What i do not understand is why Silverstone release TJ04-E when they have the TJ08-E. They should have focused on improving the TJ08-E instead. Introduced better HDD cage and made out holes for 2x120mm fans instead of 1x180mm fan as an option as well as change the bad 180mm fan.

Cant help but think Silverstone was just lucky with the TJ08-E. One is probably better of modding the TJ08-E instead of hoping for improvements based on reviews these days.

I think the main source of noise are the GPUs. Would using the bottom and second front intakes be more beneficial?
The blower-style GPU HSFs don't benefit much from exhausts (I guess).

On another note, the 4870 gets a bit old. Wouldn't a modern 7950 with non-reference HSF be more suitable nowadays?

What's the target audience / build for cases like these, with 4x 5.25" and 9x 3.5"? Quiet file servers?

Why change the gpu? Just because it's older than you like? TDP and noiselevel of this card are well known, we would render all old tests useless or would have to re-test certain landmarks. We would give up continuity for being pseudo-modern.

And further i'm not able to see why using more front intakes would lower gpu noise. Adding a second fan creates more noise too and that the gpu will be less noisy just because we put more air into the front cannot be taken for granted.

Target audience: Enthusiast and wanna-be-pros, everyone with a little bit of knowledge would be able to find "better" cases for a given use.

Because the 4870 might not be representative of modern cards w.r.t. noise production and airflow requirements.

That's why I'm asking.

What temperature is best for peak HDD reliability? I thought it was quite high like 40*C but I can't find any references. But probably putting a heatsink and/or fan on a hard disk will make it less reliable than just letting it run really warm!

I think silverstone has focused on the wrong area and in the process, forgot it's own identity along the way.

The review states that the case focused heavily into cooling hard drives so maybe it is designed as a server case or something, but they could seriously just look at the fractal r3/r4 for hints imo. This is not an ideal case for silent folks.

Pity I really like the front panel, it's a shame it is so under performing.

The Silverstone comment links to the Google study, which is perhaps the most comprehensive and oft-referenced study on HDD temperatures and their link to reliability. The sweetspot seems to be between 37-42 °C (if I'm reading this right). The link in the comment: http://static.googleusercontent.com/external_content/untrusted_dlcp/research.google.com/en//archive/disk_failures.pdf.

Running too hot seems to quickly increase the variance in probability, though this can also be attributed to the smaller sample size, so it is not a straightforward case of "hot is better than cool". It does look like "hot" is "less lethal" than room-temperature levels of cool though - but the lowest I have ever seen any of my drives (sample size is 8 drives) is just above 30 °C, and they have easily reached close to or over 50 with no dedicated airflow. Getting them to the optimal range has always been straightforward: one fan, cranked to just the right RPM, airflow over the drive and not side-on. It is indeed not hard.

I applaud the move towards smaller cases, but this particular case concentrated way too much on (could even say overengineered) a minor aspect of the design. The side intake to me is as dubious as side exhausts: a noise hole right by the noise source, and in many scenarios one directed right at the user, usually sitting by the case's side.

I'm sure that's not the one I was thinking of or saw, but it is still helpful.

I would read it as 38-46°C for peak reliability. So then it would probably do more harm than good to actively cool your hard drives in the vast majority of typical usage scenarios (read: home/office client workloads with normal 5400/7200 RPM drives).

Right now, doing nothing more than surfing and watching flash videos, my Green is at 40 °C. I just now even left the PC to idle while I nipped off to get some groceries. I'd say active is the way to go, drawing from what limited experience I have; even the massive P182 with its separate HDD and PSU compartment with an actively cooled PSU needed a fan for the HDDs not to reach close to 50 °C. Heatsinks are pure, unadulterated overkill though, no question about it, unless combined with a noise-killing enclosure of course

I can't imagine a scenario where a fan blowing or drawing air over the HDDs would not be beneficial.

I'm a bit surprised at that for a green drive.

I've never seen drive temperatures go very high. But maybe they did and I didn't notice. My Samsung green drive (single platter 500GB in a scythe quiet drive), seemed to take ages just to get to 30°C. Right now I've got this Maxtor 200GB 7200RPM and that's sitting at 38°C. Right at the bottom of that ideal temperature range.

Blowing ambient air over the drives from an intake fan would push the drive temperatures towards the mid to low 30s I would expect. This is SPCR too. Elsewhere people probably run their fans at 12v and get drive temperatures in the high 20s.

I really appreciate the way manufacturers are in contact with computer hardware websites like SPCR.

I started paying attention when one summer, long ago, I most likely fried a 120 GB Maxtor in my extremely insulated Sonata 1. Those 100-200 GB drives would be close to the 35 °C mark, as I recall, though much in my setups and monitoring behaviour has changed since then.

This Green easily goes up to 45, even with active cooling, if placed in a restricted, sideways-airflow cage like the Define Mini's. And I don't even have a graphics card in here! Either the drive is hot (judging by wattage it's not), or the cooling is just not that effective. In fact I suspect, at the risk of sounding paranoid, that the space requirements of the bigger, multiple platters and their mechanics, combined with weight management and cost savings, may have eaten away at the capacity that modern drives have for cooling.

Since this is all so case-dependent, pardon the pun, and due to my personal history, I feel it's better to be safe than sorry with the active cooling. People do need educating on picking the proper fans, though - no harm done if they were to run a 650 or 800 RPM fan at 12 V or 100%.

