The Mineral Oil PC

legoman666

Gawd
Joined
Dec 21, 2003
Messages
638
List of Updates:
1: Jan 31 2009 - Potential design for the case feet.
2: Feb 2 2009 - Process by which I make the feet
3: Feb 6 2009 - Construction Begins
4: Feb 7 2009 - Moving Right Along...
5: Feb 9 2009 - Lower Frame Completed
6: Feb 12 2009 - Something I've been Working On...
7: Feb 13 2009 - Made the Bottom Panel and Attached the Feet
8: Feb 15 2009 - More Side Panel Work, 1 Rad Mounted
8b: Feb 15 2009 - Mini Update, LEDs in Feet.
9: Feb 16 2009 - Yet More Side Panel Work, 2 Rads Mounted
10: Feb 21 2009 - PCB mounting Extravaganza
11: Feb 25 2009 - Weekday Evening Modding
12: Feb 28 2009 - Completed Upper Section
13: Mar 8 2009 - More Electrical Work
14: Mar 12 2009 - Hinged the Top Section (plus other misc work)
15: Mar 14 2009 - Made "Tank" for 8" LCD
16: Mar 15 2009 - Tank Construction Complete!
17: Mar 22 2009 - Motherboard Tray Construction!
18: Jun 21 2009 - Sanding the Panels and Mounting Stuff!
19: Jun 23 2009 - Hard Drive Mounts and Other Pics
20: Jun 25 2009 - Almost Done!
21: Jun 27 2009 - Putting it all Together
22: Jun 27 2009 - Fail! A Leak, Oh No!
23: Jul 3 2009 - Fully Submerged, Alive and Kicking
24: Jul 10 2009 - LanWar, Pics and Stuff
25: Jul 13 2009 - Final Photo Shoot
26: Jul 22 2009 - Backlight Hottness!
27: Aug 17 2009 - New Sponsor!
__________
28: March 21 2010 - Back from the dead?
29: March 24 2010 - Tubing Update
30: March 31 2010 - Prepping GPUs for Oil Bath
31: April 7 2010 - Houston, we have liftoff!
32: April 17 2010 - Leak repaired, oil re-submersion successful!
33: May 1 2010 - Overheating problems fixed!
_________________________________________________________________

It's that time again!

Petra's Tech Shop!
PTS_banner_OCP.jpg

I'd like to thank Alex Venz for the support by means of a discount. PTS will also be supplying a XSPC RX360 Radiator which should go nicely with my low speed Noctua fans.

Crystalfontz!
log_crystalfontheader.gif

The crew over at Crystalfont has kindly agreed to supply me with the CFA-635 LCD module and UberSCAB kit that I'll be using in this mod. Thanks! Be sure to give them a look.

I'm happy to announce a new sponsor: Crucial!
Kelly Sasso has kindly agreed to supply me with a brand spankin' new 128Gb CT128M225! Sweeeet. These just came out a few days ago and the reviews are trickling in.

Teaser pic:
IMG_2175.jpg
 
Last edited:
This will be my 3rd project worklog that I have posted on bit-tech.

My first was my Wooden PC, completed in early 2004:
Old_Wood_Case_1.jpg
Old_Wood_Case_2.jpg

Worklog: http://forums.bit-tech.net/showthread.php?t=38878 (pics are all dead)

My second was my Aluminum/Copper PC, completed early 2008:
Old_Aluminum_Case_2.jpg
Old_Aluminum_Case_1.jpg

Worklog: http://forums.bit-tech.net/showthread.php?t=145057

And that brings us to now, I present to you: The Mineral Oil PC!
Link to planning thread in the Modding forum: http://forums.bit-tech.net/showthread.php?t=163378 Most of the following posts are copied straight from this thread, so some stuff may be out of context.

Here are my preliminary sketches (from 22/1/09):
log_CaseMkI_1.jpg


log_CaseMkI_2.jpg


log_CaseMkI_3.jpg


log_CaseMkI_4.jpg


Description of the general layout:

Bottom compartment:
2 XSPC RX360 Radiators: Link to product
6 Noctua NF-P12 120mm Fans: Link to product
1-2 Eheim 1250 120v Pumps: Link to product
3 fans will suck in, 3 will blow out

Middle tank:
Motherboard: ASUS Crosshair II Formula Link to product
CPU: Phenom II X4 940 Link to product
RAM: 8gb G.SKILL 1000mhz DDR2 Link to product
2 HD4850's: Link to product
Silverstone ST75F 750W PSU: Link to product
Thermalright IFX-14: Link to Product
2 Thermalright T-Rad^2 GPU Coolers: Link to Product

The Mobo/CPU/RAM still isn't 100% set in stone.

