Project Goliath - 70+ pounds of H2O goodness!

[starbuck3733t]

After bouncing e-mails back and forth with Ben at Mountain Mods about getting my res laser cut, it's in progress... so that and the glue-up of it will likely be the next update.

The machinist, Chris, who I work with at the MRI was in a bad accident and will be out for quite a while, so no mill action for a bit. he's doing well, and I'll miss talking with him... I've got a lot to learn from him.

On to the show.

Part I: F is for Flow

It so happens that the eheim 1250's outlet is 3/8" NPT, so it seemed logical that the barbs I ordered from mcmaster would fit. Well... they went in about 2/3 of the way and stopped. The thing about NPT is that it's tapered so the threads near the top of the hole are the farthest apart. If it wasn't initally tapped for the proper depth of fitting, you get my problem. Luckily, I own a full assortment of NPT taps! Away we go!


No problem. Note the evil huge 3/4" ID barb on the inlet. the 1250 gains a lot more head when it's fed with a 3/4" ID inlet. (I learned this a long, long time ago on procooling - the pump shot water ~2 feet in the air with a 1/2" inlet, and ~6 feet in the air with a 3/4"!)


(Points at 6/8 tubing) "You call that tubing" (pulls out 3/4" ID tubing) "Now this is tubing". The 1/2" ID stuff may or may not get used, I accidentally bought thin wall. if it works, fine, but if it doesn't I'll trade somebody at the lab for some 1/2"ID with 1/8" walls.


Part II: Boiling Black Cauldron

Black Rit dye, meet pot, pot, black rit dye. Thanks to zittware and his project, I decided to try out dying my polypropelene fittings black to match the rest of my case. Word to the wise - this stuff smells like nasty feet when it boils, makes a huge mess, and will stain anything it comes in contact with. Don't do it with the wife's good pots. I pulled this old thing out from my bachelor pot collection...


Stainless steel tube (scrap from another project) bent to hold the fittings up from the (very hot) bottom of the pot.


In we go..


Black cauldron (peeee-uuuuuuuu this stinks!)


Before and after. The odd thing is that they came out a bit green. Which is wierd, considering the dye is black and even diluted the dye is dark purple. I'm not sure if I like it. What do you guys think?


Dark fittings...


...or light fittings?


Post your opinion, I'd appreciate it.

Until next time, kids.

[starbuck3733t]

Part I: When bad things happen to good motherboards - northbridge hoops tore out

One morning I go down into my workshop to grab a few things before I head off to work, only to discover that my northbridge block (aquacomputer twinplex) had fallen off the board. Upon further inspection, it had ripped out and took the mounting hoop with it!!


This where it used to live.


So after some solder-sucking action, I was able to force a new hoop, made out of the legs of a safety needle, through the board. I soldered the new hoop in place and left the excess sticking out and up like so.


Then I slid a piece of prototyping circuit board over the legs (copper contact side up, of course.)


Then I folded the legs over on themselves (which required a bit of force since they're steel) and soldered them closed.


There's no FRIGGIN WAY that the block can pull these out.

Part II: Better living through acyrlic (reservior)
This is not your typical reservior. It needed to supply a 3/4" outlet barb for my pump, and combine the 4 smaller loops. It mounts vertically, so in most of these shots, it's sitting on it's front - the part that faces the front of the case.

The two sides (front and back) of the reservior that I designed a few posts back, with the 'inspection port' covers placed.


IPS Weld-On #4: Strong in 15 minutes, full cured in 48 ours. Awful smell.


Syringe w\ 18 gauge needle to apply the weld-on. Weld-on is capilliary cement, meaning it draws it self into the joint through capilliary action. Remember kids, I'm a professional


All glued up. You can see around the edges where the plastic has melted and welded together.


The side wall pieces were laser cut, but the laser left a bit of a kerf, and as you can see the edges are not at a 90 degree angle from the sides, as indicated by the arrows.

