I'm upgrading my old (and on its last legs) Athlon 64 system, and being sick of fan noise I'm trying to build a quiet system. Been lurking a while, first time poster - be gentle.

This is a general purpose PC, web browsing, video watching, low- to mid-end gaming (no MW3 or anything), and I'm trying to keep the build cost under $700, but can wiggle a bit higher if needs be.

I've also got some gear from my current PC that is fit to reuse, namely:
ATI 5770 passively cooled graphics card
1.5Tb samsung eco green HDD
An antec tri-cool 120mm case fan

I want a system that'll do the job (which won't need much, given that the current machine almost still does the job) but will also provide an upgrade path.

CPU: i3-2100T (I was weighing up a G620T/G630T, but these seem even harder to come by in Australia than the 2100T, which I'll have to order from Amazon as it is)

RAM: 2x4Gb generic (current system is 4Gb and that's barely enough, plus RAM is cheap)

SSD: OCZ Vertex 3 60Gb or a Crucial m4 64Gb (I have an SSD at work and have fallen in love with the speed)

Case: Antec SOLO II (I've heard good things, and I don't need a giant case)

PSU: antec high current gamer 520 (no doubt overkill, but not by much - I estimate my rig will use about 300W at full load. Under-utilised PSU will be quieter, I expect). I was expecting to set up the PSU to blow hot air out the top of the case.

CPU cooler: will start with the stock cooler, see how it goes.

MOBO: Z68 chipset. Trying to find one with fan control, maybe an Asrock Extreme3 Gen3 or a Gigabyte GA-Z68M-D2H or MSI Z68A-G45. Mobo USB3 header for the front USB3 ports is a big plus (putting the Gigabyte and MSI in the lead).

Fans: I'll probably reuse the tri-cool, although the switch control is annoying to use - honestly, who manually flicks the switch higher when the system is under load? The SOLO II comes with a fan, and 2 front fan mounts, so I'll probably set the tri-cool and the inbuilt fan to low, and get a mobo-controlled fan to blow cool air over the GPU. I'd like that fan's speed to be controlled by the GPU temp, but maybe that's a pipe-dream.

That's about it. Questions/comments/feedback/discussion welcomed, and you have my thanks.

You don't need to spend that much if you don't want to.

I don't see why you'd need a Z68 mobo. No doubt there are cheaper models which would suit your requirements.

The dual-core T models are poor value and especially pointless in your build. I think that the 2100T comes with the low-profile cooler which is not suited for your build by the way.
If you get whatever Pentium or Celeron you can get easily in your locale, you'll be able to upgrade later at moderate cost as fast Sandy Bridge CPUs should hit the used market in about two months when the next generation is released. Or you could wait longer until they're even cheaper.

I've never used a Solo. It's a big case by my standards.
It looks like the main thing you need to watch for is to make sure a case fan blows on the GPU. A slow SB dual-core with disabled graphics would not heat up much and the stock cooler at low speed, while not silent, would probably be less noisy than your PSU.

There's no way this will burn anywhere near 300W. The PSU is overkill and probably inefficient and noisy. But you don't have to replace it right away of course.

I have an ANTEC SOLO case and also an outdated Lian Li aluminum case. From experience the stock CPU cooler is not suitable for a silent PC build. I set all case fans to Low and the loudest noise comes from the CPU cooler. The MB turns out to be a big factor in my two builds. One is from 5 years ago and one is from 2 years ago and the older one runs very hot with southbridge + graphics card at around 50C. The newer MB is around 30C with a passive cooled graphics card. It makes a big difference on the CPU fan speed. The outdated Lian li plays a part as well, with much less airflow than the SOLO.

The ANTEC HCG PSU is probably not designed for quiet operation.

Personally i think the CPU, RAM, SSD are great choices.

Not too sure about the GPU fan config you mentioned.

You might want to wait for the new Ivy Bridge Motherboards. They permit you to slide in an SSD as cache for your hard drives.

You get the benefit of traditional HDD data stability coupled with SSD speed.

Agree with most of the others. The PSU is not geared towards silent computing but gamers, so it will not be that quiet. That also applies for the stock CPU fan. You might as well buy an aftermarket cooler when you're at it. If you don't want to spend much money, then consider coolers from the Arctic Cooling Alpine series. These are marketed as stock HSF replacements that are significantly quieter than the stock HSF. SPCR testing confirms this, they can be run very quietly. Also, I wouldn't consider buyting the -T series. They seem to be only artificially performance-capped CPU's. The standard models will run as efficiently, and those extra MHz will be appreciated when gaming.

About the motherboard. I personally find ASUS fan-controlling excellent. Via FanXpert in AI Suite you can control the CPU PWM fan VERY well. 3-pin chassis fans can be run as low as 40% (i.e. 5 V). I believe Gigabyte lacks in this particular area. The Pro-versions usually have two 3-pin fan headers and one PWM fan header. That's worth investigating.

