Looking for feedback on my quad sli/cf phase change plans

[Charles Wirth]

Trying to picture what you have planned and not understanding the goal.

You want to cool how many machines and to what temp?

If you have four units, I would assemble two units in tandem and make a two stage cooler with much better performance.

I admire your vision in using the water loop as this performs better than an air/liquid HX and you would see better performance if you used a shell in tube HX for this as this favors the high flow of the water loop. The brazed plate at sizes over 1 ton still are very constricted for flow for this to work well on the first stage.

[Mythicaltech]

Trying to picture what you have planned and not understanding the goal.

You want to cool how many machines and to what temp?

If you have four units, I would assemble two units in tandem and make a two stage cooler with much better performance.

I admire your vision in using the water loop as this performs better than an air/liquid HX and you would see better performance if you used a shell in tube HX for this as this favors the high flow of the water loop. The brazed plate at sizes over 1 ton still are very constricted for flow for this to work well on the first stage.

My plan is to cool a single machine with 4 gpus on the bench. In particular 2x 3870x2 and similar crossfire/sli dual gpus. The reason I am planning on water cooling instead of an air condenser is that I can save massive amounts of desk space which is very limited in college housing. It also will remove a significant amount of heat from the room. The reason I am not going dual stage is for a lot of the stuff I want to do it is too cold and draws too much power and if need more cold I am going to either use the cascade I am getting from you or use ln2 if I am benching a fullpot platform . I will lookinto a shelltube hx.