[starbuck3733t]

Nothing a little help from the 1/8" 4 flute carbide endmill can't solve.






All true'd up now. Flipped them over and did the other side in the same way.


clamp clamp mushroom, mushroom.


Front edge lined up.


Injecting Weld-on into the joint.


Clamp another side on and tape it in the corner so it doesn't shift. Not a lot of pressure is required with weld-on, tape and a light weight or clamping pressure is plenty. No gorillas need apply.


Other side, more tape. The strips pull it toward the middle, while another strip holds it to the side. They fight against each other, and put downward pressure on the joint.


Tape down the end cap and clamp it on the sides.


The melted-like appearence of the joint once cured. Weld-on is a solvent-cement, meaning that it softens the surfaces and, as it dries/cures/evaporates/ the now semi-fluid surfaces are permanently joined.


Side 1 all done.

[starbuck3733t]

Post-hold-water-testing. No problems.


1" Diameter fill-hole cap from Mcmaster-Carr. 3/4" NPT thread, black polypropylene.


Center divider placed...


...looks a bit high, doesn't it? The water path, in green, from return to pump inlet. The space between t he top of the drain plug and the top of the divider is too small. The purpose of the divider is so that bubbles can't go immediately back into the pump intake. This res did a beautiful job of bleeding all the air out of the system in under 2 minutes!


Why don't we take off some material? Material to be removed is highlighted in green.


Halfway through. The endmill didn't have enough cutting surface to do the whole thing at once.


Action shot.


Smoothing the end, all done.


Shortend divider in place and glued up.


Top clamped on, no shortage of clamps here.




All done, just not installed. This is the "Front" because it faces the front of the case. A bit dusty in this shot... :shrug:


This is the "Back" - it faces into the main compartment of the case.

[starbuck3733t]

Front view: installed in the rear of the top-front bays of the case, looking back.


Back view: installed in the rear of the top-front bays of the case, looking frontwards.


Filler cap just barely fits vertically... Phew! :thumb:


Part III: Replumber's Crack
I've been itching for this since I ordered the Storm/G5... plumbing it, as well as my custom 4-way splitter, custom res, and custom-mounted Radeon twinplex block back into the loop for some massive parallel loop action...

Storm/G5!! (that's pure-silver (not copper) version.) We'll see more of it in a bit. This version has better performance and a different design from the Original Storm/G4 and Swiftech Storm/G4. Plus it's just so... damn... sexy.


Yummy.


Parallel loops are starting to come together. Heatercore is installed again, with foam tape around the edges to deaden vibration between it and the not-yet-installed-in-this-picture shroud. Huge thanks to Knipex, Dr. Fibbles and Pug/Infidel for their help sourcing some of the watercooling gear: Blocks, Tubing, Fittings...


Parallel loop fully plumbed! Don't worry about the ugly grey foam the pump was sitting on, it's temporary.


Black 8mm tubing courtesy of Pug/Infidel returning to the reservior. The loops split in the splitter (seen further down in this update and in previous updates) and then combine again here.


All parallel loops return-plumbed. Look really closely and you can see the one that comes up from the silentstar HDD enclosure. The fitting that the silentstar returns to was later replaced with a 90 degree fitting, due to the extremely tight fit between the reservior and the end of the fan shroud + fans.


Everything reinstalled: Radeon 9800 pro, 3com gigabit copper ethernet, Adaptec Ultra160 SCSI adapter, Pioneer SCSI slot-DVD, HP SCSI CDRW, 2x Ultra160 10K RPM SCSI HDDs.


Neat shot of the G5 amidst all the other various-size (6mm ID, 1/2" ID, 3/4" ID) tubing.


Shot from the back w\ everything installed.


Tight fit going back into the res! Don't mind the slightly-pinched looking line, it's just there if I ever need to drain excess off of the reservior.

[starbuck3733t]

Northbridge on custom super-strong mounts.


