Mineral Oil PC - 4/7/2010 Houston, we have liftoff!! *56K Warning*

[Rubycon]

Ah inline sensors should not be at an elbow. Yes it will work but...

You should have just kept a T and put it inline and put the sensor through the top.

A pretty good inline holder for probe type sensors can be made from a 1/2" NPT PVC plug. Drill out the center 1/64" larger than the probe and secure with good epoxy such as Belzona. This can be screwed into a threaded T. If the probe needs to be removed you can replace with a standard plug and keep running with little interruption.

 

[legoman666]

Ah inline sensors should not be at an elbow. Yes it will work but...

You should have just kept a T and put it inline and put the sensor through the top.

A pretty good inline holder for probe type sensors can be made from a 1/2" NPT PVC plug. Drill out the center 1/64" larger than the probe and secure with good epoxy such as Belzona. This can be screwed into a threaded T. If the probe needs to be removed you can replace with a standard plug and keep running with little interruption.

Hm, would epoxy hold up in oil?

I got a GTS250 for "Free" with my MSI Big Bang mobo. I think I'll use it for PhysX in conjunction with my 2 ATI 5850's.

So it needs to be prepped :dremel:

Don't want these cheapo underpowered fans in my oil. Besides, they make the card take 3 slots instead of 2.

I sealed around the caps with a generous amount of Dow Corning 735 silicone. This will prevent oil from getting into the rubber thats on the bottom of the caps.

Since I had the HSF off anyway, I replaced the TIM with some Arctic Alumina. Then I ran a bead of silicone around the top of the shim to prevent oil from getting to the GPU.

You can make out the squished bead of silicone on the shim here.

The card reassembled.

Both of my 5850s will get the same treatment. Here's the bare card.

There weren't as many caps to seal around; only 4. I replaced the TIM here too.

And here it is without the shroud and fan. This is how it'll sit in the oil. The natural convection currents will take care of the cooling.

Letting the silicone set. Once I'm ready to get my case running again, I'll do the same thing to my 2nd 5850.

 

[Rubycon]

Epoxy is very resistant to mineral oil attack.
The longer the cure time (generally) the better the resin. Faster curing time is nice but you give up (mainly) bond strength. 90 second work time epoxy for emergency line repairs will hold 100X the pressure you have at the bottom of your box though!

Prep is extremely important and often overlooked or improperly done. Quite possibly because it's probably one of the worst things you have to do in the whole project!

But as you know if you have to tear it down because of a leak or other problem - doing it right pays! We always say do it nice not twice AND (for those in a hurry) if you cannot find the time to do it (right) how the hell are you going to find the time to do it again?!

 

[legoman666]

Epoxy is very resistant to mineral oil attack.
The longer the cure time (generally) the better the resin. Faster curing time is nice but you give up (mainly) bond strength. 90 second work time epoxy for emergency line repairs will hold 100X the pressure you have at the bottom of your box though!

Prep is extremely important and often overlooked or improperly done. Quite possibly because it's probably one of the worst things you have to do in the whole project!

But as you know if you have to tear it down because of a leak or other problem - doing it right pays! We always say do it nice not twice AND (for those in a hurry) if you cannot find the time to do it (right) how the hell are you going to find the time to do it again?!

http://www.economyrubber.com/610-Brass-Compression-Male-Run-Tee/M/B001S57ZI6.htm something like this? And then buy a 1/2 NPT female threaded cap? (You have no idea how much work it was to find a male run tee that fits 5/8" tubing...)

 

[Rubycon]

http://www.economyrubber.com/610-Brass-Compression-Male-Run-Tee/M/B001S57ZI6.htm something like this? And then buy a 1/2 NPT female threaded cap? (You have no idea how much work it was to find a male run tee that fits 5/8" tubing...)

Yes that should work. What kind of sensors do you have?
You know they have sensor wells for this kind of thing too. They won't leak when properly plumbed and can take 1000's of PSI (for monitoring temperatures of hydraulic lines, for example).

All of my "hands on" designing has been with hard lines and fittings. The concept is the same, however.

