F5F Stay Refreshed Power Users Overclocking Out of case refrigerator

Out of case refrigerator

Out of case refrigerator

Pages (2): Previous 1 2
T
theaj42
Member
50
09-05-2016, 11:00 PM
#11
DbDBlackJack :
This is purely theoretical. I aim to achieve the coldest possible liquid around the CPU block as much as feasible, without causing any condensation. No matter the cost or difficulty. The goal here is to boost performance. A lower reported temperature might be for showing off, or a brief spike to snap a photo of that figure. In general, trying this kind of thing is about enhancing performance—not just lowering the CPU temperature. Up to a point, a CPU can run at the same speed even if the core temperature drops from 65°C to 50°C, with no real gain. The processor was fine at 65°C. If your CPU is near the upper limit and actually throttles, this could be an interesting project, though it’s unlikely.
T
theaj42
09-05-2016, 11:00 PM #11

DbDBlackJack :
This is purely theoretical. I aim to achieve the coldest possible liquid around the CPU block as much as feasible, without causing any condensation. No matter the cost or difficulty. The goal here is to boost performance. A lower reported temperature might be for showing off, or a brief spike to snap a photo of that figure. In general, trying this kind of thing is about enhancing performance—not just lowering the CPU temperature. Up to a point, a CPU can run at the same speed even if the core temperature drops from 65°C to 50°C, with no real gain. The processor was fine at 65°C. If your CPU is near the upper limit and actually throttles, this could be an interesting project, though it’s unlikely.

M
MrBeviyolo
Member
55
09-14-2016, 02:11 AM
#12
Just to confirm the setup: two radiators, one fan. The first radiator contains refrigerant, the second is a typical aio (both 120mm for simplicity). The fan sits between them, allowing sub ambient air to flow through the aio. I’m not certain the liquid inside the aio would reach such low temperatures, but with the first radiator, I thought you could control the cooling airflow.

This is purely hypothetical. My goal is to test whether such a system could function. Second, if you can maintain a temperature drop from 65°C under load (not a prime95 workload but a regular gaming load) down to 27°C or 30°C, you’d see better results than just marginal improvements. That means higher clock speeds without using LN2. Third, I’m trying to bypass the condensation problem by placing the unit outside the PC case. If you can maintain the sub ambient temperature inside the cooler, you could set up a reservoir or drain system for any water that collects there. As long as the hose returning to the PC case stays cool.

Now, if you can package this device in a box roughly half the size of a shoebox and only need extended hoses for the aio connected to your case, I’m confident it would be marketable. Actually, it wouldn’t be an aio but would include its own pump and use G14 threads, making it compatible with an EK or Swiftech block or similar.

Drea, I’ll check that link now. Thanks everyone for your responses and suggestions. Appreciate it!
M
MrBeviyolo
09-14-2016, 02:11 AM #12

Just to confirm the setup: two radiators, one fan. The first radiator contains refrigerant, the second is a typical aio (both 120mm for simplicity). The fan sits between them, allowing sub ambient air to flow through the aio. I’m not certain the liquid inside the aio would reach such low temperatures, but with the first radiator, I thought you could control the cooling airflow.

This is purely hypothetical. My goal is to test whether such a system could function. Second, if you can maintain a temperature drop from 65°C under load (not a prime95 workload but a regular gaming load) down to 27°C or 30°C, you’d see better results than just marginal improvements. That means higher clock speeds without using LN2. Third, I’m trying to bypass the condensation problem by placing the unit outside the PC case. If you can maintain the sub ambient temperature inside the cooler, you could set up a reservoir or drain system for any water that collects there. As long as the hose returning to the PC case stays cool.

Now, if you can package this device in a box roughly half the size of a shoebox and only need extended hoses for the aio connected to your case, I’m confident it would be marketable. Actually, it wouldn’t be an aio but would include its own pump and use G14 threads, making it compatible with an EK or Swiftech block or similar.

Drea, I’ll check that link now. Thanks everyone for your responses and suggestions. Appreciate it!

S
sunemoonsong
Senior Member
380
09-14-2016, 08:50 AM
#13
DbDBlackJack :
I'm attempting to bypass the condensation problem by placing the unit outside the PC case. If you manage to maintain a sub-ambient temperature inside the cooler, you can set up a reservoir or drain system for any water that gathers there. As long as the hose leading back to the PC case doesn't become too cold.
The radiator's temperature isn't relevant here.
You need to move this low-temperature material past the CPU and cool it directly. That’s when condensation starts.
From a thermodynamic standpoint, you can’t win—you can’t even break even, you can’t cheat, and you can’t give up.
The overclocking limit isn't just about temperature; it's also about the actual CPU. Just because you've used LN2 in a freezer and outside Antarctica doesn't guarantee this specific CPU will reach an unrealistic GHz figure.
In the end... the CPU performance between 65C and 30C is essentially zero.
S
sunemoonsong
09-14-2016, 08:50 AM #13

DbDBlackJack :
I'm attempting to bypass the condensation problem by placing the unit outside the PC case. If you manage to maintain a sub-ambient temperature inside the cooler, you can set up a reservoir or drain system for any water that gathers there. As long as the hose leading back to the PC case doesn't become too cold.
The radiator's temperature isn't relevant here.
You need to move this low-temperature material past the CPU and cool it directly. That’s when condensation starts.
From a thermodynamic standpoint, you can’t win—you can’t even break even, you can’t cheat, and you can’t give up.
The overclocking limit isn't just about temperature; it's also about the actual CPU. Just because you've used LN2 in a freezer and outside Antarctica doesn't guarantee this specific CPU will reach an unrealistic GHz figure.
In the end... the CPU performance between 65C and 30C is essentially zero.

Pages (2): Previous 1 2