Splitter installed. It's mounted through the back plate of the mobo with an 8/32" hex-cap bolt. And no, you're not seeing things, it's not vertical. I mounted it at an angle w\ 1 bolt instead of straight up w\ 2 bolts to prevent the tubing from kinking. I had less room between the top of the splitter and the bottom of the radiator fans than I thought. No problem, though. 6mm ID tubes leave the splitter and go to the northbridge, silenstar HDD cooler, and radeon 9800.


All my front-panel connections bound together and installed: 2x firewire and 2xUSB 2.0.


Twinplex on Radeon 9800 pro. The twinplex takes up less room and puts less stress on the video card than the old danger den block ever did.


Massive 3/4" Inlet barb. Yes, it's WHITE, I know. Wait till you see the night shots in the next update and you'll find out why.


Silentstar w\ full ultra160, power, and water plumbed in.


Drain hose for the reservior. Thanks for the plug-and-cool unions Knipex!


The bolts that hold the reservior in. They don't screw into anything, but the go in and press on the reservior to force it against the side of the drive cage. That and sitting on top of one of the optical drives keeps it quite secure, although it doesn't move, even if you do remove the optical drive.


Temperature sensor board for Speedfan (MAX1668 based) installed w\ probes for reservior, GPU temp.



Part IV: It was a dark and stormy night

Midway into leak testing the loop in my lab, I noticed that there was water on top of the northbridge block... I blamed this on a bad seal on the twinplex. I got out the metric allen wrenches and tightend the top down a bit. It still leaked. I tightend it some. it still leaked. Well WTF?? I put my finger up under the storm's universal mounting plate to find out that it was the STORM that was leaking, not the twinplex on the northbridge. WTF *@#*@(%* I said!

It turns out that the problem seems to be that the o-ring groove on the bottom part of the storm was milled too deep, or the supplied o-ring wasn't fat enough. See how the o-ring doesn't sit up above the groove at all? Well, that's bad.


Not having any other o-rings (save for the ones for my paintball gun) around, I pulled out the silicon-II and smeared it into the o-ring groove to build up a bit of a base so the o-ring had something to raise it up, and something to seal against on the bottom.




Half an hour later the stuff was non-tacky, and I reinstalled the o-ring. Now it sticks up! that's GOOD. I reassembled the block, and have had no leaks since.



Part V: Overclocking again

Now that I'm done leak testing, I've begun stress testing my new 2.8C SL6WJ P4. Since I've still got my mobile 3.06 P4 to fall back on (which will do 3.56 GHz) I've decided to go whole-hog and break the 1.7V barrier on v-core. 1.9V set in bios yields 1.86 w\ less ripple than when it was set to 1.675. Seems the higher you set Vcore, the less it droops... This seems to be in line with the Vdroop mods for the IC7, wihch lessen the droop but raise Vcore. All of that leads to this:



That's all for this time, folks

[bigben2k]

Great work SB!

Are you pressure testing the res? (not that it's needed...)

[starbuck3733t]

Thanks ben!

I should have, but didn't have appropriate gear to do so. I had a pinhole leak that was quick-fixed with some goop. It seems to be solid now.

Odd thing was that it passed the BLOW HARD test. maybe it was a weak spot that took some time to open up... who knows.

[starbuck3733t]

Part I: Lighting placement and routing

Top white CCFL with Reservoir's UV inverter hidden up underneath the frame, and drain tube from res.


Bottom UV CCFLs, 2 of them. They're zip tied together, and then to the case frame. I discovered that with the way the doors attach to the case, that I can actually zip tie into the structure and not interfere with the mating surfaces.


Bottom UV CCFL inverter. I dare you to find the wire that goes back up to the top bay from the inverter!


Top white CCFL inverter tied to front of the frame, with the drain nestled beside it. Yes, the drain line is hanging out a bit, but it'll be pushed in by the side of the case when it is installed.