Other than that your components seem to be fine. I'd choose the Crucial over OCZ since the M4 has a better reputation for reliability. The recent firmware bug seemed very simple and specific and Crucial has already released an updated firmware, so I don't believe the firmware bug will be a noticeable issue. The OCZ, however, has notoriously been known to BSOD, and if you value reliability (who doesn't?), I'd definitely recommend you to stick with the tried and true: Crucial M4, even if it's more expensive.

Thanks for everybody's replies!



The Z68 mobo does SSD caching of my HDD - hence my choice (cheaper than buying a 256Gb SSD, and almost the same performance for boot/app launch). As for waiting, isn't that always the case? If there is one thing I've learned from building my own PCs over the last 20 years, there is always something "just around the corner" that will either rock your socks off, or drop the price of whatever it is you were thinking of buying today. I've given up on that game, and approach from the position that a working computer on my desk for a couple of months is worth more than the savings or performance increase I'd get by waiting for that "next big thing".

Going with a low-end Sandy Bridge is a conscious decision to provide an upgrade path to an SB i5 or i7 in a year or two.



Can you provide some more detail on this assertion? I get that maybe the stock cooler isn't any good and I'd need an after-market model. Are you saying I'd be better off using stock cooling on an i3-2100 than buying an i3-2100T and getting a better cooler? Or are you saying the price premium ($10) and performance hit of an i3-2100T over an i3-2100 isn't worth it, and I should put that $10 into a better cooler for the 2100? or both?

Also Re: PSU, there is no PSU in the SOLO II, so I can buy a lower power one (400W?), although my concern is that at max load I'd be running the PSU at close to 80-90% (there was a site where I put in all my components and it spat out about 335W as max power), which means I'd need to spend more on a PSU that is quieter. (Plus a 400W PSU leaves little room to upgrade my GPU should I decide to go that way one day). Thoughts?

Anyone got any thoughts or experience on which particular brands/makes/models of mobos (Z68) work best in a cool/quiet setup?



I'm not married to this particular PSU, maybe I'll do some more research. As for the GPU, I'll be using the case fan to blow fresh air over the GPU (it's a Gigabyte 5770 silent cell - http://www.gigabyte.com/products/produc ... id=3584#ov).

One of the case fans on the SOLO 2 is right behind the PCIe x16 slot, which is one of the reasons I like this case. And the back-side output fan is right behind the CPU, so it'll help expel the warm air from the CPU cooler/fan - another reason I like this case. And the top-mounted PSU with a top-side exhaust vent means the PSU won't heat the interior - another reason I like this case.

Again, thanks everyone.

Thanks for that - fan control is something that is totally new to me, so I'm very much in the dark, and the various manufacturer websites seem to be quite silent when it comes to trumpeting these features, should they have them. Some don't even tell you how many fan connectors are on the mobo - just that they exist. And forget mention of controlling them via either software or bios.

Three strikes (you, HFat and tppmedia) for the PSU is enough - I'm definitely going to revisit my choice (and wattage - although that's a harder question - how much headway does one leave above full load, and what about future upgrades - another HDD, an upgraded GPU...?).

Why don't you read a review of the 400w Antec high current gamer PSU. Very quiet up to 50% load. I doubt your system will use more than that.

Thanks for that - I didn't even know there was a 400W Antec TCG model - it's certainly not available at any of the number of stores I searched online (here in Australia) where I'm going to be buying my parts.

I've almost settled on a Seasonic M12II-520W. It powerful, and it's quiet. Tried to find a recommended PSU in the 400-450W range, and they are just not available here.

Take a look at my signature. With a 65W TDP CPU and an HD 5770, I see 145W AC while gaming, 235W AC with prime95 and furmark going. With a 65W TDP CPU and an HD 5770. Translating the 235W AC at the wall to system load, it's 200W DC.

You'll see similar load with your build and a non-T part.

Like others have said, don't bother with the 2100T. You won't see a power savings at idle or at normal loads...and you get some extra oomph when you need it.

The Seasonic M12 and S12 are nice. As are the X series.

I have to say I'm shocked that your wattage is so low under load. I found a couple of websites that estimate power ceiling, and all have spat out numbers in the 350W range. And the Gigabyte website says 450W power for my 5770 - although that's clearly fanciful as the card uses a single 75W 12V plug, making a total 100W max for the card assuming it sucks max power from both the 12V plug and the PCIe socket.