 

[legoman666]

http://www.crystalfontz.com/product/WRDOWY17.html

 

[cbn]

Legoman666,

This is some very nice work you have done (to say the least). Actually I am blown away by it. Thank you for sharing.

 

[Rubycon]

Wow a T092 package immersed in oil.

That just screams custom pot job!

Sleeve the leads (polyolefin heat shrink) and slip down into the tee so the top of the package protrudes about 2mm into the bore of the line so it will be actively submersed. Pack in epoxy putty and allow to set up making adjustments as needed. Once it cures you have an inline sensor that's leak proof!

Make as many as you like!

You can also make proof of flow/pressure sensors using the same technique as well.

This kind of stuff is actually fun.

 

[legoman666]

Wow a T092 package immersed in oil.

That just screams custom pot job!

Sleeve the leads (polyolefin heat shrink) and slip down into the tee so the top of the package protrudes about 2mm into the bore of the line so it will be actively submersed. Pack in epoxy putty and allow to set up making adjustments as needed. Once it cures you have an inline sensor that's leak proof!

Make as many as you like!

You can also make proof of flow/pressure sensors using the same technique as well.

This kind of stuff is actually fun.

So would you recommend spending $45 on new elbows with caps or just go with my current jobbie? I can easily remove the silicone from my PVC fittings and replace it with epoxy putty.

And I'm not sure my mental image of what you have in mind is what you're intending. Can you draw a diagram? (Nothing fancy, mspaint is fine )

Legoman666,

This is some very nice work you have done (to say the least). Actually I am blown away by it. Thank you for sharing.

Thanks! It's been a learning experience for sure... Even though this is take 2, I'm apprehensive about dunking a new $300 mobo, a new $200 cpu and 2 $250 video cards into oil.

 

[Apocalypse23]

That's a very impressive setup legoman666...I wish I was as skilled a handy man as you! cheers.

 

[Rubycon]

And I'm not sure my mental image of what you have in mind is what you're intending. Can you draw a diagram? (Nothing fancy, mspaint is fine )

Hope this helps!

 

[legoman666]

That's a very impressive setup legoman666...I wish I was as skilled a handy man as you! cheers.

Cheers! It's easy to learn, just start working!

Hope this helps!

Indeed! Thanks.

 

[Rubycon]

Next up!

Oil quality monitoring!

http://www.igcar.ernet.in/benchmark/Tech/23-tech.pdf

Simple yet effective.

Now we're getting into overkill territory as the oil in a transformer (or enclosed azipod! ) can suddenly take a turn for the worse if it goes unchecked.

In the case with our azipods it's especially critical to monitor the oil for contaminates (since they operate submersed in the ocean!) as well as periodic metallurgic analysis for traces of metal from thrust bearings, etc.

You certainly don't have all THAT going on inside a PC case. The idea may even be novel kind of like diatomaceous earth filtration to the casual aquarist. But the techie drones thrive on this stuff, right? Right?

 

[legoman666]

Next up!

Oil quality monitoring!

http://www.igcar.ernet.in/benchmark/Tech/23-tech.pdf

Simple yet effective.

Now we're getting into overkill territory as the oil in a transformer (or enclosed azipod! ) can suddenly take a turn for the worse if it goes unchecked.

In the case with our azipods it's especially critical to monitor the oil for contaminates (since they operate submersed in the ocean!) as well as periodic metallurgic analysis for traces of metal from thrust bearings, etc.

You certainly don't have all THAT going on inside a PC case. The idea may even be novel kind of like diatomaceous earth filtration to the casual aquarist. But the techie drones thrive on this stuff, right? Right?

That would be neat! However, I was under the impression that oil only degrades if there is arcing taking place. I certainly hope I don't have any arcing! Hehe.

Excuse the low quality pics, took them with my cell phone.

My motherboard is going to crowded.

Leak testing the tank! If all goes well, I'll move my pc into the oil on Tuesday.

 

[legoman666]

Putting everything into the case.

I had to modify the bracket on the GTS250 on the end. It was a double height, now it's single height. :dremel:

Fired up the system for the first time. It has no active cooling so I only let it run long enough to see that it worked.

