![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Silverstone SST-SX600-G 600w Gold SFX
- Thread starter m2h
- Start date
AFD
2[H]4U
- Joined
- Aug 7, 2013
- Messages
- 2,154
It opens doors, and can be quite a cheap hobby to start off with
Not the way I solder. End up damaging as much stuff as I fix
JoepieDeSnoepie!
n00b
- Joined
- Jan 14, 2014
- Messages
- 50
I see. Interesting, none the less.
Try a pci-e riser maybe, as a temporary solution for testing?
I've done some digging but I haven't come across any suitable replacement fans yet. Unfortunately I haven't found any 80x15mm fans with FDB or something similar at the RPMs needed to work here
Hopefully something will come up. If there were a simple way to positively offset the voltage (i.e. add a constant +3V or something), it should be a lot easier to find a suitable fan (?). But I can't think of any simple way.
That's a really good idea! But, I'm guessing there'd still be a problem with finding a suitable 120mm fan then?
That's excellent, especially if the power measurement is on the AC side?
Aibohphobia
[H]ard|Gawd
- Joined
- Nov 16, 2013
- Messages
- 1,340
Hopefully something will come up. If there were a simple way to positively offset the voltage (i.e. add a constant +3V or something), it should be a lot easier to find a suitable fan (?). But I can't think of any simple way.
Answering myself, it should be possible to use a summing amplifier to do this, adding the voltage from the psu fan controller to 3,3V from the psu itself? I'm not an electronics engineer though, so there are probably several problems with this...
Aibohphobia
[H]ard|Gawd
- Joined
- Nov 16, 2013
- Messages
- 1,340
Answering myself, it should be possible to use a summing amplifier to do this, adding the voltage from the psu fan controller to 3,3V from the psu itself? I'm not an electronics engineer though, so there are probably several problems with this...
Tony must think we're crazy, considering recasing the PSU and tapping voltage from an internal line in order to use a slim 120mm fan so it can power a hexacore and SLI 980s as silently as possible
veryrarium
Weaksauce
- Joined
- Apr 12, 2014
- Messages
- 125
That would be nice but Sanyo and Nidec Servo only make 2BB fans, no sleeve or sleeve-variation bearings, I think.I looked into it when I learned the Gentle Typhoons were being discontinued because my dream fan would be a GT with PWM and FDB with a black frame with red blades
Anyway a fantastic work, Aibohphobia, a lot of useful measurements! And I'm also looking forward to your case! (Hopefully a prototype next month?)
Aibohphobia
[H]ard|Gawd
- Joined
- Nov 16, 2013
- Messages
- 1,340
Nidec does but as far as I know it's been limited to specialized applications like ultrabooks. I believe some of the MacBooks use them.
http://www.nidec.com/en-NA/technology/story/uff/
Hopefully.
Now that I know such a high-end build is feasible I can't wait to give it a proper SFF home instead of a test bench
I'd buy that
I've been thinking about it, but there's really no way to make it a kit unless you don't mind the PSU socket positioned as I showed it on the previous page. That's not a very handy place, especially for NCase M1 which has a bracket on either side.
Here are the things that will cost some money.
1. The right fan, adapter cable.
2. Steel sheets.
3. Enclosure needs to be laser cut or stamped, because of the modular sockets. I've cut and filed them out by hand twice and it took hours.
4. Steel needs to be folded on a proper press-break.
5. Standoffs, screws, etc.
If it weren't a kit, but a conversion service, to desolder and get a new socket in the right place. Then additional costs are: mailing the PSU back and forth, a new socket, wires and solder.
Overall I think if you really really want a 600W+ SFX-L then you'd be much cheaper off importing the Dirac from Japan.
JoepieDeSnoepie!
n00b
- Joined
- Jan 14, 2014
- Messages
- 50
As you may have read before at another place WiSK you know I got all the machinery and skills to pull it actually of... if I really wanted. Also desoldering is no issue at all... the problem is time right now unfortunately. With the lasercutter and folding machine it would be a piece of cake (or even an own case for that matter, but the reason I bought a ncase was time again though the design was pretty similar to mine).