Top Compartment:
2x WD 500gb GP Hard Drives (maybe)
2x WD 250gb Hard Drives (maybe)
The top is an exit for cables from the"back" (top in my config) of the motherboard. If I don't go with SSD's, I'll have my hard drives in this compartment too. As you can see in the pics, the tray will have 2 handles for easy removal from the oil. Everything that will be submerged in the oil will be attached to the tray. The top of the case will also be hinged for easy access to the oil.

More Preliminary Sketches (from Jan 24 2009)
log_CaseMkII_1.jpg


Just playing with the model more.

Are MCP655's submersible? I can't fit my pump(s) in the bottom compartment. I realized this when I replaced the radiator models with model's of the radiators I'll actually be using. If I can't submerge the MCP655, I'll have to stick with Eheim pumps...

Anyone know?
 
More Preliminary Sketches (from Jan 27 2009):

Today it snowed ~5-6" overnight, so I wasn't able to make it into work. So I whittled away the day doing absolutely nothing productive. It was great. Looks like I might be stuck at home tomorrow, since it's been freezing rain/sleeting for the past 2-3 hours and we're supposed to get 4-6" more snow by the morning (it's 9:45 PM).

Here's what I did:

log_CrosshairFormulaII.jpg

I modeled the Asus Crosshair II AM2 motherboard. At this time, it looks like I'm going to go with an AMD setup, unless some other sponsors come through. The latest hardware list:
ASUS Crosshair II Formula
Phenom II X4 940
8gb G.SKILL 1000mhz DDR2

It also turns out that I was seriously uninformed about TECs. The Phenom 940 is 125W TDP which means about 150w at full load. To get a deltaT of 30C between the hotside and cool side of the pelt, I would need to have a pelt with almost 400w of cooling power, which translates into something ridiculous like 600w worth of pelts. The 226w pelt I was going to buy would have produced a deltaT of 11.5C with almost no room for overclocking.
(If this doesn't make any sense, it's because I'm copying these updates from my initial log in the Modding forum)

So it looks like I'm going to stick the most massive heatsink I can find on my CPU: the Thermalright IFX-14 Extreme

log_CaseMkIII_1.jpg


log_CaseMkIII_2.jpg

Getting rid of the pelt allows me to have a single PSU instead of 2 and make the case smaller and thus cheaper.

I still need to do a lot of work on the model, it's just very tedious. I'm using the Windows 7 Beta at the moment and the hardware acceleration of Google SketchUp doesn't work. Even with a Q6600 @ 3.55ghz it's wrist-slitting-slow.

log_CaseMkIII_3.jpg


log_CaseMkIII_4.jpg


With any luck, I'll order the raw materials for the case (aluminum sheet, bar and angle, and the plexiglass) by the end of the week. Construction will commence when it arrives.
 
More Preliminary Sketches (from Jan 28 2009):

Turns out I wasn't able to make it into work again today either. Yesterday we got 5" of snow, then 1" of sleet/freezing rain. Today we got another 4-5" of snow around 6AM. As of 5:30PM, most of the streets were still in terrible condition.

So I celebrated my free day by spending a little time making a model of the HD4850 without the cooler.
log_atihd4850.jpg


I need to model the GPU cooler I'm using and add it to the model, (the T-Rad^2) but they're actually in use right now and I'm not taking them out of the case.

I also played with the model more. I made it shorter and changed a few things.
log_CaseMkIV_1.jpg


log_CaseMkIV_2.jpg


Tonight I placed the orders for the raw materials!
I ordered the following:
18" x 24" x 3/8" Plexi - This will be the left side of the tank
24" x 24" x 1/4" Plexi - This will be the front and back of the tank
18" x 24" x 1/4" Plexi - This will be the right side of the tank
12" x 24" x 1/4" Plexi - This will be the bottom of the tank

The left side of the tank is thicker because it will have screws going into it.

