• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

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

Page 7 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
From what I've heard, IX works great if your IHS and HSF base/block/etc are flat. If they're bowed or uneven somehow, it still works, but not as well as it could.

Be absolutely sure to follow the instructions so that you get a proper reflow before you dunk it in the oil, of course.
 
Legoman666, this whole thread is great. EXCELLENT work with all the fabrication you've done and way to be persistent and stick with it. Incredible stuff!
 
Legoman666, this whole thread is great. EXCELLENT work with all the fabrication you've done and way to be persistent and stick with it. Incredible stuff!

Thanks! It's been a frustrating journey.

IMG_2225.jpg

My Indigo extreme kit came in this week. Been busy with senior level enginering courses, so I haven't had the chance to mess with it until today.

IMG_2223.jpg

The problem is that the indigo extreme needs to be applied when the motherboard is flat. Seeing as how I can't just tip my case over (bad Bad BAD), I have to remove the motherboard first. So I drained a few gallons from the tank, pulled the motherboard out of the oil, and am now waiting for as much oil as possible to drip off before I remove the motherboard from the tray.
 
WARNING!

Be absolutely sure to follow the directions to the T with that Indigo! You need to make sure you're getting a good contact AFTER reflow. This cannot be stressed enough since it's a PITA to remove the board from your box...

Not sure if it will improve CPU temps much - remember the best TIM is no miracle if you have nothing "upstairs" efficient at pulling the heat away from the block. 😉
 
The Pepsi and little Easter Island head are the key components of this project.

Dammit I knew I was missing something.

Great news! Got the Indigo Extreme installed and made a few modifications to the GPUs. Temps are excellent now! Woohoo, I can actually use my computer for its intended purpose.

IMG_2230.jpg


IMG_2232.jpg

Following the very specific directions for the indigo extreme, I applied the TIM to the surface of my chip. The TIM itself comes in a hard plastic case and has protective plastic on both sides of it. Solvent (HFCs from the smell of them) wipes are provided in order to get the base of the heatsink and the CPU as clean as possible.

IMG_2236.jpg

This elaborate setup is needed because dumb#$%&#$ Windows 7 apparently installed some important files on my RAID array instead of the SSD. Le sigh. Rather than move the hdds to the motherboard (which I would have rather set up on the floor than on top of my case...), I moved the mobo to the hdds.

IMG_2234.jpg

Indigo extreme requires high temps to melt it in order to spread it evenly over the CPU. Prime95 /w 8 threads on a passively cooled heatsink. Toasty.

temps-1.png

The moment of truth! These temps are after 100% CPU load and 100% load on both 5850's for ~45 minutes. Oil temp is 37C. Most excellent. The whole machine is pulling 445W from the wall.

IMG_2245.jpg

To help push the oil through the CPU heatsink, I moved the fan that was pushing oil onto the GPUs onto the back of the CPU heatsink. I don't think it was doing any good in its old location.

IMG_2229.jpg

To help reduce the GPU temps, I removed the silicone seals, reapplied AS Ceramique and reattached the fans and shrouds. I remove the grill on the side of the shroud in order to remove oil-flow impedance.

IMG_2246.jpg


Here you can see the GPUs. I removed the GTS250, It was too close to the edge of the tank to get adequate circulation. Additionally, It was blocking the oil intake on the second 5850. So out it went!

IMG_2250.jpg

Oil in/out temp after about 1.5 hours of full load on the CPU and GPUs. Not bad.

IMG_2239.jpg

Now... who's going to clean up this mess for me?

Before:
CPU: 60-70C idle, 98C under load and throttling to prevent overheating
5850s: 80C idle, 99C under load and throttling to prevent overheating
GTS250: hit 127C idle before throttling itself down to 157mhz and hovering around 90C

After:
CPU: 35C idle, 65C load
5850s: 35C idle, 55-60C load

All that's left to do is see how this beast overclocks...
 
Yes o/c and stress for several hours and note the starting oil temp and cpu/gpu temps and their relationship, etc.
 
Yes o/c and stress for several hours and note the starting oil temp and cpu/gpu temps and their relationship, etc.

I got a second data point for calculating the pump flow rate.

Properties of oil:
Specific gravity: 0.9 g/cm^3
Heat capacity: 1.966 J/(K*g)

Power draw of computer idle: 270 W = J/s
deltaT of oil In/Out: 1.55C

Power draw under load : 480W
deltaT of oil In/Out: 2.60C

(270 J)/s * (cm^3)/(.9 g) * 1/(1.55 K) * (K*g)/(1.966 J) = 98.45 cm^3/s -> 93.6gph
(480 J)/s * (cm^3)/(.9 g) * 1/(2.60 K) * (K*g)/(1.966 J) = 104.34 cm^3/s -> 99.2gph

And if you want to get more accurate, we can take into account the convection on the tank itself to the air:

Tank outside walls are .453m^2 and .00635m thick.

k (air) = .025 (assumes no forced air movement over tank surface)
k (plexiglass) = .20
k (oil) = .138
k total is the inverse of the sum of the inverses = .0191 W/(m*C)

W = k*A*deltaT/L, So heat lost through tank at load is 20W and 14W at idle. (deltaT between oil/air of 15C at load and 10C at idle)

Recalculating the flow rate with a new power value gotten by subtracting the heat lost through the tank from the total power consumption yields:
load: 99.99cm^3/s
idle: 93.34cm^3/s
Which isn't really much more consistent, but I'm not sure if I'm taking into account the convection correctly. Rubycon, any input?

CPU temp is rock solid at ~65C after 100% load for about 4 hours.
 
Last edited:
Great news! Got the Indigo Extreme installed and made a few modifications to the GPUs. Temps are excellent now! Woohoo, I can actually use my computer for its intended purpose.

Very nice! It's good to see that you got the reflow going properly. Some folks never get reflow, though it's not very common.

Should be interesting to see how that monster OCs.
 
Where are you getting your flowrate from?

The mass of the oil is fairly significant so you would need to test under full load for perhaps a few hundred hours and you will see a rise. 😱
 
SG ratings of pumps are for water. Oil has a high viscosity which will pull this number down!

Try for another 4 hours - it will eventually stabilize. Then again how stable is your room temperature?

The numbers will be comparable to mid range air on the cpu and low water on the gpu.
 
Here's some pics that I forgot I took. I snapped them before I installed the Indigo Extreme though, so the GTS250 is still there.

mineral&


IMG_2211.jpg


IMG_2213.jpg


IMG_2216.jpg


IMG_2221.jpg

Slow shutter without the nasty flash using a tripod is really the way to capture the glow.
 
Something interesting:

Uploads&


Oil is wicking along the inside of the display port cable for one of my monitors. It's wicking 6ft from the inside of the case to the display port input on the monitor itself. Interestingly, the other two monitors cables, which are DVI, are not experiencing this phenomenon.

I only discovered it when I touch the base of my middle monitor and it came away covered in oil. I had a big WTF moment before I figured it out. The reason that only the display port cable is doing this is because it is a male connector about 5/8" long, DVI connectors are female. Thus, the display port cable is dipping into the oil on the video card side whereas the DVI cables do not.

Strange.
 
Wow you need to seal/pot the inside of any connector that is submerged! At least oil is an insulator. 😉
 
Back
Top