Rear UV CCFL inverter and a bit of the top white CCFL. The rear UV CCFL is sitting above the exhaust fan, you can just barely see the end sticking out above the exhaust fan.


Play spot the inverter. You can also see (more clearly this time) the top rear UV CCFL above the fan.


Part II: I'll shut up and let the pictures do the talking...
...with the occasional caption.








(and this is why I ran with the white barbs, they glow!)




(Looking down from the rear)




(Splitter)


(G5 in and out)

[starbuck3733t]

Reservior Lighting

From the front:


Dark shot. Silver things are the plug and cool G1/4 returns.


Closeup of the backside of the fititngs:


The light source:

[starbuck3733t]

Remember those touch sensor keys? well, at some point I had to actually put connectors on them so I could install them... which is what I was hoping to take care of tonight, but as fate (or my painting apparatus) had it, that was not to be. The bezel dropped off my hook and cracked the edge. Thankfully, with minimal effort and some of the IPS Weld-on 3 I was able to stick it back together. The curse of Zittware strikes again.

So, on with the show. Why did it take me a good 3 hours to do all 4 of these? Keep reading.


Part I: The monotony of it all!

Attach 82 ohm resistors to cathodes of LEDs.


Attach black wires (my color scheme... ;-)) to the anodes.


Attach wires to resistors.


Mark which black white does what, the cathodes were stripped on one end, and the tape helps me keep them paired up.


Sleeve.


Sleeve some more. All 5 wires coming out are sleeved, and the center wire (going to touch sensor itself) gets its own sleeve. I could have bundled it in with the LED lines, but why risk noise and false triggers on the touch sensors if you don't have to! The LEDs come out on their own short lenghts of sleeving, then another length of sleeving covers all 4 of the LED wires, which goes down to... (see next pic.)


The plug end. The single line has no plug on it because I'm not sure what I'm going to use to connect the touch sensors into the chassis control board.


Make sure everything still works (white LED in this case, but the blue works too ;-))


And 3 hours later I have 4 of them! Hooray. >_>



Part II: Paint until nightfall, that's when the bugs come out.

Top primered. I have to do a few more coats after the inital wetsand w\ 400 grit, but this is the beginnings of the paint job.


Had to really sand the crap out of the OEM finish on the bottom of the case, all sorts of scratches and gouges in the factory paint. Mostly due from me using it w\ no feet on it for several months last year (shame on me!)


That's it for now. Hopefully I can get the front bezel painted and the sensors intalled sometime soon.

[starbuck3733t]

Part I: "Industrial Black"
When this project was started in 2003 (yes, 2003, around February/March time, with the first photo-documented post being on August 31st, 2003), I conceived it as being a "black monolith" when off... In fact, the project was nearled called "Project Monolith", so it may not come to those who have followed the project since its inception to see a non-bling type paint job. Flat black was chosen to go with the industrial feel of the mod... after all, when was the last time you saw shiny gloss paint on industrial equipment? The shiny parts are usually on the inside, well lubricated, and working well together, and this mod is no different. From the polished gloss black shroud to the polished aluminum fans, to the gloss black and crinkle black powdercoating, this case screams industrial.

On to the show, already in progress...

The bottom has been primed and will probably get black paint tonight (shortly after I finish posting this )


Flat black: the anti-bling (a close-up of the top shell of the case.)


Again, with mesh integration into the top panel. You'll get an overall shot when it's finished drying.


The front bezel, same drill: FLAT BLACK, but will get the some gray plexi top-half, as seen here.


And a little trick I came up with to keep dust and other back-yard under-deck crap out of fresh paint. Put a box over it, but with the ends open so the solvents can still evaporate out of the paint.


To be continued...

[starbuck3733t]

Part I: Continued...
The paint continues to fly...

Completed top. Sorry for the blurry shot, but the light was fading fast, soon it would be no-paint time (due to bugs) hour was upon me.


Completed bottom.


Bezel...


