Our guide to a mid-price/performance gaming PC, Dec 2010.

http://www.silentpcreview.com/Silent_Mid_Gaming_PC

Nice, very nice !! Thanks for another great homemade build review !!

The article left me confused about what fans were used. The case comes with a 200mm top fan and a couple 120mm front fans, but they aren't mentioned. The article states that six fans are used - I count only five mentioned in the article (1 case, 1 CPU, 2 graphics, 1 power supply).

We know the front fans aren't used but good question about the 200mm fan.

[EDIT : how about a folding rig ? That would be a challenge : a computer running at full load 24/7 with the best noise/power ratio !!! Let's see what it would be like at 2 noise levels : 18dB and 25dB for example ]

Luke M -- you're right, it's 5 fans as you counted, not 6. Corrected.

No case fans except the back panel 120mm fan was used. There is no need for any of the others. I thought this was made clear on the fine tuning page:


Couple of other things I should have mentioned in the article:
1) When the system is being worked hard, you can feel the heat rising up from the top panel vent for the big fan, so that vent is very useful, otherwise the back panel fan would have to move all that heat too. There is naturally hot air blown out by the back panel fan.
2) The Seasonic X650 fan was probably not spinning under any system load at stock clock speeds. It did contribute to the increased noise when the system was OC'd and stressed with Furmark/CPUBurn.
3) The temperature of the WD Green may have hit 40C at some point, but only after a couple hours of OC'd Furmark/CPUBurn. The CPU and back panel fan move enough air across/around the HDD so it stays pretty cool w/o direct active cooling.

Great guide.

One other question: is there any significance to the relative speeds of the CPU fan and case fan, or is this just an accidental consequence of using two different fan models? Would it be any better or worse to run them at the same speed?

Hmm interesting choice of components

First why AMD cpu? I5 750/760 runs cooler and will be faster after oc

Second - why OCZ fake low voltage memory (1,65V) when on newegg gskill ecos sell for less than 50$ and they run @1333 cl9 with 1,35V
http://www.newegg.com/Product/Product.a ... -_-Product

Third - why such a huge case? Wouldn't it be better to for example use P180 mini ? Or if we go with big ones perhaps fractal/P183 would be up to task ?

In our build guides, it is not always a question of the very best single components. There are many factors:

1) Unlike other sites' build guides, we actually build our systems so that the acoustics can be really heard and measured -- along with all the other testing that we do. That means the goods must be in our lab. We cannot always get every part we would like for every build in good time.
2) pricing/budget. We wanted to keep this system as close to $1000 as possible -- the i5-750 sells for $220 vs $145 for the Phenom2 x4 955. An i-760 was featured already anyway, in the SFF gaming PC guide. The 900 case isn't any bigger than a ton of other gaming cases, actually about 5% smaller than a P183. I think the Fractal might be even bigger than that... and a mini P180 does not fit an ATX board. Also, the 900 has been seen on sale for as little as $70 around Vancouver. No such pricing on the cases you mention. Besides, do we want to see yet another P183 build? Variety is good for the soul.
3) I don't understand your RAM question/comment -- again, it's what we had on hand that works well. "High performance" memory with high voltage requirements are such a PITA to work with. So many motherboards simply cannot post with such RAM, never mind run well.

Mostly accidental. If run at the same speed, there might have been a little less heat felt up at the top panel fan vent -- depending on just what speed those fans were set to. In a more closed case, I would try to run them closer in speed so that one does not end up impeding the other -- when two fans are in a "closed" push/pull configuration, it is the slower fan that sets the maximum airflow throughput.

I meant why you used 1066 ram running at 1,65V when for same price you could get better faster memory which will run at 1,35V.

Also 1,65V is industry stand so OCZ calling their memory "low voltage" seems like a sad joke to me.

The part I'm most interested in is motherboard choice. It is really difficult to get any handle on what type of BIOS fan control any given motherboard provides, the old sticky in the motherboard section hasn't been updated for years, and barely any mentions on the forums themselves. Oh, and Abit is also dead.

Has everyone moved to hardware controllers now?

GPU fans (Gelid Icy Vision): through Zalman FanMate 2 on max speed, on 3-pin Asus motherboard Chassis 2 header, set to Silent in BIOS fan control.

I don't quite understand the GPU fan control. Is the fan speed related with GPU temperature or something else?

Great post. But I'm curious - why did you go for the x-650 over the fanless x-460?

I have an overclocked 2.66ghz e7300 chip, hd5770 graphics but will ulimately swap to the amd 555 overclocked (retaining the current gpu) - would you recommend the x-460?

It's good to see someone read the article carefully enough to notice.

