Project: Intel White Box

[Vapor]

Just getting started on a new case project....well, it's been in planning for a few months but is just now starting to pick up steam.

In the next week or two I'll be taking delivery of a custom Mountain Mods case based on their Duality. The Duality is a two-system 18x18x18 case that packs a decent bit of airflow and organization. Overall a good case, but it just wasn't *quite* right for me

Breaking it down, the duality has a few basic problems as far as I'm concerned: 1) 80mm fans are undesirable (and I have no knowledge of which 80mm fans are the ones to get),
2) I have no use for four PSUs,
3) 6 front 120mm fans for intake as opposed to PSU and 4x80mm exhaust is an airflow nightmare,
4) the PSUs on the top really weren't 'my thing'...I've had a G70 for one of my systems for awhile and just loved putting my PSU on the lower spot

So...what did I come up with? Well, fortunately one of the sample designs was exactly what I wanted!



Side panels will remain solid (window and fanless), but for the top, I came up with this:

(this side [bottom of the image] faces the rear of the case)

The left system will have the CPU above the GPU while the right system will have the GPU sunny-side-up.

Now, what will I be putting in this?

Left system:
QX6700, XBX2, 2x2GB of G.Skill, X1600Pro with an Accelero S2 passive, TRUE + San Ace 1011, and 5HDDs

Right system:
E6700, P5W DH, 2GB of some RAM, some DX10 GPU + HR-03+ with a San Ace 1011, IFX-14 + 2xSan Ace 1011, and 2HDDs.

Each system has 1 optical drive. The QX6700 system is a dedicated encoding machine that also doubles as a network server (1.7TB or so between the 5HDDs), the E6700 system will be a gaming machine.

You might have counted the number of fans so far: 12x120mm external and 4x120mm internal...yeah, it's a lot. And it will be quiet The San Ace 1011s will run 23.9/7 at 5V with a switch to put them to 12V if I ever need to. The external fans will be a mix of 10 Yate Loon SLs and 2 Yate Loon SMs. They'll be split evenly over the two PSUs and two fan controllers.

How will airflow go? Backwards. That's right....I don't want air coming into the case, warming up over the HDDs and then slowly meandering toward the CPU and GPU. I want the all the air to go to the most important part of the system first As such, the fans on the mobo trays will be intakes, as will the 4 fans on the top of the case. The six fans on the front of the case? You got it, exhaust.

Since the two systems have very distinct priorities: left needing better cooling on the CPU, while the right system should have the GPU being the priority to cool...the top intakes are perfect as they'll aim virgin air to those components first

Why is this project called Intel White Box? Well, the case is all-white and has Intel components in it....1+1

Why is this in air cooling? I'll be doing a minor study on positive/negative pressure, local airflow around components, and system noise vs. performance Besides, I'll have 16 120mm fans in the case, that's air cooling

Over the next few posts/days/weeks I'll be adding more progress as different parts come in and I prep them for installation.

Next post: prepping/modding the fan controllers
x2

[Vapor]

Alright....one thing I really wanted to grab by the horns in this build was fannage. I wanted to keep it quiet with 16 120mm fans, but not get dinky 800RPM fans...no easy task!

I've always had rheobus fan controllers, which are nice, but have a few problems:
1) No 'settings.' i.e., you never really know what voltage your fan is doing unless you have a DMM hooked up, there's no semi-arbitrary 1-2-3-4-5-6-7-8-9-10-11 numbering either Just a knob that turns...
2) LEDs suck.
3) Finite loads. i.e., you can't hook up two high powered fans on one circuit without overloading it
4) Droop. 11.3ish volts when you want 12? Yucky-ucky.

Since around May I've wanted to build my own fan controller...a switch-based fan controller with 0V, 5V, 7V, 8.7V, and 12V settings. A switch-based fan controller would not have ANY of the aforementioned weaknesses....and with 4 different running settings, it would also have enough variability. I thought the idea was perfect

I knew it wouldn't be easy, considering I'd have two different feed voltages (12V and 5V rails), with four different 'ground' poles (ground, 3.3V, 5V, and open). About a month later into the planning, I realized I needed two

Then about another month later, I realized I'm not so great at DIY, and I still had almost zero luck finding the ultra-complicated switch I needed. I moved on trying to find pre-made switch based controller.