PS. The Scythe Quiet Drive looks like it would be efficient at conducting the relatively small amounts of heat that HDDs output, at least moreso than air, so I'm not surprised if the Samsung stays low in there. It is enclosed, yes, but enclosed in an efficient conducting material with a large surface area - exactly what you need for passive cooling.

PPS. I should point out that the workspace has a high ambient of 24 to 26 °C (with a +/- 2 °C error margin, according to the $5 thermometer's specs), but that is to be expected in a small-ish space with powerful electronics.

Edit1: typo.

So, I have learned that they wanted to create a mid-tower size case that holds 9 3.5” hard drives and six 2.5” drives that does not hassle graphics card compatibility.
Why is it necessary to have so many drives? If the future is to continue as is, I feel that we will eventually see 3 tb drives at affordable prices. We will eventually get cloud computing. SSD will get larger sizes. If true, the role of the "hard drive" area in terms of size is to be relegated to possibly 4-5 hard drives maximum. Of course, this is for most personal users. The need for so many drive slots would make sense once it is used as a server (splitting many datas into many physical harddrives to maintain data lifespan). I would possibly be ok with that, but as said before, what is wrong with the fractal design r3/r4?
It is also pointed that it does not hassle graphics card compatibility, as in it will accommodate most, if not all graphic card lengths. Afaik, there are many graphic cards of various specifications that fits all, if not most cases, whatever the size. I'm not sure if this advantage is as great as it would seem.

In fact, I think they wanted to create a case that is mid-sized yet can be highly accommodating for any component of any size. For this, they seemed to have forgone some qualities that im sure some of us here value - hdd vibration stability, cooling and silence. I don't think the trade-off is worth it, considering the many cases on the market now.

What do you mean?

I apologize for my Danish idiom. I'm glad to see that manufacturers such as SilverStone are present in computer hardware websites such as SPCR.

Ironic? If silverstone was in contact with SPCR they would be able to make the ultimate ATX, mATX and ITX cases. Unfortunately they probably arent.

As far as i have seen most manufacturers don't recognise the group of quiet computing fans as target audience. The few that do are already comunicating with spcr, but not necessarily through our forum.

Most caveats on reviewed cases were comparatively easy to correct or could have been avoided in construction/design. So i can only conclude that the companies expect no additional value/sales if their product is engineered to be quiet and therefore are not willing to spend more in production or development.

Can't help but agree. I mean the "side panel bracing", as seen in mod form in this review, is a straightforward solution to a common problem that remains common only because the issue is not taken seriously by manufacturers.

The lack of flawless cases was why I stuck with my P182 for so long (two rigs' lifespans). For the longest time there was no value to be gained from rebuilding in a smaller case, as all cases seemed to have some unacceptable error or compromise in their design.

Supposedly, what Kuzzia is referring to is the comment from Silverstone at the end of the review.

The observed failure rate is base on average operating temperature. Without the raw data and some statistical analysis there's no way to tell if the different failure rates observed between 30°C and 45°C are statistically significant -- the differences maybe within margin of error.

Even if the failure rates are significant and real, the Temjin TJ04-E would need to have the ability to read the HDD temperatures from SMART and adjust the fan accordingly to keep the drives within the 35-40°C band in order to benefit from the observed lower failure rate.

Also the HDDs in the study are from a data farm and tend to be used much harder than ones in desktop systems. HDDs in home users' systems tend to idle a lot (remember the early WD Greens Intelli-Park causing SMART errors) and run at much lower temps.

Lastly the Google paper was presented at FAST07 and the study included drives that are 5 years old, there no way to tell how relevant data from 5-10 years old HDD are to todays models.

Personally I think citing the Google study is a bit of a marketing smoke screen from Silverstone. <G>

Ah! Why didn't that update get posted to this topic? I missed it completely.


How can we reach you (him / Silverstone)?

Even though it might not be the same person that wrote the comment in the review you could try to send a PM to SST Guy.

Myself and our product manager are long time fans of SPCR and thankfully Mike has been great to us as well when we needed opinions from him. Usually the review articles themselves already provide a lot of feedbacks so we didn't have to bother him too much.

We do try very hard to make SPCR worthy cases whenever possible and currently have more SPCR recommended and editor's choice cases than anyone else:

http://www.silentpcreview.com/article75-page5.html

We'll probably never make the ultimate case for everyone unfortunately, but SPCR's influence will continue to be a major factor for our future designs. In the case of TJ04-E, it had different design priorities that aren't as SPCR focused, so like Lawrence said in his conclusion, we can't win them all!

Thanks for making your first SPCR post a SilverStone one, but I think you've misunderstood our intention here!

The comment that was later added to the TJ04-E article was actually sent to SPCR before they reviewed the case. It wasn't really intended for public consumption but I think it's fine now that SPCR decided to post it. Other than this and in similar e-mails we've sent to other reviewers, we have not used the Google study link in our website or in other end-user specific marketing materials.

When we first showed prototypes of TJ04-E/KL04 last year, we noticed quite a number of resellers/distributors (and reviewers) were concerned about hard drive running hotter due to their design of tightly packing hard drives together. So we used the Google study as a way to show them that hard drives don't behave the same way as CPUs or GPUs do (which run better and more reliably the cooler they become), running hard drives cooler doesn't provide the same benefit. We simply wanted to prove our design for TJ04-E/KL04 won't cause hard drive to run worse than in traditional cases that focused their primary intake fans over the hard drives.

Have these ultimate cases been defined already? Different (SPCR) people have different requirements and wishes that might conflict.