This was not really a simple build, despite appearances. We knew 18 dBA@1m was easily reached; if the SFF Gamer could do it, it would be no problem in a case with so much more space and ventilation. But to achieve lower SPL, all the fans needed to be carefully tweaked for minimal speed at lower power. This was straightforward with the back panel and CPU cooler fans, both of which worked well in Silent mode in the Asus BIOS fan control. But the Gelid ICY Vision fans ran at >1000rpm with the same setting. Many devices were used in between the fan and the header to reduce the speed, and there were some other devices which actually worked a little better in some ways -- but they lacked one simple trait necessary for a build guide: Ready availability. The Zalman fanmate2, on the other had, is ubiquitous anywhere Zalman stuff is sold, and cheap. You have to understand that even at full speed setting, the fanmate2 reduces voltage by about 1.5V; typically you get around 10.5V at this setting. And that was all the speed reduction needed to keep the Gelid fans down to <15 dBA.

Yes, the Silent mode for the case fan header in the Asus board is tied to CPU temperature. This is not such a bad thing, as there are no games which stress only the GPU without also stressing the CPU. There's lots of headroom for GPU cooling under this arrangement even in Furmark/CPUBurn, so a normal game is a cinch to cool quietly in the system.

This is one of the issues with any aftermarket VGA cooler. If the fans are not PWM, then thermal control functionality off the video card fan header is generally lost; all of the recent cards have 4-pin PWM fan headers.

Surely you noticed that at max lab test load the system drew nearly 420W AC? That's around 380W DC power delivered. Not unreasonable for the X-460 to handle as this is only under stress test conditions, but the fan-cooled X-650 would handle it better, especially if you consider longevity in hotter climes. Again it's a matter of thinking through on a design that will work for the largest group of users. Obviously, the X460 will work fine, too, and your proposed load is no problem for it, as the total DC power demand should be something like 60W lower than in this system.

FWIW the i5 760 is $160 at microcenter... picked one up a couple days ago...

the fan speeds during idle are impressive, and the controller on a controller is a great idea!

i've also recently discovered another trick for small fan speed tweaks:
quite often pwm fans react differently to voltage and pwm control, and i'm guessing that motherboards have different logarithms for each too, you can take advantage of this by manually setting the control method in the bios(my gigabyte was on auto). for example my fan spun about 5% slower when set to pwm compared to voltage, it's not much but sometimes it's all you need.

Novice builder... SPCR has motivated me to build PC's... anyway I built my own version of the silent "mid" rig:

2x - gtx 460, one msi cyclone, one msi twin frozr both sub 200 after rebate (based on slickdeals sales)
12gb ripjaws 1600 ddr3 - $50 / 4gb newegg BF sale
i5-760 @ 3.92 - $160 at microcenter a couple days ago
antec 1200
scythe kaze 4 fan controller
2x - 64gb microcenter brand sandforce SSD's
p55-gd65 - msi motherboard - ~120 open box on newegg
hyper 212+ - $25
5x noctua nf-p12 - got them oem for like less than $10/each at endpcnoise.com
AcoustiBlocks for empty drivebays
samsung oem sata dvd+r

Also I'm not sure if you can tell by the picture, but I have a 120mm scythe slipstream (took it off a mugen I had lying around) mounted in the middle of the case position (at the back of one of the drive bays) that blows directly on both graphics cards... playing bf2bc at full settings or just cause 2 and seeing > 90%+ utilization on both GPU's, temps don't go above 60, which is high, but all the case fans are at around 600rpm, and motherboard puts cpu fans on "silent." Also, I have the fan set on both cards to never ramp up over 40%.. obviously it took a lot of testing to get that ceiling number, but that's as loud as it will get.

unfortunately I got 2 radeon 6870's for xmas so I'm going to be doing another build soon. I love these build guides, I also built the SFF File Server (lian li pc-q08) with the exact specifications on spcr. unfortunately this gaming machine isn't as quiet as that one (it would be if I didn't do SLI). I'm thinking of migrating this PC to a p193. lemme know your thoughts.

http://www.terrapinsgonewild.com/interior1.jpg

Thanks for your clarification! Now I can understand your configuration, though it looks a bit strange

Another question is regarding the absence of sound proof material, which is widely used in SPCR built systems before.

Previously I encountered a strange resonance/whooshing noise issue in my Solo build, i.e. after reducing the system noise to a minimal level, the noise can be significantly louder with side panel closed than that with side panel open. This is just the opposite of common experience (still not resolved due to lack of high quality sound proof material in local market).
http://www.silentpcreview.com/forums/viewtopic.php?f=23&t=60171&hilit=resonance+noise

In idle mode (14dB), do we have similar issue in this built? i.e. if we open the side panel, will the noise level be even improved? I expect a different sound spectrum at least I don't know whether the similar issue exists in other system widely since I seldom heard similar complaints, generally people experiences much smaller noise with side panel closed.

netmask254 --


Actually this is not quite right. It's common in SPCR Certified Silent PCs which are built by commercial system integrators and offered to consumers. It's not so common in DIY builds we've done ourselves and detailed.