...With smoke acrylic on sitting on top. The light does wierd things with the paint, and it's hard to photograph, so I shot a few angles.




A wise man once said "you can never have too many clamps." (it was dutchcedar, btw ;-)) I'm clamping it so the IPS Weld-on I use to glue the acyrlic together can harden.


Niiiiice fit.



Special Feature: Teaser of what's coming soon...

Touch sensor (and yeah, I need to windex the plexi.)


A bit closer.


Oooh...

[Karlmcwade]

WoW this is a truly great project well done

[starbuck3733t]

Thank you!

[starbuck3733t]

Part I: (the Paint Saga) Continued some more...
The top panel is 95% done, I've got to go shoot one or 2 more coats, but it's just to cover some unnoticeable-to-anyone-but-me thin spots. I'll fix it, because I KNOW ITS THERE... /anal-retentive.

Painted top on


Flat, baby. (My finepix 3800 has no aperture setting, hence crappy depth of field...)


Sticky situation: Touch sensor, front panel... front panel, touch sensor.

Time to glue the touch sensors into the front panel... The adhesive you see is plumbers GOOP (yes, I love this stuff and use it for almost everything, it's great versatile stuff!)

A dab of GOOP.


Spread with plastic cutlery until roughly even.


Put into spot on front panel and press like a mo-fugger for a minute or 2 while it starts to set up, then tape it down.


Lather, rinse, repeat until you get this:


What? you wanted to see them all lit up? too bad, so sad, the glue has to cure. I'll probably light everything up tomorrow, but I've tested it this evening and there was a small problem. The GOOP is a solvent glue at heart, and ruined the frosted look of the white ring around the sensors. I've got a trick up my sleeve though *Winks at scopEDog*

VFD Signal Cable: Extreme(ly) Ugly Cable Makeover!

The horror, 2 lengths of gray CAT5! Hacked up ribbon cable! ick!


5/8 sleeving... (Mental note to purchase huge 1" 4:1 shrink heatshrink... )


(Realizes he's out of electrical tape, how embarrassing, runs to wallyworld and buys a firesafe (for 24 LTO-2 tapes) and 2 rolls of electrical tape.)

A little electrical tape never hurt anyone.


A cut-down slot cover + some of mnpctech's now famous u-channel to dress up the cable's entry into the rear of the case. Why not drill a hole in the slot cover and feed it through? Because the cable is too thick for that.


Entry...


The whole shabang from parallel port on down to entry.


Till next time....

PS: Do ya'll like the shorter, more frequent updates? or do you like the less frequent, big juicy ones that seem to flow a bit better?

[Karlmcwade]

i like quick updates myself

[starbuck3733t]

Sticky Situation: continued + touch keys finally installed
The GOOP is all dry, and everything is stuck together and ready to rock! Here we go!

Connections bound together on the back of the bezel


All lit up!


Now you see it..,


...Now you don't.


A little bit of custom tooling... This solves the fact that the frosted effect was destroyed when the goop solvent glue was applied to stick the keys to upper front of the bezel.


And after that, we've got frosting again.


Not as good as the original, but still pretty good.


Door installed on case, again.


Hrm, where'd those keys go?


Booyah!


The bottom was finished, and reinstalled on the case along with the feet.


Industrial black again


USB/firewire front ports all painted up w\ the same flat black.

[starbuck3733t]

Front panel indicators + indicator bracket

Ben @ mountainmods cut this bit along with all the bezel. The holes are for fiber optic cables to fit through, so that the indicators are completely hidden when they're off. This piece mounts behind the front panel on a bracket (yes, a bracket for a bracket) that we'll see fabricated shortly.

Bracket (not bracket bracket, that's still coming.)


A bunch of LEDs with holes drilled in their tops w\ fiber optic cable strands installed. I made them long for a reason, you'll see shortly.


