The thermal conductivity coefficient is based on the thickness of the material, not the surface area. That's why it's measured per meter and not per square meter. You have to do calculus to figure out how much energy will be transferred (but in the case of a flat surfaces it's not difficult)
Yeah flat surfaces are super simple. My emphasis on surface area was in a sitution where you've got a simple bath, say an aquarium or other large rectangular enclosure without an exchanger. The surface area of the liquid will dissipate more heat than the glass, as long as the surface is being churned. The outer medium is a given, you could use fans or another cooler material (water/etc) passing over it to help dissipate heat. (some use glass/acrylic mineral oil box inside of another aquarium, as is one of my designs) But with the liquid surface providing the main cooling, it's surface area which counts. In all of the passive units (ie, not using a proper heat exchanger) I've seen and played with, this was the more important factor. This is how I understood the OP's perspective, simply dropping the boards into a large tank and circulating the oil.
It's why I was encouraging they not do just that. It's doable with a low power computer, but it's just going to go the way of fried chicken with ASICs without a great dissipation plan.
Anyway, you would definitely need to keep oil circulating in order to keep things cool. And I think anyone dunking a brand new KnC in a vat of oil is pretty stupid.
I wouldn't think they're stupid per-sé, but they'd need really amazing reasoning to do it. Honestly I'd find a simple watercooling setup (even the pre-built ones) to be just as effective, quiet, and lower maintenance than air cooling. I'd go that route before I dunked ASIC boards.
You should be able to use standard CPU coolers for this so just get some waterblocks if you don't want to go with air cooling.
Yup! Much easier for sure.
Simply pointing out that it is both difficult and expensive to "properly plan & execute" an oil cooling system -- a likely reason for no consumer implementations being on the market.
I think that's part of it, but mainly there's no real market for it. Only a small corner of the tech world really gets into watercooling to the extreme, and only a fraction of them look into submersion--and usually just for the neato factor. There are some consumer kits you can order for submersion of desktop systems, it's just not a popular item at all. Some datacenters use it, as do some other server centers. It's a niche thing.
Not sure where you're getting $2/gal, the price (bought in 55-gal drums), about as "bulk" as anyone here is likely to buy, is over $10/gal, more like $20, all said and done.
I couldn't tell you the exact distributor, just that I've got access to resources to obtain it for super dirt cheap. Maybe I'm not paying the full price, that could be it.
The radiator (cooler, heat exchanger) also needs to be much larger. While aquarium pumps are likely to work (would have to be several -- this stuff is not water, and we're not filtering for fishies -- we want actual flaw), and i've used them for water cooling (the actual pumps/rotors are often exactly the same), i sure wouldn't sleep well with a Rube Goldberg contraption cooling a big chunk of my monyz.
For what its worth, you can get very high quality, positive flow aquarium pumps. Most of the high end submersion systems use precisely that. I was suggesting the simple pumps for flowing the oil across/around the electronics, not the plumbing.
Finally, having made many all-water/no noise gaming rigs which proved as practical/maintenance free as air cooling, i've never come close to making practical submersion-cooled boxen. Doing it "right" would cost a fortune.
I definitely would say it's not practical to submerge when considering the other options! Again for asics, air cooling in open case systems makes great sense, especially if they're in a room with air handling. Other than that, I'd go simple waterblock setups if I had an amazing reason to do so.
So it seems you could buy a couple of 600w (or reuse ones you already have) rather than spending mega dollars on one massive overspecced one.
I dunno why so many people are nitpicking about the price of a good PSU. These systems are thousands of dollars, unique, invaluable in ways, and do not have off-the-shelf user servicable parts. But folks are going to be like "aw man, $150-$200 for a power supply? Ugh! man this sucks!"
Seriously folks?
This isn't a knock on you.. I just don't get the mentality.