The Seasonic X series look wonderful, but at $199 for a 460W vs $120 for a 520W M12, it's hard for me to justify. As for the 430W S12, I can get one for $80, and I just now found the spcr review of it. I never considered the Seasonic S series as it wasn't on the recommended PSU list on spcr, however it seems like it's a pretty good one, especially for the cost. And I'm pretty much convinced now that 430W will do the job. I've had a pretty hard time coming to that conclusion, because my current PC has a no-name 500W PSW and now dies when I tax the GPU too hard.... (the main reason for this upgrade happening now).

I've given up on the idea of the T-series i3, considering the feedback I've received here and my ongoing research. Thanks guys.

So as it stands, my parts list is looking like:
Case: Antec SOLO II (seems to be a good quiet case for a good price)
CPU: i3-2100 (cheapest LGA1155, nice wide upgrade path for the future)
MOBO: Asus P8Z68-V LE (Z68 gives SSD caching of HDD, USB3, Asus provide good PWM fan control, and the cheaper LX lacks the USB3 headers for the front ports on the case)
RAM: 2x4Gb PC1333 (Why buy 4 when you can buy 8?)
SSD: Crucial m4 64Gb (because I can't afford a 256Gb SSD)
PSU: Seasonic S12II-430W (seems to be a good quiet PSU for a good price)
Case Fan: Gelid Silent 12 (quiet, quality PWM fan)

[plus left overs from current PC:
ATI 5770 silent cell
2Tb Samsung eco green
Antec 3-speed case fan (not PWM)
]

I'm going to start with the stock cooler, and upgrade later if CPU fan noise is an issue - I like the scythe ninja 3, and at 160mm it _should_ fit inside my case. I'll also look at upgrading my existing and the bundled case fan if needs be - will just have to see how noisy/hot the system gets with the two fixed-speed fans and the third PWM fan.

What do you all think?

The case fan that comes with the Solo II is pretty darn quiet. Given your system specs, you probably don't need a front intake fan. You might consider strapping a slow rpm fan to your passive 5770. Might as well build 'er up, run some stress tests and then modify as needed.

If you do have the spare cash, consider moving up to the 128GB M4. Write speeds are a lot higher. 4k random write speeds seem to be the weak link in many SSDs and are the most common activity.

Regarding your old PSU: Older ATX specs had more amperage going to the 5V and 3.3V rails. However, power management architecture changed over time. So today, PSUs supply up to the rated wattage on just the 12V rail. The mobo and addon cards then use local voltage regulation to get the voltages they need. That said, your PSU may not supply enough current via the 12V rail.

This is one reason why ATI and NVidia overstate the PSU wattage required. They don't know what ATX spec your PSU meets/how old it is. It's easier for them to overstate the total wattage and hope there's enough amperage on the 12V rail, than to deal with customers complaining about their PSU not working...

All the Ivy Bridge chips use about 4 watts at idle. The only difference between the 2100 and the 2100T, is that the 2100T has, in essence, a speed governor on it.

Almost ll of the time of a CPU chip is spent at idle or near idle.

That brings up an interesting point that I'd neglected to mention amongst the bigger issues of "what should I buy"... namely, that of a positive pressure case vs a negative pressure case. With two front intakes and a single rear exhaust, I'd be in positive pressure land without doubt. With one of each, could go either way depending on what speed I set my fans. And obviously with no front intake fans, I'll be in negative territory. So, I ask those with more knowledge and experience on this topic - what difference does it make either way?

One think that strikes me as obvious is that of dust control. My PC sits on the floor, and the dust that builds up in its corner of the living room can be a bit nasty. With positive pressure, I would expect that most of the dust would be caught by the filters on the front intake vents. And I would expect that with negative pressure, dust would more readily enter via the various gaps and holes in the case.

I would also expect that with positive pressure and front intake fans, I could more easily direct cool air flow over the fins of my passively cooled GPU - effectively strapping a fan to my GPU, but making use of the mobo PWM fan control to keep things cool (I hope).



Few things would please me more. I'll see if I can convince the wife to let me spend an extra $90- although as it is I'm looking at about $670 between two stores, so the extra could be a hard sell. The small SSD is one of the reasons I'm going the Z68 chipset. If I went a bigger SSD I could be 100% confident I could install windows and all programs on it for the forseeable future, and could get a different chipset.... Choosing amongst the various incarnations of a particular chipset across multiple vendors is enough of a nightmare - the thought of adding different chipsets to that mix makes me despair!

The Z68's SSD caching is called Intel Smart Response. The SSD caches reads (and writes in high performance mode) to a raid array. You need at least to hard drives for a raid array.

Another one might be that nVidia certifies PSUs for SLI. No doubt money is involved so the kickback is overstating the necessary wattage by ~50% to help with sales of high wattage (high priced) PSUs.

Are you sure about that? From what I heard a single HDD will do, no need for an array. That said I don't think much of SRT, too much complexity and reliance on driver support.

I am unaware of any benefit to negative case pressure.

I read that you need one HDD and one SSD... but you have to select a configuration option titled RAID to configure it to work.