Unfortunately, all the switch based ones I found switched between a closed (normal) 12V circuit and a circuit with some resistance via a resistor/transistor pair. This still had a finite (and SMALL) load, damn LEDs, and droop on higher loads. And only two settings still

So I was stumped for awhile....

Then I found the Zalman ZM-MFC1.




It has two hard-wired 12V/5V/open switches (great) and four rheobus channels that are fairly weak (meh). It's also butt ugly with the graphics, LEDs and clear plastic. It's rheobus does have one GREAT advantage over the Sunbeam's: the minimum setting is ~5V....so it does actually have two distinct settings, as well as a variable middle-ground

And on further inspection I noticed something slightly bothersome: the 12V/5V/open switch doesn't have a 3-pin connector....instead needing bare wires. Not too difficult to cut and strip a wire, but I'd rather not.

Fortunately, some quick thinking has lead me to a few solutions to most of the problems

1) Switch connection is fixed!

Performance-PCs sells great sleeved wirings, and after a bit of digging their website, this fan adapter became something of import to my plan: http://www.performance-pcs.com/catal...oducts_id=3596

So in addition to my two Zalman MFC1s, I bought four of those

And ta-da! Stripped off the connector on the 2-pin side and they work well Highly recommended!


2) The ugliness is fixed!

Unfortunately no retailer has a fix for this: the fan controller is butt ugly.

Fortunately, I have a fix

You shake your hand for a couple minutes, white stuff comes out, let it dry, and you have a white fan controller

BTW, I was talking about white spray paint, perv.

So I bought some of that Krylon Fusion spray paint for plastic....

took off these pieces (from each controller):


sanded off the ugly arrows and words on the front of the controller, and had my way with them with the Fusion paint...

I must say I liked the results


I even did the back of the plates


2.5) Requisite PCB shots!

Something I learned when getting into computers....when you have a PCB of any sort, you take pictures of it and post it

Front:


Back:


(of note is that although Zalman says only the blue-socketed headers can do RPM sensing, all the circuits are identical, just the wires they include are different...cheap company)

3) Those damn LEDs....hmmmm, LEDs look good when OFF!!!

I was originally going to pull the LEDs (again, I hate LEDs....my monitor has 2.3Million pixels, I don't need 4 more on the front of my computer, thank you), but the way they're installed would have made that difficult....and I found no good filler for the plastic panel, so that was off. Then I considered just leaving them....but they FLICKER when you undervolt the fans! So in a move of desperation, I shorted one out while I was running a fan....and it worked! The LED went off, the fan kept going on Now all 8 LEDs on the rheobus are shorted....haven't tried the 4 on the switches, I'm fearing they're hard wired, we'll see.

Next post: the wheels on the bus go round and round? Oh wait, those are fans....a lot of them!

[Vapor]

Okay...fans....I have 12 external (case) fans, 1 for my Ultra-120 eX, 2 for my IFX-14, and 1 for my HR-03+. So that's 16. I happen to have a little experience with fans and I know at least this much: when you have a lot of fans, some will be bad apples.

I also know a few other things: noises you can't hear are often the most annoying. May sound like some cryptic thing you'd find in a fortune cookie, or you may think I have voices in my head, but I assure you neither are the case.

What I mean is this: we have cases for various reasons...among many reasons: animal/kid proofing, aesthetics, portability, and noise proofing. When a fan makes a slight murmur or a faint click, chances are the noise will not leave the case. You will not hear it from a few feet away through a metal box. Granted, any sort of loud click or scuffing you'll hear, but minor noises can be dismissed. What is of large concern is when the fan impacts the metal box.

