed ourselves?
 
You must log in or register to reply here.
Back
Top