96" 1.25" Angle Aluminum - This will be made into .75" pieces for the joints
192" .5" Angle Aluminum - This will be used to strengthen the corners of the plexiglass
32' .75" Bar Aluminum - This will be used structurally in the bottom of the case and for the mobo tray
9' sq .05" Sheet Aluminum - This will be used for the panels

When I say "angle aluminum" I mean this: http://www.onlinemetals.com/merchant.cfm?pid=1019&step=4&showunits=inches&id=63&top_cat=60

Here is how I will use the 1.25" stuff:
log_angleexample.jpg
 
Here's something I just found:
log_CFA-635_TMF_Front_Backlight_On_.jpg

https://www.crystalfontz.com/product/CFA635TMFKU1.html in conjunction with

log_UBERSCAB_large_photo.jpg

https://www.crystalfontz.com/product/UBERSCAB.html


I've used a CrystalFontz screen before on my first mod. I liked it well enough. This would allow me to use temperature curves to control the fans on the radiators. I could also monitor several oil temperatures with little sensors. I'll use this guys idea: http://www.crystalfontz.com/forum/showthread.php?t=4452 to stick sensors on the inlet and outlet of the rads and perhaps in 2 other locations (top and bottom of the tank?). I could also use a flow meter to sound an alarm if the pump fails. This would be great since I won't really be able to see the oil flowing nor will I be able to hear the pump.

I made a simple sketch of what I had in mind:
log_simple_sketch_2.jpg


This is the bottom section of the case. I was going to use 90 degree elbows for these corners since it's a tight fit, but I can simply change them to T's and stick the temperature probe in the branch pointing downwards and seal it with caulk or silicone or something.
 
Does anyone know how these work? http://search.digikey.com/scripts/DkSearch/dksus.dll?Cat=1966259&keywords=pressure switch My plan is to stick this on the pump output and attach the sensor in parallel with the power button. I don't know what the Normally Open refers to. Is it NO when there IS pressure, or is it NO when there IS NOT pressure? I would like it to be OPEN when there IS pressure. That way, when there is no pressure, that is to say, when the pump is not running, the switch will close and will essentially press the power button.

Anyone know what type I should buy?

Actually I just answered my own question. http://content.honeywell.com/sensing/hss/hobbscorp/pdf/cat_pav.pdf

"Normally Closed (SPST-N.C.)
- A normally closed switch conducts electricity until the actuator is moved by the media causing the contacts to open."

This is what I want. Now I just need to figure out how much PSI my pump will output in the oil. I don't want to guess wrong and buy the wrong type of switch, they cost ~$25. There are two that might fit my needs. One operates from .5 to 1 PSI, the other operates from 1.1 to 3 PSI (they're adjustable). Anyone have a guess as to the output PSI of a Eheim 1250 when moving mineral oil?

Edit:
PSI = 0.434 x h x SG
h = max head in ft. The max head on the Eheim 1250 is 6' 7" = 6.58'
SG = specific gravity. The specific gravity of mineral oil is .818 to .905 according to its msds sheet.

Thus the max PSI is 0.434 x 6.58 x .85 = 2.43 PSI. I'll buy the 1.1 to 3 PSI model unless someone knows something I don't. Speak up!

Math is fun.
 
I hope you know that the motherboard you selected is NOT a crossfire motherboard, If you want your 4850's in Crossfire, your going need either a P45 or X48 based motherboard. Not NF200.
I know ;) I have 2 HD4850's, but I don't run them in crossfire. I have 4 monitors, so even if I wanted to run crossfire, I couldn't :(

p0pe said:
i think you would get a lot more performance if you made som sort of quick fitting that allowed you to get the mineral oil to go from the pump to a cpu block, and then the gpu´s so that you could jet it over the blocks. wouldnt that give better performance and less weight?
I'll be doing something similar. The outlet for the radiators will be near the video cards so I can supply them with constantly cool oil. I'm not 100% certain where I'll put the pump inlet. It'll either go near the CPU or just intake from the bottom of the tank.

I would also like to announce a third sponsor: Crystalfontz!

log_crystalfontheader.gif

The crew over at Crystalfont has kindly agreed to supply me with the CFA-635 LCD module and UberSCAB kit that I'll be using in this mod. Thanks! Be sure to give them a look.
 
Got some goodies in the mail today. Nothing too exciting though.

log_goodies_switches.jpg

3 rocker switches. Not sure if they light up or not. Not too sure about the green.

log_goodies_relays.jpg

2 relays. One for the pump and 1 for the 2nd PSU that I will no longer be using.

log_goodies_powerplugs.jpg

A bunch of plug inlets. Exciting, I know.

log_goodies_LEDs.jpg

30 red LEDs.

log_goodies_eheim1250.jpg

Eheim 1250.

