How many digits of pi would the computer need to drive Redjac out?
Let me put it another way: I don’t think there’s any entropic difference between a computer running a program to calculate the billionth prime and a plain-vanilla heat generator doing no computation.
Fans keep it from overheating by moving hot air out and cold air in. If there’s no cold air to grab from anywhere, there’s no cooling. Your suggestion implies that cold is air is coming in from somewhere–which is adding energy to the system from outside.
Just in case I’m not being clear, I do not believe the thought experiment is violating the 2nd Law.
I’ve wrestled with this and for a while I agreed but now I’m not sure. Say the computer is suspended in a vacuum (in the box). The CPU eventually gets so hot that it radiates the heat away as photons. The inside of the box is coated with 100%-efficient photoelectric cells and pumps the electricity back into the computer. Wouldn’t this be a way to reclaim the heat? (This would assume that energy is pumped into the box until the CPU gets hot enough to create an equilibrium between the energy used and the energy radiated.)
I hesitantly disagree. It seems like the computer is doing work but I admit I can’t quantify why I think so. If you are correct then my theoretical computer-in-a-box can run for a million years (assuming there is a way to recycle 100% of the energy) and I have a hard time believing that.
Your thought experiment is violating the second law. Heat cannot be recycled with 100% efficiency. Period.
Exactly. Assuming there is a way to recycle 100% of the energy. If you could build a break-the-laws-of-thermodynamics machine (a machine that can recycle 100% of its lost heat), then it is going to break the laws of thermodynamics (by running forever without the input of energy).
If you have no difficulty believing that a bit-flipper or a direct electricity-to-heat machine can run for a million years (under the counterfactual assumption of 100% efficiency recycling of heat), why such difficulty believing that a prime calculator can do the same?
The 2nd Law says heat cannot be completely converted into work. If calculating primes is not work then theoretically the computer could continue running. Right?
That would be true if you had no heat source. The CPU is a heatsource. After air travels over it, it’s hotter then the air that hasn’t traveled over it. There’s your temp difference.
Strike that last post I don’t know what I was thinking with the not needing refrigeration track.
That said dividing up the box into sections would work with refrigeration. Hot air into one chamber, cold into another.
It still wouldn’t work because of the 2nd law, but before the power ran out it would still be possible to have warmer and cooler sections without outside intervention
It doesn’t work like that. You’re using energy to split the air into hot and cold. The same hot/cold then needs to power both the refrigerator plus the computer, both of which are adding heat into the system. With heat building, you have to consecutively use more energy to cool it, which means even more energy use. Since the idea is for the device to be self-supporting, “more energy use” is outside of the realm of possibility.
Energy exists as the differential between states, like hot and cold, high and low, or positive and negative charge. The OP is trying to posit a circumstance in which the energy of the differential being released not only powers returning the differential, but also does work along the way. That’s asking for more energy than there even was to begin with.
Computer runs on electron movement. That’s work.
I never said op’s system was feasible. I likened it to plugging a power strip into it’s self. I called it magic up thread too.
My only point is it is possible to have warmer and cooler regions in the box.
I think this related to what I consider the more interesting part of the question: if you have a running computer (disregard all moving parts, assuming it’s just doing random calculations); what happens to the energy you put in it? Is all the energy consumed converted to heat, or is some part of it lost anywhere else?
If all of the consumed energy/electricity is turned to heat, then the answer to the OP would be: it would run forever (assuming some impossible way of turning the heat back to energy 100% efficiently).
I suppose some minuscule fraction might escape as light or vibration, but the heat generated by computation is enough of a problem that an entire cooling industry grew up alongside the advent of computers. Not only is the energy converted to heat, it’s converted so readily to heat that chips must be cooled or they destroy themselves.