Excuse the poor cell phone pics

 

[ehume]

Those are great shots. Big Bang indeed. That's what you hope you don't have.

This is one of the more impressive things I've seen. No active cooling? What are your temps at idle?

 

[Rubycon]

That would be neat! However, I was under the impression that oil only degrades if there is arcing taking place. I certainly hope I don't have any arcing! Hehe.

The oil prevents arcing (corona) in addition to cooling...

Problem is it does degrade over time. Exposure to moisture, contaminants, you name it. Quality goes down as soon as it hits the air when poured out of the drum.

 

[legoman666]

Those are great shots. Big Bang indeed. That's what you hope you don't have.

This is one of the more impressive things I've seen. No active cooling? What are your temps at idle?

Thanks. I don't know the temps yet, I only let it run long enough to take a picture of the POST screen. There's no oil in it at the moment so I didn't want to leave it on for too long because the video cards don't have any active cooling.

 

[legoman666]

Worked on the case a bit more tonight but not as much as I had wanted. I really wanted to finish it today, but my electronics design lab is sucking hardcore, so I didn't get home until lateish.

I decided to dedicate a fan to circulate oil directly over the GPU heatsinks. Seeing as how there are three GPUs, there is a lot of heat that needs to be dissipated. So this is the mounting bracket I made for this purpose.

I bolted it to the top plexiglass "lid".

And from the front.

I made installed these little pieces of bent aluminum to hold the wires going into the tank in place. I was having problems with them not staying put and the zip ties getting snagged, so these should fix that.

I also made a new hard drive mounting bracket. The only difference between this one and the old one is that the drives are now mounted top side up (when the lid is closed). Previously, with the drives upside down, the sata and power plugs were so close to case top that I had trouble getting the plugs in. This fixes the problem. :thumb:

It's almost ready! There are a few things left to do: mount the SSD, clean the tank, clean the aluminum, and probably a few other things I'm forgetting at the moment.

Till tomorrow...

 

[legoman666]

Filling... (~3 gallons)

Filling... (~5 gallons)

Full! (~7.5 gallons)

Houston, we have liftoff!

Fan spinning slowly (~300rpm is my guess, no RPM sensor though)

Weeeeeee.

Installing Windows 7 at the moment. Can't wait to OC this behemoth. 270W idle power draw.

 

[ehume]

Is that red glow at the bottom Hades peeking up to have a look? Or just a crack in the Earth's crust?

 

[alfa147x]

Nice build! How thick is that top padding?

 

[legoman666]

Will definitely try to get a video up soon. However, it sprung a leak late last night (right before I went to bed) so I drained it. The tank is not leaking, and I doubt one of the fittings is leaking, so it's probably one of my temperature sensors or the bulkhead fittings. I won't be able take a look at it until I get home from class later today.

Neither is difficult to fix, but if it's the temp sensors, I'll need to take some time to redo the seals with epoxy (instead of silicone) which would require me to completely clean the polypropylene T's. If it's one of the bulkhead fittings, I'll just need to tighten the nut or re-seat the o-ring. However, if it's something else (I don't even know what) it'll take more time/effort to fix.

 

[Rubycon]

Will definitely try to get a video up soon. However, it sprung a leak late last night (right before I went to bed) so I drained it. The tank is not leaking, and I doubt one of the fittings is leaking, so it's probably one of my temperature sensors or the bulkhead fittings. I won't be able take a look at it until I get home from class later today.

Neither is difficult to fix, but if it's the temp sensors, I'll need to take some time to redo the seals with epoxy (instead of silicone) which would require me to completely clean the polypropylene T's. If it's one of the bulkhead fittings, I'll just need to tighten the nut or re-seat the o-ring. However, if it's something else (I don't even know what) it'll take more time/effort to fix.

Most adhesives (epoxies included) don't adhere well to polypropylene. Roughing the surface may help but in the end a poor bond still exists.

 

[legoman666]

I roughed up the interior surface of the fitting with a file before I globbed silicone in there. Maybe I didn't get it rough enough.