We've been getting around that with silicone gaskets (which I will be using), silicone screws, silicone sealant....just about everything silicone except Pam Anderson. But the root cause is the fan and is actually fairly easy to remedy. The remedy also allows for a build with fans that don't scuff, murmur, or click in the slightest

1) The fans

Here are 14x Yate Loon D12SL-12s, 4x Yate Loon D12SM-12s, and 4x Sanyo Denki "San Ace 1011s"


They're the principle fans I'll be using in this build, and for good reason:
1) Yate Loons (REAL ones from Petras) are VERY good fans. They are not the best, but they're damn close. They come with flexible wiring options and have a very plain all-black design (except the stickers). Oh, and they're only like $3-4 a pop.

2) Yate Loon SLs push their rated 47CFM (actually, a tad bit more on average) and are fairly quiet. At 5V, they're well below the noise floor of my dorm room (heck, at 12V, a bunch of them are still below the noise floor when I have my A/C on).

3) Yate Loon SMs push well below their rated 70CFM (the highest I've ever seen over 5 different samples is 55.1CFM IIRC). Their CFM levels (54ish) are very close to the SL's 47CFM, which may scare some people away, or many just may find them irrelevant. However, they do have a purpose for my build: balancing airflow at my various voltage 'presets' (12V, 10.5V, and 5V).

4) San Aces are just incredible at 12V and 5V, but expensive at nearly $20 per fan. Hence only being used on the HSFs.


2) The fans need some work.

First, Yate Loons have a lot of wire connected to them....I'll be using 3-pin connectors throughout, so I can ditch the double-molex. Also, even up to the 3-pin is fairly long. So while I had plans to sleeve (and would still like to), I'll only be braiding the wires for now. It's not that big of a deal, as I have no windows in my case, but it's something I set out to do and I've bought a lot of sleeving materials and even had Alex from Petra's Tech Shop dig in his warehouse for a length of very narrow heatshrink as the included wasn't shrinking very much. I feel I owe it to him for being such a good sport and to myself.

In the meantime, I've braided them (and lost an SL and a San Ace in the meantime ).

(the Yate on the right in cootie-land is the one I messed up on the wiring and do not feel comfortable putting into a system 24/7).

I've also removed all the stickers (front and back) to make the fans all-black. On the back, in order to seal the motor, I've applied black electrical tape.

The San Aces, other than having their wires braided, are unchanged (internal only after all).

Now, I started with 18 case fans, (down to 17 now), but I only need 12. Why is that?


3) The Cure

Let's face it, fans aren't perfectly identical, some exhibit unwanted noise, others just don't spin at their rating, and some are just perfect. Real Yate Loons, although inexpensive, are very good fans and typically are very consistent.

I also have the luxury of being able to test fans empirically.

All things considered, empirical dBA testing was out of the question for me at this time, but CFM testing is easy and my ears are as trained as anybody's. Plus, I have another test up my sleeve.

First, I had to identify what made a good fan and where I wanted my fans to 'run.' Ultimately, if you plan on running a fan between 5V and 7V all the time, why should you care how it performs at 11V? No need to really. May be nice to know I suppose, but it's not critical. I identified the three voltage points my fans will always be running at: 12V (12V switch setting), 10.5V (rheobus max), and 5V (5V switch setting AND rheobus minimum). I really doubt I'll ever have a fan at a middle-ground setting on the rheobus.

Next, because all the stickers were removed on the Yates, I had to label all of them at the connector. Fortunately I did this before I removed the stickers though As I said, I lost an SL, so "L7" is absent from testing.

How I tested:

1a) Does it startup on the 5V switch setting?
1b) Does it startup on the (5V) rheobus minimum?

2a) CFM at 5V (switch)
2b) CFM at 10.5V (rheobus max)
2c) CFM at 12V (switch)

3) Subjective noise rating (letter-grade style) on the rheobus (where more problems are likely to exist).

4) Vibrational tendencies (letter-grade docking).

Startup tests are easy....if I have to explain those, please leave XS.

CFM tests, I used my very familiar anemometer setup and got very consistent results. Old news. None of the voltages were checked, as ultimately I won't run it that way....I just used the 12V switch (12V rail), the 5V switch (5V rail), and the rheobus's max (on the same channel for testing).