Hopefully next week will turn up the good stuff.
 
And in other news, scroll saws are not ideal for cutting plexiglass. I was trying to make a plexiglass back for a AC Ryan Radgrill and light up the edges with LEDs. So I printed off the template and taped it to a piece of plexiglass. I cut around the outer edges with a band saw no problem. Then I tried to use a scroll saw to cut the middle out. No good. The cut just fused back together behind the blade.

log_scrollsawssuck1.jpg


log_scrollsawssuck2.jpg


Lol, it was like I hadn't even cut it. I hate scroll saws. I couldn't even break the "cut" with a hammer. When I hit it with a hammer, the thing shattered (not even along the cut). The fused together cut seems just as strong as before I cut it.

I was just trying this today on a spare piece of plexi to see how well it would work. Looks like I'll have to use a jig saw to cut the middle out.

Yeah slow and steady with acrylic. Drop the RPM down on the scroll saw and it will cut the plexi not melt it.

It has 2 settings, high and low. It melted the plexi on both :(
 
You have a very-cool (unique) project underway..

Thanks.

Then it looks like you may have to use a Dremel or a Jigsaw for the inside cuts.

Yarrrgg

What kind of blades are you using? Maybe try getting some specifically made for plastic. My scroll saw (Delta) worked great on plexi. But I used the wrong blades so the cuts were ugly.

It was probably a blade designed for wood. I didn't even think about it. There a quite a few other blades I can try; there's a whole container full next to the saw.

I was bored today so I messed around in the shop trying to come up with a decent design for the case feet. Here's what I came up with.

log_Foot1.jpg


log_foot2.jpg


log_foot3.jpg


log_foot4.jpg


It was a lot of work making those 2 pieces of 1/4" aluminum perfectly round... But I think it turned out pretty well. I won't use this particular foot on the final case as it's not quite up to scratch. However I will probably use the design. Thoughts?
 
Here's the process by which I make the feet. It's a lot of work.

log_feetprocess1.jpg

First I made a template so I don't have 4 slightly different feet on my case.

log_feetprocess2.jpg

I traced the outline onto a big piece of 1/4" thick aluminum.

log_feetprocess3.jpg


log_feetprocess4.jpg

I do the rough cuts with a band saw.

log_feetprocess5.jpg

Then I drill a hole in the center so I can stick a bolt through it.

log_feetprocess6.jpg

I made a little jig on the belt sander so I can easily do the rough sanding. I just stick the bolt with the rough cuts into the wood and slowly spin it to a circle.

log_feetprocess7.jpg

It gets hot enough that I often have to dunk the aluminum into water. You can see the water boiling in this pic.

log_feetprocess8.jpg

All done on the belt sander....

log_feetprocess9.jpg

...Now onto the drill press. I moved the belts around to get around 2000 RPM. On the press I use various grits of sandpaper to get the edges nice and smooth.

log_feetprocess10.jpg

Then I put the template back on and drill the remaining two holes.

log_feetprocess11.jpg

And tap them.

log_feetprocess12.jpg

I use a bigger drill bit to countersink the screws.

log_feetprocess13.jpg

I repeat the process with the piece of plexiglass and then sandwich it between the two discs of aluminum.

log_feetprocess14.jpg

All done and polished. (Or at least as smooth as I care to make it.) 1 foot took about 1.5 hours to make.

log_foot3.jpg

Final product.
 
And that brings the log up to date. I'm currently waiting on 5-6 shipments which should arrive on Feb 5th. If anything is confusing or out of context, please bring it to my attention. I just copied almost everything verbatim out of my Bit-tech log.
 
Thanks guys.

Got some more goodies in the mail today!

log_LCD_2.jpg

The package from Crystalfontz arrived today. Thanks guys! Lots of stuff inside: the LCD, the SCAB board, 4 3 pin fan cable extensions, 4 temperature sensors, 1 cable to allow the LCD to turn on the computer, 1 molex to 4 pin floppy power connector, 2 USB cables for the LCD, 2 different length cables to connect the SCAB to the LCD and a protector for the LCD. Jeez!

log_LCD_1-1.jpg

I couldn't resist hooking it up and messing around with temperature controlled fan curves.

log_pressureswitch1.jpg

I also received my pressure switch. I couldn't resist testing this either, looks like it'll work great.

log_foot5.jpg

Got 2 feet done. And yes, they're different sizes. I made the one on top first. I cut it out using the band saw. The bigger one on the bottom was cut out using a hole saw. I'm not sure which method was a bigger PITA. I may make 1 more smaller foot and 1 more bigger foot . They're not different enough that it'd be noticeable as long as the big and small ones were grouped together. I hope.
 
