My Unnamed Theoretical Project

[NakuraD]

What stage are you at right now?
Any problems you are running up against?
I was just thinking, is there an issue with the acrylic flexing once the oil is in?

I'm interested to see what you are dealing with design wise. I would love to build an oil cooled setup some day when I have the space and tools. Living in an apartment really makes you miss having a shed.
I like your modifications compared to the (forget the name, P-something) version, yours looks better and will be more sturdy.

[burntheland]

I have a neat idea, but I'm not sure how much prep it would take to get it right. You could "feasibly" have two hoses exiting from the front/right bottom hand side of the water portioned part of the tank. Get a pump to pull the died water through the hoses (which would look sweet as all get out) and pump them to the back of the tank where at the top there will be something that is set to drip. Like a cylinder shaped colander so to speak. And from the side of the tank you could see the silver dyed water fall down back to the bottom. This has absolutely NO functionality. But personally, if I had the means I would definitely steal that idea from myself. lol Although where to mount the pump poses an issue. I don't think a short would be possible as long as the "drip station" were flush against the far inside wall of the tank.

[mDust]

Originally Posted by NakuraD

What stage are you at right now?
Any problems you are running up against?
I was just thinking, is there an issue with the acrylic flexing once the oil is in?

I'm interested to see what you are dealing with design wise. I would love to build an oil cooled setup some day when I have the space and tools. Living in an apartment really makes you miss having a shed.
I like your modifications compared to the (forget the name, P-something) version, yours looks better and will be more sturdy.

I'm still experimenting with materials and trying to 'evolve' my original design. I have had so many crazy-awesome ideas that I had to scrap for budget reasons. However, I figure I can build it and mod it over time to perfect it...I can't wait to clean up after that mod session.
My current problems:
1) I cannot find a silver dye anywhere. This is actually turning out to be harder than I thought it would be to find.
2) Bubbles in the welded joints.
3) No dedicated workspace.
The acrylic actually doesn't flex that much although I haven't built a full scale tank yet. When I do, I'll be using 3/8" or 1/2" thick sheets. The 1/8" and 1/4" scrap I've been working with so far is plenty strong. When the tank is built, I will fill it with water and let it sit for a day or so. If it holds the water, which is roughly twice as dense as the oil, I'll know that it's safe to hold oil. I honestly have no worries about this after experimenting with the weldon. There have been several times where I could not break the weld by hand. It's very strong...

Originally Posted by burntheland

I have a neat idea...

Yeah, I thought about stuff similar to that. It would look pretty badass. I may incorporate that into it at a later time. I was thinking about a very small pump from an aquarium or something and punching holes in the hose so it sort of bubbled out and floated back down all the way around the top four edges. I'd rather not risk oil currents sending the water bubbles towards anything with electricity in it, so I would use 1/8" clear acrylic to make a protective wall around the submerged parts. I'd hope that such a thin edge would go mostly unnoticed in the oil as the edges are really all that will be seen anyway. We'll have to wait and see though, I'm currently throwing around so many variables and new ideas that it's actually hindering my progress.

[burntheland]

Yeah, when you first stated "silver dye" i thought to myself..."self, where the hell is he going to find something like that?"

But really man the idea you've got here is great. This thing is going to turn out beastly!

[pheenix]

When I saw the first pic, I thought "hey cool". Then when I saw the next pic with the hood lifted, I thought "I want one".

[mDust]

Originally Posted by pheenix

When I saw the first pic, I thought "hey cool". Then when I saw the next pic with the hood lifted, I thought "I want one".

Haha, thanks.

Although I've left this thread to fend for itself for a long time, I've still been modifying the design of this thing. I don't think it will need the pump and rads anymore, but it will still be actively cooled...just using more extreme methods. Due to this I'm having to build in some interesting workarounds. I'll probably not be using mineral oil anymore. The substitute oil is still clear but it's water-thin at room temp and doesn't freeze until -200*F or so depending on which grade I end up with...hint, hint... I'm not entirely sure the newest design is feasible yet, so if it fails I will fall back to the simple mineral-oil-pumped-through-rads design.

I've been looking for very detailed info on acrylic such as thermal expansion/contraction and some sort of rating for how brittle it becomes with decreasing temperature. I did find the contraction information which turned out to be 'scary amounts', but since I'm not an engineer, I don't know what terms to search for in Google to find information on brittleness. Perhaps someone here can point me in the right direction?

Also, I can get argon gas from a welding supply store, but does anyone know where I can could purchase krypton gas? I'd just like to compare performance with prices.

[Konrad]

Awesome concept mDust ... I especially like the enclosed lid. Good CAD models, too.

Have you looked at engineered dielectric fluids like Midel 7131?

You obviously want to keep your mechanical drives and I/O connectors dry. You might need to hermetically isolate them from the liquid sections or (carefully) apply conformal sealants to prevent condensation. Not a good mod to display an exposed HDD hack, lol.

I'm personally not certain your passive-circulated cooling scheme would be enough ... it seems like a good concept, but basically I'm a little skeptical about all the "hot" oil moving away from the hot parts into the rads. I think I would personally force circulation with some kind of pump, and maybe even use some strategic peltier spot cooling.

[mDust]

Originally Posted by Konrad

Awesome concept mDust ... I especially like the enclosed lid. Good CAD models, too.

Have you looked at engineered dielectric fluids like Midel 7131?

You obviously want to keep your mechanical drives and I/O connectors dry. You might need to hermetically isolate them from the liquid sections or (carefully) apply conformal sealants to prevent condensation. Not a good mod to display an exposed HDD hack, lol.

I'm personally not certain your passive-circulated cooling scheme would be enough ... it seems like a good concept, but basically I'm a little skeptical about all the "hot" oil moving away from the hot parts into the rads. I think I would personally force circulation with some kind of pump, and maybe even use some strategic peltier spot cooling.

Thanks.
I was looking at Halocarbon .8 Oil for its low temperature performance and cloud-point, but I'll look into Midel as it's biodegradable and non-toxic. Although, Halocarbon isn't toxic either. I just don't want to have dangerous chemicals in my apartment because if a spill were to occur, I'd really not want to explain to my landlord and neighbors why I was letting Hazmat into the building.

And that original design includes a pump with a top intake so the hottest oil pumps into the rads and is returned to the bottom of the tank.

[Konrad]

lol, I use Midel fluid in my capillary heatpipes. Vapour point is about 50C, so I didn't need to put much vacuum into them; plus I'm not especially worried about electrical failure if the tubes ever leak.

Flourinerts used to be popular, used in SGI machines I hear. Second to none in efficiency, but nasty toxic I think, and $$$ per gallon, lol. Pure oil is still a good choice; gotta keep it away from oxygen (it starts breaking down and getting "rancid"), and be careful that it doesn't have any contaminants. (Whichever fluid you use, test resistance and capacitance with meter.)

Aaah, didn't catch the pump in your description. I thought it was passively self-circulating. Impressive work.

[mDust]

Originally Posted by Konrad

Flourinerts used to be popular, used in SGI machines I hear. Second to none in efficiency, but nasty toxic I think, and $$$ per gallon, lol.

More like $$$$$$$$$$$$$$$$$...it would cost around $17-18,000 for about 10 gallons. No thanks!

Whatever fluid I use must be completely clear and as inert as possible. It will also be a fluid that is designed for this purpose so testing shouldn't be a problem.