For the noise tests, both close-range and distant subjective noise testing was done. I listened for both unusual noises and amplitude. I also tried to target any sort of noises that may become a problem in the future.

Vibrational testing I thought was rather creative. First I'd stand the fan up so it was only standing on a 120x25mm side with the wire sticking straight up. Then I'd torque the wire so that it pulled on the fan slightly (wanting to twist it) and then I also aimed the fan to blow against the torque (helping it twist even more). I'd turn it on for 12V and watch the fan crawl across my desk. Every fan moved across the desk, but the ones that veritably shimmied or chattered their way across were the ones I was looking for.

Findings: only one of my fans exhibited any sort of extraneous noise. "M1" is a proper nomer for what it did to my ears. Okay, it wasn't so bad, but among a bunch of straight-laced fans, this murderous little guy was a standout. Only one of my fans exhibited any sort of chatter. "L1" was the problem fan and I have no joke for it. Interesting that both my "1" fans were the only big problem ones (numbers were arbitrarily assigned).

Even the slightest shimmy or murmur was docked considerably (2/3rds of a grade). If I even thought a fan was slightly louder than it should be, I'd dock it either 1/3rd or 2/3rds, subjectively.

Here are the results (red = will not be used, green = a 'go', yellow = definitely good, not sure what I should do with it):


Now that I have my fans selected, the next major update will be the airflow layout. After that, the grand finale: the case arrives and will be used.

[Vapor]

Oooh! Another!

[Vapor]

Last one, I swear

[Spawne32]

Reserved.

[Gautam]

Reserved.

[for mad props whenever this project is done ]

[Movieman]

I'm not waiting. Excellent project and I love the white and what you did with the controllers!
Also a good read, like a friend telling a story..

[Spawne32]

I'm not waiting. Excellent project and I love the white and what you did with the controllers!
Also a good read, like a friend telling a story..

Gah, thats almost as bad as my mom telling me my car looks "cute"

[Movieman]

Gah, thats almost as bad as my mom telling me my car looks "cute"

Ok, your car blows... j/k...j/k...j/k
seriously, I like cube cases, I've been wondering what to do with a high amp fan controller and I like articles that are written smoothly with no BS or " this fan controller could handle point 02649735v more than the other 32 tested.."
Like a good book, it flows from post to post..

[Spawne32]

Ok, your car blows... j/k...j/k...j/k
seriously, I like cube cases, I've been wondering what to do with a high amp fan controller and I like articles that are written smoothly with no BS or " this fan controller could handle point 02649735v more than the other 32 tested.."
Like a good book, it flows from post to post..

i usually choose a case that flows with the rest of the room, in my case, my room has no flow, so i just go with whatever works.

[Vapor]

Just started sleeving my four San Aces....not going so well.

1) my mom's hairdryer doesn't get hot enough to really shrink the heatshrink
2) I've never done it before...it's not as easy as it looks!
3) I tried to get away w/o zipties

Any ideas for a better heatgun? Like a household thing....lol

EDIT: Gah'lee, I'm REALLY a stupid! I used the wrong sleeving....lol....heatshrink still ain't shrinking much though

[Spawne32]

Just started sleeving my four San Aces....not going so well.

1) my mom's hairdryer doesn't get hot enough to really shrink the heatshrink
2) I've never done it before...it's not as easy as it looks!
3) I tried to get away w/o zipties

Any ideas for a better heatgun? Like a household thing....lol

4 dollar butane powered aim a flame thing at walmart, make sure you hold it far enough where it is just putting heat on it, not enough to light it on fire

[Vapor]

I'm not sure if it's the heat or the heatshrink....basically, I can shrink it fairly easily, but not enough.

[lad]

Hey, Vapor...

That XBX2 can drive 1.5 amps off of the two variable-voltage 3-pin fan connectors. The two 4-pin connectors ( 3 amps each) could be used with 4-pin fans. You could then use Speedfan to control your fans.