LEDs installed in the bracket, you can just barely see the excess fiber hanging out the bottom. I trimmed the black pieces off, they were really just there for extra fiber to hold on to, to make stripping the black coating off easier.


LEDs all installed, and then the excess fiber trimmed, and sanded flat.


An example of the point-source nature of the light.


The bracket-bracket: mushroom, mushroom

A little bending break action.


And shortly thereafter, we have this:


Trimmed with rubber u-channel.


And installed.


As far as mounting the bracket on the bracket-bracket, that'll have to wait till tomorrow... I'm beat.

[Maviryk]

I wish I had as many tools as you do.

I think my e-peen shrank after seeing your project compared to mine...

Great job on your case, can't wait to see it finished. Where did you get the mesh for the little bracket? I'm trying to acquire a 1'x2' section of it for my front panel(G5 rippoff).

[starbuck3733t]

Lots of people say I have lots of tools, I'm still gaining mastery of them all.

The mesh came from www.mnpctech.com, though you may be able to find it elsewhere. The holes are a LOT bigger than the ones on the G5. He's also got hexagonal mesh with better percentage of airflow vs. metal for use in grilles. I used the roundhole stuff becase...well... I just did.

[starbuck3733t]

Mild setback

Grump grump grumpy I am...

So I was fitting the stealth activity LEDs yesterday (which I got most of which done, SCSI HDDs, IDE HDDs, DVD-read) and made the cables and fitted them into my highly modified HP CDRW-8X SCSI optical drive only to find out that the only LED that lit was the red one for "WRITING". Of course I know exactly why this is happening, and the fact that I didn't think of it when I was doing the mod is a bit shocking... I usually think of these things.

The problem was that the drive originally carried Green and Red LEDs to indicate disc in/activity and writing w\ another red one for tray lock (meaning you try to eject it whilst it's writing)... well, what's the diff between a green, red, or orange LED and a blue/white, other than the color? The forward voltage and current required to get it to light. Dooy! :duh:

So I crack open the drive and track down the current limiting resistors (small ass surface mount ones) which are 330 ohms being fed 5V. Well no wonder the white/blue ones aren't lighting! So I solder in a 150 ohm resistor in parallel w\ the existing 330, making ~103 ohms. This is a fine number, seeing as I usually wire them up with 82 ohm resistors for maximum brightness. Stuff the whole thing back together and fire it back up in the rig. No lights still. WTF. Oh .... utoh.... crap! NO BOOT! (Hung at the drive being detected on the 29160 controller, passed the post OK.)

Pull the drive and disassemble... detect distinct smell of cooked electronics. crap. Not sure what I cooked (I can't find which part the smell is coming off of) but the tray won't eject. I suspect I fried the some of the motor drivers, which is causing the main controller chip to hang on the drive's motion test (which all drives do as soon as they get power) and then not be detected by the SCSI controller as the drive controller is still hung up in the motion test.

Short story..I killed it.

But no fear... I rooted through my drawers and find a 48X Yamaha CDRW-F1 w\ SCSI backpack! (IDE -> SCSI converter) WOOOOOT!!! Cracked it open this morning at work after sorting out the backup software issues, and discovered that its already got a blue activity LED in it, meaning it should happily light my white/blue ones w\o modding the current limiting resistor. However, it doesn't appear to have a tray lock LED, or if it does have a spot for one, it's not installed. W\ any luck I can get that installed tonight while watching the Redskins game on the TV.

:thumb: :dremel:

[Karlmcwade]

I hate smelling cooked electronics (sulphur) proven to make grown men cry, exp when its the third time this week.

[starbuck3733t]

Another S**t G** D***it moment

I finished off the stealth activity leds around 1:00 AM yesterday, everything works fine... all 8 of them: SCSI HDDs, IDE HDDs, Optical 1 Write, Optical 1 Read, Optical 2 Read, Network 1 Act/Link, Network 2 Act/Link, Network 1&2 @ anything other than 1000mb/s (fault.)