Your observations about the noise of closed vs open cases is not unique. I know I have mentioned air resonance many times in the forums -- any enclosed or even semi-enclosed space exhibits natural resonances, and if there is sound at or near those resonance frequencies, then they will be amplified. Here's an educational set of explanations and terms: http://hyperphysics.phy-astr.gsu.edu/hb ... on.html#c1

In essence, acoustic damping materials such as foam tend to shift (the frequencies of) and reduce the amplitude of resonances. Usually they cannot eliminate them, but they can help reduce the effect.

But if the noise level (amplitude) is low enough, then the effect is not worth trying to eliminate. In this system, the base noise level is so low at idle that the air resonance effects are insignificant. At full artificial CPU/GPU load, damping on the side panel could probably help reduce the noise a bit. It's even possible that the Nine Hundred case has less of the air resonance effect because it is so much more open than the Solo. Of course, if you use all the fans in the case (as intended), then resonance will be the last thing you need to worry about hearing.

Based on the power consumption (@stock speed) reported in this guide, is it possible to build a similar gaming system with my actual PSU?

Assuming you refer to Enermax Modu82+ 425W... your PSU fan may ramp up high at times but still, max load in a real game rarely got to 300W AC which is ~275W, or ~65% of your PSU capacity, so yes, it is doable.

another great gaming build by spcr! gotta like that overclocking as well.

nice job... i would have gone nvidia, just because of the better drivers, but i'm thinking that it would have generated more heat and more noise, with a bigger power draw.

I did another build of this machine, this time in a nzxt tempest evo build for silence

2x radeon hd 6870 both with scythe musashi

I had major issues with cable management as I couldn't fit everything behind the motherboard plus SLI cable routing is inherently difficult. if anyone has any suggestions on how to improve the cable management it would be appreciated.

specs are:

amd 965 be
mugen
8gb gskill 1333
scythe 4 fan controller
6870+musashi x2
64gb microcenter ssd
2tb seagate green drive
msi 890fxa-gd70

pic is attached, let me know any suggestions!

Your photo is blurry so it's hard to tell exactly... but it does not look bad. One simple trick is to bundle loose cables tightly with ties and tuck it to one side or another. This page shows examples of this w/ unsleeved cables inside a now ancient PC -- http://www.silentpcreview.com/article51-page2.html You can also use the space to the right side of the PSU to tuck portions of long cables.

I'm surprised you could not get all the cables to fit under the right side cover. It is deliberately sculpted outward to leave more room -- and this was very effective for our build. Maybe it's a matter of spreading the cables out more evenly & sticking them in place with tape or ties, to avoid bunching up on that side.

So is it quiet?

it's quiet for sure at idle with all 6 stock fans undervolted... (5 internal + 1 on the side window)

one thing I don't understand is why the musashi has two knobs for the fan controller... does anyone run the fans at different speeds? it makes no sense to me since they are either both on low, or both on high... I sacrifice a little bit of noise and set the window fan (which blows directly on the graphics cards) at full speed, and leave the FOUR musashi fans at about 20%. with all other fans at about 700rpm, the computer is what I would've called quiet BEFORE i learned about SPCR. now that I read SPCR regularly, there's no such thing as "quiet enough."

also, the top card is on average 10deg hotter which is not surprising due to how the musashi exhausts. with furmark extreme burning, the gpu temps in the setup above (fairly quiet) top out at about 81 for the top card and 71 on the bottom.

I'll try your suggestion about the grouping the cables together... it's a little difficult in this case as opposed to my other build in an antec 1200 (behind the motherboard looks like a war zone).

oh, additionally I read that the corsair power supplies haven't been considered SPCR material, however I have to say I don't hear mine at all, to me it's as quiet as the antec CP-850 I have in the other gaming machine I built.

also, props to nzxt for shipping what seem like pretty good to excellent quality stock case fans... IMO WAY better than the antec tri-cools that come with the 1200, and for the price, the tempest evo is a much better value. however, the 1200 feels more "substantial" and has a certain je ne sais quois that exudes a bit more quality.

that looks like a sweet build, and you have plenty of space between the video cards.

my sli rig also had a similar temp differential between the video cards, and i ended up going with the ft02 case, which helped a lot, but did not fully equalize the temps.

you could put a 25mm fan on the top card, but if you aren't overclocking the video cards, it's probably not worth worrying about.

Not true. All the Corsairs built by Seasonic are just as quiet if not quieter than the Seasonic models. But Corsair has some built by Channel Well Technology, which are definitely not as quiet, although the overall technology / build is pretty good. They are intermixed liberally without notice in the Corsair lineup, which is why Corsair gets mixed reviews by SPCR (and SPCR forum members). But you're safe (for noise) with the HX and AX series -- afaik, all of them come from Seasonic.