Quote from: 2112 on June 26, 2014, 08:01:06 PM

Quote from: Lincoln6Echo on June 26, 2014, 03:35:38 PM

well, the ''magic'' of two phase immersion cooling is that the change from liquid to vapor state consumes a lot of heat energy! That's thermodynamics, too.
So you need a lot less cooling equipment than with air cooling or normal water cooling.

This phase-change occurs only in the "primary loop". I use "loop" somewhat inaccurately because it is a natural convective boiler-condenser setup, not a pumped loop. The Novec is condensed using a normal single-phase water loop with normal radiator and fans at the cold end. They work in series.

Novec immersion cooling is simply a replacement for lack of proper engineering of the mining boards. Bitcoin mining for sure doesn't require large flat boards. All the heat-dissipating components can be directly mounted onto the secondary cooling loop using larger quantity of smaller PCBs. This whole Novec tank is just a distraction here.

Quote from: Lincoln6Echo on June 26, 2014, 07:11:09 PM

As stated before a huge amount of heat energy gets absorbed for changing phase (liquid to vapor/gas) of 3M Novec. That's the advantage to one phase watercooling.

The phase change occurs only in Novec. Then in series with this works a regular single-phase water cooling system.

I'm getting flack for calling people semi-literate or semi-numerate. So what is the better word to describe a person like Lincoln6Echo who seems to have completely missed that the Allied Control cooling system is a series connection of two cooling systems? A mark? A dupe?

I blame this partially on him and partially on deceptive marketing from antirack, and partially on Allied Control being unclear in their marketing materials.

Immersion cooling is still in experimental phase, b/c it is not yet clear what secondary benefits can arise. That's the gamble investors take.

Secondary benefits include:
- rapid deployment
- less/no impact from ambient temperatures
- operating temperature may be adjustable
- higher clock rates may be possible (e.g. 10% increase reduces the capital cost requirements for hashing power, adding to the ROI of the DataTank)

However, to be fair investors need to determine for themselves whether spending bitcoins is worth the gamble. BTC strikes me as undervalued right now, which may confirm what Spondoolies is asserting. But then, by the same logic any investment into bitcoin mining is questionable at the moment, if you don't have access to equipment at cost.