I can't wait till this is done. I have always wanted to do a mineral oil based computer, but haven't ever really come up with a plan on how to get rid of the heat. This is looking very good.
 
I like the design of the case, just for a water cooled setup. Cant wait to see more
 
Gonna keep an eye on this. Been banging around in my head for a while the idea of running a phase chiller for the cooling loop and using submersion as a means of condensation proofing.

Your space planning looks pretty good thus far, keep it coming. ;)
 
Got a couple more packages in the mail today. A couple more should arrive tomorrow.

log_radiators1.jpg


Petra's Tech Shop
came through on their sponsorship. One of these nice looking XSPC radiators was provided free of charge. At low fan RPM, these out perform Feser's TFC. There's a review of them here: http://www.skinneelabs.com/Radiators/XSPC/RX120/RX120.html

PTS_banner_OCP.jpg

I'd also like to thank Alex Venz for the support by means of a discount (and the free rad). Give them a visit!

I also got a bunch of screws from Small Parts Inc.

log_screws1.jpg

These 8/32 3/4" phillips head screws will be used to hold my L brackets in position.

log_screws2.jpg

These 8/32 1/2" flat socket head screws will be used internally. They will be countersunk to keep them out of the way.

log_screws3.jpg

And lastly, these 8/32 1/2" button socket head screws will be used to hold the aluminum panel to the frame.

log_the_plan1.jpg


log_the_plan2.jpg


log_the_plan3.jpg

I made a quick sketch to show how I will hold the two different sizes of L angle to the bottom of the frame. The green stuff is 1/2" and will provide strength to the acrylic frame. The red pieces are 1.25" and will be used to hold the corners together.
 
I got a couple more packages today. One of them contained this:
log_progress1.jpg


log_progress2.jpg

All of the aluminum I ordered! There's a 3' x 3' x .05" sheet, 32' .75" x .75" bar stock, 16' .5" angle and 8' 1.25" angle.

log_progress3.jpg

I also received 3 of the 6 Noctua fan's I'll be using. I only bought 3 because I already own the other 3; they're currently in use.

log_progress6.jpg

Using these nifty little guys that come with the fans....

log_progress8.jpg

...I mounted them on to one of the radiators. It's a lot of trouble to get those little rubber things through the back of the holes so I can mount the fans. I did this so I would know the exact dimensions of the rad with the fans mounted and I can size the case accordingly.

log_progress7.jpg

My first corner complete! This is one of the many corners of the frame.

That's all for today! I did some more work making the frame, but I kind of messed up. There was a slight miscalculation that resulted in screws hitting each other when they shouldn't. I'll fix it tomorrow.
 
I had planned on just posting one update tomorrow, but I couldn't keep the pics all to myself! So here they are:

log_progress9.jpg

I mentioned in my last update that I completed a little more than I had shown. Well, here it is. I was not happy with the quality of my construction. As you can see, 2 of the holes don't have screws. That's because the screws coming from the other direction ran into them. I didn't measure the spacing well enough. The other 2 arrows point towards where I was able to get the screws in, but I had to grind down the head so they didn't collide. Needless to say, it looks poor and all around unprofessional.

log_progress14.jpg

I decided to start over with proper planning. I made a list of the sizes of all of the pieces I will need and their locations. Each piece gets its own identifier so there will be no confusion later.

log_progress15.jpg

With the plans in hand, I turned to my trusty mitre saw and cut to length all of the pieces required.

log_progress17.jpg

Then I labeled them accordingly.

log_progress16.jpg

A quick mockup of how the pieces will go together.

log_progress18.jpg

Drilling commences.

log_progress24.jpg

I made a small template that allows me to drill 2 different sets of holes in the exact spacing I need with no variations. I either use the two far holes or the two middle holes.

log_progress21.jpg

Using the template, I made a bunch of my L brackets. Each one bracket has 2 far spaced holes and 2 closely spaced holes.