Aibohphobia
[H]ard|Gawd
- Joined
- Nov 16, 2013
- Messages
- 1,340
I don't know about that. There's a company in Japan called Tenso that offers a buying service where they'll buy the item for you and have it shipped to their warehouse and then it'll get forwarded on to you anywhere in the world.
The 650W Dirac including tax, fees, and shipping would end up costing around $240
Now if only you had the same user name here, as at that other place
I don't know about that. There's a company in Japan called Tenso that offers a buying service where they'll buy the item for you and have it shipped to their warehouse and then it'll get forwarded on to you anywhere in the world.
The 650W Dirac including tax, fees, and shipping would end up costing around $240
So $110 more than the SX600-G. Where can you get steel cut accurately and folded, and a good fan, and return-postage to someone who can desolder for you for $110?
Aibohphobia
[H]ard|Gawd
- Joined
- Nov 16, 2013
- Messages
- 1,340
You're right, I forgot about the case, I was just thinking of the cost of redoing the inlet
Tony must think we're crazy, considering recasing the PSU and tapping voltage from an internal line in order to use a slim 120mm fan so it can power a hexacore and SLI 980s as silently as possible
Or maybe he should just listen carefully to what we're saying?
The 650W Dirac including tax, fees, and shipping would end up costing around $240
$240 is acceptable to me, but what was the plan here; to cut a larger hole in the case and replace the 92mm fan with a 120mm? Or just hope that the 92mm is actually any better than the 80mm in SX600-G?
Aibohphobia
[H]ard|Gawd
- Joined
- Nov 16, 2013
- Messages
- 1,340
$240 is acceptable to me, but what was the plan here; to cut a larger hole in the case and replace the 92mm fan with a 120mm? Or just hope that the 92mm is actually any better than the 80mm in SX600-G?
Either. I think the extra room in the case will allow for better airflow so maybe the 92mm will be enough to quietly cool it. If not it's nothing a dremel can't fix
Zangmonkey
Supreme [H]ardness
- Joined
- Jul 6, 2005
- Messages
- 5,065
A good recommendation over all with a jab at some of the third-tier capacitors (exactly as Paul Johnson said).
I'm still waiting to see if anything happens with the L but these reviews are good indicators that this PSU is a strong performer.
I see SilverStone releasing a new revision (v1.1 or 2.0) later to replace the lower tier parts with better ones.
link?
I really hope they end up doing that with the crappy caps
I see SilverStone releasing a new revision (v1.1 or 2.0) later to replace the lower tier parts with better ones.
Scan now have them in stock.
I think he was saying that he could see them coming out with a revision at some point to address this issue.
Yup, that's exactly what I meant.
Either. I think the extra room in the case will allow for better airflow so maybe the 92mm will be enough to quietly cool it. If not it's nothing a dremel can't fix
True. Hmm. Well, I should get my SX600-G (and GTX 980) tomorrow, so I'll know soon enough if I have to do something about the noise.
Would be interesting to see if a high priced SFX in the Zeus series would have a market. Server grade components and 5+ year warranty. Special attention given to the build quality of the fan: heavier hub, well balanced, less airflow / more torque ... not a 4500rpm fan running at low volts, but a 2500rpm that uses the full range of power. I'm not sure how feasible heatpipes are in a PSU. The modular daughterboard connected directly to the main board instead of wasting space of all those wires. I could go on but really I know nothing about high end PSU tech.
Aibohphobia
[H]ard|Gawd
- Joined
- Nov 16, 2013
- Messages
- 1,340
They could probably fit a 92mm in there too. It would fit in the SX600-G if the AC filtering board was a little smaller or moved elsewhere
Aibohphobia
[H]ard|Gawd
- Joined
- Nov 16, 2013
- Messages
- 1,340
Sure, I'll give it a shot after I get off work.
Hmm. The working range of the fan controller seems to be roughly 3 to 6 V? If the fan controller voltage was added to itself, i.e. by Y-splitting the wire and using a summing amplifier, the range would be transformed into 6 to 12V. There should be a bunch (well, one is of course enough) of quality replacement fans in that range?
Yes, there should be a bunch. But until now, the fan mod has been fairly easy. Either solder the wires, or buy the adapter from ModDIY, or make your own.
To add an additional component introduces extra difficulty and costs. It needs to be tiny to fit, easy to buy and connect.
Edit:
What about a 5V fan? Qualtek make 80x15 fans either 2-ball, hydro or sleeve bearing. Unfortunately the spec sheet doesn't say if 5V is the upper limit or just the recommended operating voltage.
http://www.qualtekusa.com/images/Fans/pdf_files/FAD1-08015.pdf
Last edited:
Yes, but as a last resort, if all else fails...
Good find! 5V should be the upper limit (I guess?). This could be exceeded by the fan controller in the SX600-G on heavy loads of course, but I for one wouldn't have a problem with that (unless the fan burns out, but ... shit happens).
Aibohphobia
[H]ard|Gawd
- Joined
- Nov 16, 2013
- Messages
- 1,340
Ok I got some readings. I didn't get the full range because with Prime and FurMark going the fan wasn't ramping up to full speed and some of the components inside were getting to 80°C! I think not having the top panel on was messing with the airflow causing the temp sensor to run too cool so it wasn't increasing the fan speed as much as it should.
What I got matches the results I got testing the fan by itself though. It started at 1V and slowly increased until 3V at which point the fan starts turning. As it got hotter the voltage increased until it got to 4V @ 1600 RPM (which matches the chart here). That's when I stopped the test.
So the RPM and voltages I got testing the fan by itself match what the PSU header is doing, no funny business going on. So we're looking at essentially a 1-5.5V range.
Edit:
Just noticed that with the machine idling the voltage is 3.05V, right at the fan's starting voltage. I wonder if the voltage/temp curve is slightly off and it never drops quite enough for the fan to shut off after it's come on.
Edit 2: Pro tip, when screwing down the top cover on your PSU, cover the fan grill so that the screw can't fall inside
What I got matches the results I got testing the fan by itself though. It started at 1V and slowly increased until 3V at which point the fan starts turning. As it got hotter the voltage increased until it got to 4V @ 1600 RPM (which matches the chart here). That's when I stopped the test.
So the RPM and voltages I got testing the fan by itself match what the PSU header is doing, no funny business going on. So we're looking at essentially a 1-5.5V range.
Edit:
Just noticed that with the machine idling the voltage is 3.05V, right at the fan's starting voltage. I wonder if the voltage/temp curve is slightly off and it never drops quite enough for the fan to shut off after it's come on.
Edit 2: Pro tip, when screwing down the top cover on your PSU, cover the fan grill so that the screw can't fall inside
Last edited:
Ok, this means that the "semi passive" functionality might not work with a 5V fan, if it actually spins at 1V. But it should still be very silent.
It also means that the fan controller is extremely simplistic. Hmm. What about replacing it altogether? He he, I just won't stop, will I?
Oh, and good work with the measurements!
veryrarium
Weaksauce
- Joined
- Apr 12, 2014
- Messages
- 125
Thanks for the crrection, I just did some googling now and indeed found that they used to sell this Nidec 80x80x15mm fan, Beta SL D08R-12TH-19A, that span at 3100rpm/34dB@12V with what they called an NBR bearing, a type of fluid dynamic bearing, back in 2001. Can't find this on the Nidec sites but they do list what seems to be its successor, UltraFlo U80R12MHAB-51. I don't know if this fan starts at a much lower input voltage than what's said on the spec sheet (like San Ace fans often do) but regardless 3100rpm@12V is probably too low as an SX600-G fan replacement where we want something like 4800rpm@12V so it's probably not a good candidate.