Can you share more of your thoughts on your caution... and perhaps what experiences have contributed to your caution.

I am struggling with the need for speed vs a Seagate whitepaper cautioning that data is more easily subject to corruption over time on an SSD. Given how flash memory works, as well as my own experience with SSDs, that caution sort of makes sense to me. SRT seems to represent an optimal solution to this problem.

Here's a nice review on SRT at Anandtech. Tradeoffs:

- Z68 and small SSD = can overclock, use SRT, and Quick Sync (the latter via Virtu). You may not have any apps that make use of Quick Sync. Z68 boards are pricey.

- P67 and larger SSD = can overclock. Mobo cheaper than Z68. Don't have to muddle with Virtu. Can load all apps on SSD and media on HDD.

- H67 and larger SSD = no overclocking, has Quick Sync (via Virtu). Mobo cheaper than P67. Can load all apps on SSD and media on HDD.

One funny quirk about having all apps on SSD and media on HDD for me. iTunes will cache a song or more in RAM when it comes up on the play list. The play time is long enough that my HDD spins down between songs. When iTunes is in random play mode and needs to fetch another song, the spin up time of the HDD noticably affects the next song's start time. Seems like the stupid software should look ahead and pre-fetch...

If you are in a very dusty environment, I'd see why you'd want positive pressure. Remember to take into account the PSU and the Solo II top grille for it's air access.

Requirements from the user guide:
Intel® Z68 Express Chipset-based desktop board
System BIOS with SATA mode set to RAID
...
12. Select the drive (or RAID volume) to be accelerated.

Sounds like it'll work with single disk. Never heard of a single disk in Raid mode before.

That's what I read on the Internets. But it sort of makes sense. The instruction says selecting something... it refers to RAID.... but what you have is an HHD and an SSD hooked up together. So it is sort of two drives coupled together in a way that you direct a read or a write to both of then and they figure out what to do with it between themselves.

That's my understanding too. I have a mate that uses a small SSD as a cache on a Z68, and he doesn't use RAID.

"In every other way, using SRT on the SSD 311 Series is a snap. Just install the software application (likely on your motherboard's driver disc), point it to the SSD you want to use, then select either Enhanced or Maximized as the type of caching. The former mirrors data to the hard drive to provide a gentle speed increase, but offers protection in case the drive fails (think RAID Level 1); the latter, as with RAID Level 0, supercharges your system by caching data, but you may lose whatever is in its cache if a problem develops with either drive. You can switch between the two methods whenever you want, but doing so scrubs the drive. (Interestingly, when SRT is activated, Windows only recognizes a single drive rather than two.)"

http://www.pcmag.com/article2/0,2817,2387384,00.asp

The reason I don't like it is that it's a bit like RAID-0, if something happens to one drive then it's a good chance that the data will be irretrievable. It's also done in software via the Intel driver, so I don't think it's possible to use a Live CD or similar to access the data in case something goes wrong with Windows. This is just my personal feeling, I'm sure there's nothing wrong with Intel Smart Response per se.
I would keep it simple. OS and apps on the SSD, storage on the HDD (preferable on a NAS tucked away somewhere).

It's also typical Intel behavior to artificially limit this to Z68, pisses me off.

I think maybe you are right about keeping it simple.

I don't know about you guys, but I regularly back up my data, to another PC and to an external HDD. If my SSD cache died, and this took my system with it, my data would be safe. If I were to buy a bigger SSD and separate data and apps, then sure my data is safe if the SSD dies, and my system is safe if the HDD dies, but in both cases there is still a replacement/recovery effort required.

In my mind this SSD cache business _is_ keeping it simple. No worries about trying to decide what apps/data goes on the SSD and what doesn't. No worries about the SSD getting full of useless or seldom used stuff.

That is what I thought originally, but:
"SRT is managed by Intel Rapid Storage Technology software version 10.5 or later[2], and implemented in its device driver and the Z68 motherboard's firmware (option ROM). It is available only when the (integrated) disk controller is configured in RAID mode (but not AHCI or IDE modes) by implementing a style of RAID-0 striping. Write-back (Maximized mode) or write-through (Enhanced mode) caching strategy can be selected by the user. The maximum utilizable cache size on the SSD is 64 GB. Caching is done at the logical block addressing (LBA) level, not the file level.[3]" Wikipedia
http://en.wikipedia.org/wiki/Smart_Response_Technology

It sounds like it is using a RAID type of striping driver. I don't hear much about this feature... whether it is reliable or not reliable. I also keep my data on a separate hard drive and back it up. So the likely worst problem I will have with an SSD is to have to reinstall the operating system and the application software.

But once you start parsing data by logical block through some RAID type of driver it scares me a little... it makes me concerned that I could get my data scrambled and not even be aware of it for a while.