log_progress23.jpg

A complete corner. It turned out a lot better the second time around. My little template worked quite well.

log_progress20.jpg

Two corners done. These take a long time to do. For each corner I have to do the following:
Drill 12 holes through 3/4" thick aluminum.
Tap the 12 holes. (my arm hurts)
Make 3 L brackets.
Drill the holes in the L brackets.
Put everything together and hope to god that it all lines up properly.

log_progress22.jpg

This is what I completed today on the frame. My arm was starting to aches from all of the holes I had to tap. I can tell the complete frame is going to be ridiculously strong. It barely has any flex in it.

log_progress10.jpg

I also worked on making another foot. For this, I turned to the drill press. I think I've decided to use the hole saw method to make all of the feet. They turn out a lot better and require a lot less sanding.

log_progress11.jpg

This hole saw is super crappy. I have to go very slowly or it gets stuck. It also shakes the press so much that the chuck on the drill press comes loose. I have to stop and retighten the chuck several times while cutting these.

log_progress12.jpg

Two pieces immediately after the hole saw is done with them.

log_progress13.jpg

The feet I've completed so far. The middle two were cut with the hole saw. The right foot was my first, and the left foot was the second.

And that's where I stopped this afternoon, my arm was aching too much. I recieved my acrylic today, but it's nothing exciting. The sheets are all covered in paper anyway. I peeled off a little from a corner just to peek, stuff looks great.

What do you guys think?
 
You are making really good progress.

Thanks ;)

I am a Electrical Engineering co-op (intern-ish) at a major power utility. I went to the shop where we service broken and old transformers. Guess what transformers are filled with.... mineral oil!

log_0206091056a.jpg

We have a gigantic tank of the stuff. I wasn't allowed to steal any :( It would be nice to use the machine shop to construct my case; there's a bunch of nice metal working machines.
 
Hurray! Update time! :clap: Summary of work since Saturday:

log_progress25.jpg


log_progress26.jpg

I finally completed all 4 feet. I gave in and made them all using the hole saw; they just come out so much better this way. They're all fairly close to the same size; no differences big enough to discern once they'll be in place at the bottom of the case.

log_progress27.jpg

All of the L brackets needed to complete the lower frame assembly.

log_progress28.jpg

The top portion of the lower frame all put together.

log_progress29.jpg

Decent macro shot of a corner.

log_progress30.jpg

Hole tapping fun. :LOL:

log_progress31.jpg

Completed drilling and tapping the risers.

log_progress32.jpg

And then put it all together.

log_progress34.jpg


log_progress35.jpg

Close ups of some corners.

log_progress36.jpg

It's quite sturdy. I weigh 130lbs and it didn't flex at all.

log_progress37.jpg


log_progress38.jpg

Checking the spacing just to make sure I didn't make the frame too small!

log_the_plan1.jpg


log_the_plan2.jpg


log_the_plan3.jpg

Remember this?

log_progress39.jpg

Doing a few test cuts on a piece of scrap.

log_progress40.jpg

And now doing the cuts on the real thing.

log_progress41.jpg

From the other direction....

log_progress42.jpg


log_progress43.jpg

Checking the fit, looks good.

log_progress45.jpg

View of both cuts. When I did these cuts, I backed out the screws about 1/4" to make sure I didn't chop parts of them off. If my calculations were correct, I wouldn't have cut into any screws even if I didn't back them out, but I wanted to be sure.

log_progress44.jpg

Oh noes! The jig slipped! This piece is unsalvagable. I've already cut a new piece of bar stock to length and am preparing it to take the place of this one.

And that's where I stopped tonight. I'd once again like to thank Petra's Tech Shop, Danger Den and Crystalfontz for helping me out!

Thoughts?
 
All of the work you have done so far is fantastic!

One question though: Would it be more feasible to put the radiators above the mineral oil chamber? That way, the weight is at the bottom, which would put less stress on the brackets and make it more stable (less top heavy). Also, since the hotter mineral oil would rise to the top, make it easier to cool?

seems like a good idea to me.

this one is gonna be one of the most heavy case of the world :D

Yes, that would be the ideal design. I struggled when I designed the general layout. The problem is that the cables coming from the motherboard have to go in the top of the case. My intent is to make the motherboard tray semi-removable like this:

log_slidingrails.jpg


29094-01-500.jpg

Using a drawer slide similar to this. The top aluminum portion is going to be on hinges so I can pull the slide out. I wouldn't be able to do this if the radiators were on top.
 