I wonder why Enhance/Silverstone decided to go with the combination of a high rpm 12V fan with very low input voltage control rather than a medium speed fan with an input voltage ranging like 6-12V.
Aibohphobia
[H]ard|Gawd
- Joined
- Nov 16, 2013
- Messages
- 1,340
Nice find on the fan! [Strike=Option]Don't see anywhere to buy it though
[/s]Edit: Nevermind, found it. $5.17/ea
Minimum order quantity: 6000
http://www.futureelectronics.com/en/Search.aspx?dsNav=Ny:True,Ro:0,Nea:True,N:4294938264-4294900385
Edit 2: Found the PWM version for $25 from AliExpress: http://www.aliexpress.com/item/Nidec-u80r12mhab-57-12v-0-22a-8015-8cm-silent-cooling-fan/1459746105.html
If we can figure out a way around the voltage issue I'm tempted to order one. Anyone have any experience buying from AliExpress?
I was looking through Tony's previous posts while writing up the SX500-LG thread and I think it's a combination of relatively low noise, doesn't click when it hits it's starting voltage, while still able to go to a high enough speed to cool the cramped, hot interior of the power supply. Apparently it took them a while to track down a fan that met all those requirements.
While some of us here would be willing to pay a hefty premium for the power supply if it had a something like a Nidec FDB fan in it but the reality is that Silverstone (and any other OEM for that matter) has to sell enough units to justify the R&D expense. It's easy to think that it's just another $5-$10 or whatever but you have to remember that every $1 increase in the cost to manufacture increases the retail price by $3-$4 (exact numbers probably different but you get the idea) because Silverstone needs to make money, the distributor needs to make money, and the retailer wants their cut too.
$140 is already expensive for a 600W PSU but we accept it because there is no alternative for SFX but if it was $180 that would drastically shrink the potential market for an already niche product and Silverstone may not be able to sell enough to make it worth their while.
Last edited:
$240 is acceptable to me, but what was the plan here; to cut a larger hole in the case and replace the 92mm fan with a 120mm? Or just hope that the 92mm is actually any better than the 80mm in SX600-G?
In that Japanese magazine DOS/V I mentioned earlier in the thread, the 92mm fan in the Dirac unit was noticeably loudest of the three compared SFX class PSUs. Below are data from the article.
Environmental temperature = 27C
Background noise ~ 30 dB
Test system: Intel Core i7-4770K stock, ASUS Z97-K, 4GB x 2, AMD Radeon R9 270X, Plextor 256GB SSD. (Maximum power draw likely around 300W)
SilverStone SX600-G (80mm fan)
Idle: 30.4 dB
Load: 33.2 dB
Scythe SPKRG-S500P (120mm fan)
Idle: 30.6 dB
Load: 32.1dB
Dirac TESLA CUBE (92mm fan)
Idle: 42.0 dB
Load: 42.1 dB
Also of note, Dirac's casing depth is 120mm, so it won't fit a 120mm fan.
Aibohphobia
[H]ard|Gawd
- Joined
- Nov 16, 2013
- Messages
- 1,340
Hey Tony, great work on the SX600-G! While we may like to see a better selection of caps and a even quieter fan, I am super impressed with it nonetheless. It handled my i7-5820K and SLI GTX 980s build like a champ
Do you think we'll see a 600W or higher SFX-L unit in the future? Whether it's a higher wattage version of the SX500-LG or a recased SX600-G.
Anyway, thanks for pushing the envelope of what's possible in SFF.
Do you think we'll see a 600W or higher SFX-L unit in the future? Whether it's a higher wattage version of the SX500-LG or a recased SX600-G.
Anyway, thanks for pushing the envelope of what's possible in SFF.
So, this is pretty strange: I'm using my sx600 on my gigabyte z87n board for a week now and always noticed this tedious buzzing sound it makes when the pc is powered off but still plugged into a power outlet. Turns out disabling ERP in the bios fixes this problem. Does anyone know the reason behind it?
Sounds like capacitors discharging?