I'm using the two 3-pin connectors to drive 4 Panaflo 800 rpm fans (as they each draw only 180 mA. That way, I can run them really slow (Speedfan lies at 130 rpm) and inaudible, until I load up the 6600.

Just a thought!

[Vapor]

Lad....thanks but no thanks.

1) I don't plan on running Windows on my XBX2, ever (and no, I don't run crappy Linux either....I really, really like Vista, but it's not for this machine--partially why I have two machines in one case ). Speedfan runs on Windows.
2) Software and I never get along...if something can be done in hardware *kinda well*, I'll take that over software doing it 'amazingly' (because it isn't amazing and usually stops working for me)
3) Still really no control. RPM sensing circuits are neither accurate nor consistent in my experience on the P5W and XBX1.
4) It requires a 1-2 min waiting period for the the OS and software to start before all the fans spin down.
5) It's PWM....
6) Wiring mess, I'll have 7 fans to power on the XBX2's system and 6SATA devices....I want wires as far away from that board as possible and that wouldn't help.
7) I don't want more noise as I load it...I want constantly semi-quiet except in rare cases when noise is 'acceptable.' Hardware-based control can hold a setting ad infinitum and allows me to individually tune without interrupting or stopping what I'm doing.

Sorry if that seems harsh...but the on-board fan headers were never even a consideration (for more reasons than listed too...).

[cozwin]

Lad....thanks but no thanks.

1) I don't plan on running Windows on my XBX2, ever (and no, I don't run crappy Linux either....I really, really like Vista, but it's not for this machine--partially why I have two machines in one case ). Speedfan runs on Windows.
2) Software and I never get along...if something can be done in hardware *kinda well*, I'll take that over software doing it 'amazingly' (because it isn't amazing and usually stops working for me)
3) Still really no control. RPM sensing circuits are neither accurate nor consistent in my experience on the P5W and XBX1.
4) It requires a 1-2 min waiting period for the the OS and software to start before all the fans spin down.
5) It's PWM....
6) Wiring mess, I'll have 7 fans to power on the XBX2's system and 6SATA devices....I want wires as far away from that board as possible and that wouldn't help.
7) I don't want more noise as I load it...I want constantly semi-quiet except in rare cases when noise is 'acceptable.' Hardware-based control can hold a setting ad infinitum and allows me to individually tune without interrupting or stopping what I'm doing.

Sorry if that seems harsh...but the on-board fan headers were never even a consideration (for more reasons than listed too...).

linux and xp are "crappy"? but vista isnt hehe ok mate
but good job like the fan controller aswell.

[Vapor]

XP is awesome...Vista is better, IMO....P5W DH system will be running Vista Business for sure. On any desktop where I'd do more than file-serve and encode, I'd run Vista or XP.

Linux...I won't go there.

[Omastar]

I'm not sure if it's the heat or the heatshrink....basically, I can shrink it fairly easily, but not enough.

Try a lighter. Just hold the flame at such a distance that it doesn't burn the wires/heatshrink, but provides ample heat. Worked great for me with zipties.

[Omastar]

XP is awesome...Vista is better, IMO....P5W DH system will be running Vista Business for sure. On any desktop where I'd do more than file-serve and encode, I'd run Vista or XP.

Linux...I won't go there.

What's wrong with Linux?

[Vapor]

Try a lighter. Just hold the flame at such a distance that it doesn't burn the wires/heatshrink, but provides ample heat. Worked great for me with zipties.

Cool...should be able to find some old lighters. Kinda surprised that'd make enough heat tbh (maybe my heatshrink is shrunk all the way and it's just not small enough?)

[lad]

Sorry if that seems harsh...but the on-board fan headers were never even a consideration (for more reasons than listed too...).

The truth is never harsh. Thanks for the background, I'm always eager for knowledge!

[BNDMOD]

OMG, really nice.
I have one of controll of zalman and it's great... 5 fan and 1 kit of neon

[Omastar]

Vapor, if the heatshrink has shrunk to its maximum, use a bigger ziptie.

[The_Beast]

looking good so far