Someone needs to make a mineral oil cooled computer similar to how a transformer is designed.

0210091118a.jpg

Like this. Transformers use natural convection to keep the oil cool. This particular transformer is made for 100% passive operation.

0210091119a.jpg

I got a shot of the stuff we (the utility I work for) use in transformers. Here's some data I looked up on it, thought other folks might find it interesting:
MSDS Sheet: http://www.equivashellmsds.com/Display_MSDS.asp?ID=202481
Specifications: http://www.nttworldwide.com/docs/diala-ax.pdf

The highly refined mineral oil contains <3% (w/w) DMSO-extract, according to IP346.
Highly refined mineral oils and additives.

I need to get myself a drum of it. :D

Edit: Sweet, I asked the dude if I could buy 10 gallons of this stuff: http://cgi.ebay.com/Shell-Diala-HFX...re-Resist_W0QQitemZ230315732658QQcmdZViewItem That would be awesome.
 
As an eBay Associate, HardForum may earn from qualifying purchases.
Something I've been working on...

log_progress46.jpg


log_progress47.jpg

LED chaser lights! Weeeee. It took me forever to get that thing wired up. There are two potentiometers, one controls the brightness and the other controls the speed.

Short video demonstration:
http://www.youtube.com/watch?v=qNzpDnNeZ3w
Yea, for some reason, the pinout for the 3rd LED doesn't work on the IC. Don't know what's up with that. The plan is to arrange the LEDs in a circle and use the circuit for the HDD activity light. I'll probably stick the LEDs behind some frosted plexiglass to help diffuse the light as well. Should be pretty cool.

Here's the circuit I used: http://www.electronics-lab.com/projects/games/003/index.html
Until next time...
 
Good news! I was talking to my boss at work and somehow we got to talking about this project. I explained what I was doing and why (lol, that's harder to explain than you might think) and the subject of mineral oil came up. I mentioned that I found an auction on eBay that was for the same type of stuff that we use but it would cost me ~$200 to get a hold of 10 gallons. Boss man took me to see the guy in charge of construction of new substations and told me to explain my project to him. The guy called someone up at the warehouse where we keep the oil and arranged for me to come pick up 10 gallons of the stuff for free some time next week!

Huzzah! Saves me $200. My only concern is that the stuff won't be clear. However, I've seen the inside of one of our freshly oil filled transformers and it was as clear as any other mineral oil I've seen, so I should be set.
 
log_progress50.jpg

As tired as I am tonight, I decided to do some work.

log_progress51.jpg

I cut a piece of aluminum left over from my last mod to the width of the bottom of the frame. This piece will not be visible unless you somehow.... well no I can't even think of a way.

log_progress52.jpg

Checking the fit on the frame, looks good.

log_progress53.jpg

Using my trusty right angle...

log_progress54.jpg

...I cut the piece to length. (Dumb camera wouldn't cooperate for this pic)

log_progress55.jpg

I used my little puncher device to mark spots to drill. (Yes, this one isn't on the line. The line is in the wrong spot, not the punch)

log_progress56.jpg

Drilling the holes for the screws that hold the sheet onto the frame.

log_progress57.jpg

I learned on my previous mod that it is impossible to drill the holes on the sheet and frame separate. No matter how well you measure the spacing, the holes will never all line up properly. The best method is to tape the two pieces together and drill them at the same time.

log_progress58.jpg

More holes to tap :wallbash:

log_progress59.jpg

Using the drill to screw the panel to the frame.

log_progress60.jpg

Not really sure what my camera was trying to focus on... the reflection I guess.

log_progress61.jpg

Looks pretty good.

log_progress62.jpg

Marking out the spacing for the case feet.

log_progress64.jpg

Drilling the holes through the feet and then into the bottom panel.

log_progress63.jpg

Then I screwed them on. The middle hole is for the LED to light up the feet.

log_progress65.jpg


log_progress66.jpg

All four feet mounted.

log_progress67.jpg

View from the front.

log_progress48.jpg


log_progress49.jpg

Also got a few more goodies in the mail; an aluminum ATX back panel frame, slim slot loading USB DVD burner, a 7 port USB hub that will be internalized in the bottom compartment, and a 4oz bottle of Weld-On #4 acrylic glue.

Until tomorrow...
 
Back
Top