So I went to hook up the light show to shoot a few pics (as I didn't take any of the nic card mod, who of you wants to see me hack @ a dual port gig ethernet card, anyway!) and discover that some of the key backlights didn't work. Wierd. I fiddle with the connection on the 1 and find that something inside the heatshrink had shorted, but I wiggled it and it happily un-shorted.^_^ So I attempt to do the same on the other one, only to find out that it's oddly dim (not the right color of white as the others) after I wiggle it. So I figure its a very, very minor short and continue peeling away at the heatshrink and sleeve until I've got the whole thing loose... and BROKE THE LEGS OFF THE LED which is EPOXIED INTO THE KEY. Damnit! :duh:

So now I have to take the bezel off (which isn't a big deal, 2 8-32 threaded machine screws) and DRILL OUT the LED, and epoxy a replacement. Damnit Damnit Damnit!!!

[starbuck3733t]

FINISHED! Front Panel Indicators + Indicator Bracket

I love that F word. Finished that is... This should really be a huge juicy update, but it isn't. There was a TON of work done behind the scenes to make the "blinken' lights" happen. In addition to the repairs need to the 3rd (starting on the left) key's backlight (which really only took about 45 minutes), there were several other mods that had to happen to make this all possible:

1.) Modified dual-port gigabit Ethernet card for external link/activity LEDs & combined speed LED.
-link/act lights when there is an established link
-link/act blinks off when there is activity
-there are two link/act lights, 1 for each Ethernet port on the card
-the combined speed LED lights RED in the case of either card not having a link @ 1Gb/s.
2.) Modified Yamaha CDRW-F1 for external LEDs and & eject switch
- disc in/act lights when there is media in the drive
- disc in/act blinks off when there is activity
3.) Another eight custom cables to run the front panels, sleeved, and heatshrunk (for quality!)
- SCSI LED (white)
- HDD LED (white)
- CD/RW Writing (red)
- CD/RW Reading (white)
- DVD Reading (blue)
- NIC 1 Activity (white)
- NIC 1/2 Low Speed (red)
- NIC 2 Activity (blue)

We now rejoin our regularly scheduled program, already in progress...

Finished bracket-bracket & LED bracket.


The backside w\ the LEDs mounted & legs trimmed. I'm using .100" spaced connectors on the cables that go to the devices. The devices take care of the current limiting for the LEDs, keeping these compact.


indicator panel installed in the case. Note that I had to move it up from the last time you saw it... I didn't like the LED indicators so close to the touch keys... it looked crowded. This is better.


Wider angle of same thing. And I <3 the photoshop Lens Correction filter! It's how I got this one perfectly straight.


DVD Hanging out. You can barely make out the mesh behind the smoked plexi...it's a neat effect. The VFD isn't installed yet as I didn't feel like hauling my laptop down to the basement (Yeah, I'm lazy, an IBM T43 is so heavy, NOT) to drive the screen.


Labels are good. The ">" indicates which direction the connector should face, so the polarity on the LEDs is right should I ever have to disassemble everything. This is only 3 of the eight cables, though.


Yikes... A shot from the top with the door closed, with the wiring for the touch keys and the indicators below it.


Money shot!


MONEY VIDEO (1.2MB, DivX)
The first thing I point at is the activity LED for the CDRW drive, note that it lights up as soon as the drive registers that there is a CD in the drive, stays on as long as there is a CD, and blinks off for the activity. The second thing I point at is the same thing for the Pioneer slotload DVD-ROM... which normally stays off and blinks on for activity... that was a fun mod to come up with.

From left to right, the indicators are:
- SCSI LED (white)
- HDD LED (white) - not lit in video at all
- CD/RW Writing (red) - not lit in video at all
- CD/RW Reading (white)
- DVD Reading (blue)
- NIC 1 Activity (white)
- NIC 1/2 Low Speed (red) - not lit in video at all
- NIC 2 Activity (blue)