Thoughts on Refrigeration
Thoughts on Refrigeration
And this 7C provides additional benefits beyond confirming a stable 7C processor speed.
USAFRet:
This 7C means more than just confirming a stable 7C system—what about the extra costs and the worry of contributing to climate change?
USAFRet:
And this 7C gives you more than just confirming your PC is stable at 7C?
That's a good point. If I can keep it steady before pushing it overclock, then I have about 93 C left before the Intel chip starts slowing down. The goal is to offer better cooling than regular air coolers, not necessarily lower temps. For example, if room temperature is around 23C, a refrigerant cooler would give me 16 degrees more headroom than an ambient cooler. Remember, all cooling solutions are limited at the component level—this is why chip temperatures climb until balance is reached (heat output equals heat removal). The aim is to boost heat dissipation by lowering the surrounding temperature by 16C, enabling overclocking by that extra amount. It's not about making a chip cooler, but turning that into more performance. Some people spend days fine-tuning their systems for a 4% speed boost, and there are competitions focused on this kind of challenge.
I prefer a cooler that can handle 1800W at 50C rather than one that only manages 600W at 7C. With the latter, I could push the motherboard further, but I still have to think about power consumption and noise levels.
Another option could be to add an AC compressor to a regular liquid cooler, cooling the glycol only when the sensors indicate temperatures over 60°C. This approach may be more effective as it complements the usual heat removal from the water cooler.
My concern is whether lowering the temperature to 7C provides any real benefit.
If you're pushing the CPU to its maximum performance at 45°C (which is what a decent water cooler can handle), and also want extra headroom for overclocking, then it might help.
However, if you're already pushing the design to its limits in terms of power, noise, or reliability, continuing to increase OC could be risky.
I'm unsure how this would add value in practice.
For an engineering project or a personal build, try it out.
And if you need a real-world result, build it and test it.
Yeah, I just took some time looking at mineral oil aquarium designs. A big benefit is that it covers the entire surface for heat transfer. However, as others pointed out, the heat removal isn’t sufficient for long-term overclocking. I’m considering combining an aluminum heat sink with fans on all sides. Sure, it’ll cost a lot of expensive aluminum, but it could fix the cooling issue.
Some people suggested placing a refrigerator condenser in the oil, but the compressor failed. Maintenance and upgrades would then become quite complicated.
I’m still in the early stages of thinking about this. Clearly, any compressor would need to be able to run continuously.
Regarding your question, if it can operate at 100% efficiency at 7°C, you should be able to maintain a steady 130% at 45°C since the system could handle an additional 600W of heat. At least that’s the concept.
This makes me lean more toward enhancing a typical liquid cooling setup rather than using a freezer-style design. It brings up a whole different set of factors to consider.
Good decision to skip oil immersion, it’s a complete mess to work on.
Even with cooling methods like liquid nitrogen, an overclock hits its limits depending on other variables—temperature being the main factor. Voltage settings, component stability, and especially VRM cooling requirements all play a role.
A solid air or closed-loop liquid cooling system is simpler to set up and maintain.
If you opt for air cooling, a high-end cooler paired with liquid metal thermal paste works well. A Kaby Lake i7-7700K at 5GHz will only reach around 80°C under heavy load, which is uncommon except during tests, as shown here.
Keep in mind that liquid metal demands more frequent replacement compared to regular thermal paste.
Still, if you’re aiming for lower temps during intense overclocks, it might help to consider the cost and effort involved.
The main concern remains whether your needs and budget support the extra work, though the hobby aspect can add to the price.
The temp v performance curve isn't completely linear. At some stage, additional factors begin to affect the anticipated performance. Things like voltage, temperature of other parts, and stability come into play. Raising the voltage and frequency usually helps, and with a solid CLC it's unlikely the maximum temperature will be reached before other elements become the limiting factor. Not the least of these is the meat sack behind the keyboard.
The temp v performance curve isn't completely linear. At certain points, additional factors begin to affect the expected results. Things like voltage, temperature of other parts, and stability come into play. Raising the voltage and frequency usually helps, and with a solid CLC it's unlikely the maximum temperature will be reached before other elements become the limiting factor. The biggest issue here is the meatsack behind the keyboard.
EDIT: Most of that was covered in the previous message...lol
^ yes, exactly this point, and as I mentioned before "the meatsack sitting behind the keyboard." :lol: