Hi all,
First of all, I haven't read through every post and so may have missed the following point already having been made in an earlier post, so I apologize if I missed it. And I'm sure you all already know this anyway, but I thought I'd mention it just to be clear.
As long as 100C is far from a phase transition temperature (e.g., the melting point of Si), which I believe it is, then comparing 30C to 85C to 95C to 100C might sound like a large percentage difference in temperature, but it is actually much smaller than it might at first appear. Atom migration (diffusion) in a solid matrix has an activation energy and the kinetics are, I believe, controlled by an exponential dependency on the temperatures. But it is not the temperatures given as degrees Celsius, rather it is the temperatures given by degrees Kelvin that are important. Thus the actual temperatures should be listed as 303K to 358K to 368K to 373K instead of the ones given in degrees C above, and now there is a much smaller percentage change in the pertinent temperatures that go into the exponential functions than when these temperatures are listed in degrees C.
For instance, the difference between 95C and 100C appears to be a mammoth 5% temperature change which sounds like a large percentage change, but in reality it is only a 1.3% temperature change (from 368K to 373K) for the temperatures that count (the physically pertinent absolute temperatures that are included in the exponential functions describing the kinetics of the diffusional process).
An aside example:
A number of years ago an NFL quarterback who shall remain nameless was ensnared in "deflategate" where his game balls were under-inflated allowing for a better grip on the football. A number of writers, including some publicly famous physicists and science writers and engineers, calculated the temperature difference required to yield the measured deflated pressures, but their calculated temperatures were incorrect because they employed degrees F (or the equivalent degrees C) instead of the proper degrees K (or the equivalent degrees R) in their computations. They also employed the gauge pressures instead of the absolute pressures. Their argument was that their computed temperature differences could not be achieved at the ball game and thus cheating (manual ball deflation) had to have occurred. The proper Clausius-Clapeyron equation, employing the proper absolute temperatures and pressures, yielded entirely different temperatures required for the "deflategate" ball pressures, and ones that indeed could be attained in a locker room setting. I, and about a dozen other physicists, pointed out these errors in the computations at that time, and it was eventually publicly admitted that the original computations were indeed incorrect. This was a case of not grasping the underlying physics of the problem and thus simply plugging the wrong numbers into an equation to yield incorrect answers.
But back to the M1 chip temperatures ... my personal opinion, and I'm no expert in solid state physics and silicon and so my opinion is probably worthless, is that the minor increase in atom migration in the silicon of the M1 chip caused by a 1.3% temperature increase is probably a very minor effect on the lifespan of the computer than running the fans to keep the temperature at 95C where the fans's failure mode is caused by mechanical wear on bearings, a much greater risk for failure I would think than the minor increase in atom migration in the M1 silicon. What do you guys and gals think? I'm probably wrong, so please correct me if you have further information.
Regards,
Solouki
First of all, I haven't read through every post and so may have missed the following point already having been made in an earlier post, so I apologize if I missed it. And I'm sure you all already know this anyway, but I thought I'd mention it just to be clear.
As long as 100C is far from a phase transition temperature (e.g., the melting point of Si), which I believe it is, then comparing 30C to 85C to 95C to 100C might sound like a large percentage difference in temperature, but it is actually much smaller than it might at first appear. Atom migration (diffusion) in a solid matrix has an activation energy and the kinetics are, I believe, controlled by an exponential dependency on the temperatures. But it is not the temperatures given as degrees Celsius, rather it is the temperatures given by degrees Kelvin that are important. Thus the actual temperatures should be listed as 303K to 358K to 368K to 373K instead of the ones given in degrees C above, and now there is a much smaller percentage change in the pertinent temperatures that go into the exponential functions than when these temperatures are listed in degrees C.
For instance, the difference between 95C and 100C appears to be a mammoth 5% temperature change which sounds like a large percentage change, but in reality it is only a 1.3% temperature change (from 368K to 373K) for the temperatures that count (the physically pertinent absolute temperatures that are included in the exponential functions describing the kinetics of the diffusional process).
An aside example:
A number of years ago an NFL quarterback who shall remain nameless was ensnared in "deflategate" where his game balls were under-inflated allowing for a better grip on the football. A number of writers, including some publicly famous physicists and science writers and engineers, calculated the temperature difference required to yield the measured deflated pressures, but their calculated temperatures were incorrect because they employed degrees F (or the equivalent degrees C) instead of the proper degrees K (or the equivalent degrees R) in their computations. They also employed the gauge pressures instead of the absolute pressures. Their argument was that their computed temperature differences could not be achieved at the ball game and thus cheating (manual ball deflation) had to have occurred. The proper Clausius-Clapeyron equation, employing the proper absolute temperatures and pressures, yielded entirely different temperatures required for the "deflategate" ball pressures, and ones that indeed could be attained in a locker room setting. I, and about a dozen other physicists, pointed out these errors in the computations at that time, and it was eventually publicly admitted that the original computations were indeed incorrect. This was a case of not grasping the underlying physics of the problem and thus simply plugging the wrong numbers into an equation to yield incorrect answers.
But back to the M1 chip temperatures ... my personal opinion, and I'm no expert in solid state physics and silicon and so my opinion is probably worthless, is that the minor increase in atom migration in the silicon of the M1 chip caused by a 1.3% temperature increase is probably a very minor effect on the lifespan of the computer than running the fans to keep the temperature at 95C where the fans's failure mode is caused by mechanical wear on bearings, a much greater risk for failure I would think than the minor increase in atom migration in the M1 silicon. What do you guys and gals think? I'm probably wrong, so please correct me if you have further information.
Regards,
Solouki
Last edited:
