Overheating Cpu

[benjiro]

The Pandora uses a new revolutionary and patented technique to deal with heat.

"heat > /dev/null"

Problem solved.

Lets pull in another example, my old time favorite. The Dell x50v. 624Mhz Xscale ARM cpu, dedicated graphics hardware etc, build upon a old manufacturing process ( 0.13-micron ). Gets a little warm to the touch, but aint going o burn your hand after using it a few hours. The Omap3530 is build upon a 0.065-micron process... Its going to be mighty strange that it gets extreme hot. Ofcourse, there is also the option open for 0.045-micron in the future.

Also, take a look at the layout. The Omap is placed linear to the SD cards.Even when there are cards inside, its not going to be perfect 1:1 match. In other words, there is also a slight amount off opening there for hot air to escape. For maximum heat dissipation, a metal or magnesium based case will have been best, but thats going overboard.

 

[multiplex]

Argh. Believe me, it will all end up as heat _eventually_. Including the light, when it gets absorbed somewhere (i.e. some of that will end up as heat on your walls, your retina, anywhere where the light gets absorbed. And it _will_ get absorbed.)

Unless, of course, the Pandora actually gets heavier with use.
I theorize that it will in fact suck thermic energy from its surroundings, and expel the energy as small pellets of plutonium. Users will complain about cold and irradiated fingers, triggering a recall.
Pandora, Inc. will be left with a large amount of unsellable stock, and bankruptcy will seem all but certain when they at the last minute strike a mutually beneficial deal with Lead Mittens, Inc., setting the stage for a rerelease.
It will of course launch to the rave reviews we all know it deserves, the big players will finally see this large untapped market, and 2010 will finally be the year of the heavy metal finger-wear.

... Ah, sorry. I think I may have veered slightly off topic. :unsure:

 

[Chip]

How do we reverse Entropy?

You don't. It's against the law.

But seriously, armchair physics aside, actual observations always trump "theories". Hopefully Squidge won't mind me quoting his findings from the dev forum:

I've tested the OMAP+RAM (as it's stuck on the top) temperatures today using a non-contact temperature sensing gun.

@ 550Mhz (standard board frequency from 1.0 Kernel), after 2 hours I get 31c (88f)
@ 925Mhz (plus vdd1 opp value of 5) after a further 2 hours I get 32c (90f)

I then measured the temperature of my cup of tea... 88c (191f)

 

[Chaosstorm]

What about googling a bit about the cortex, i found some info on the arm website saying "With the ability to scale in speed from 600MHz to greater than 1GHz, the Cortex-A8 processor can meet the requirements for power-optimized mobile devices needing operation in less than 300mW"

Heres a link Click!

i hope this was enough information

 

[Phawx]

How do we reverse Entropy?

You don't. It's against the law.

But seriously, armchair physics aside, actual observations always trump "theories". Hopefully Squidge won't mind me quoting his findings from the dev forum:

I've tested the OMAP+RAM (as it's stuck on the top) temperatures today using a non-contact temperature sensing gun.

@ 550Mhz (standard board frequency from 1.0 Kernel), after 2 hours I get 31c (88f)
@ 925Mhz (plus vdd1 opp value of 5) after a further 2 hours I get 32c (90f)

I then measured the temperature of my cup of tea... 88c (191f)

Just in case, I was referencing This Short story

Or you could ask Multivax: Here

 

[hch]

Every type of energy ends up as heat in the end.

sorry to nitpick and be ot, but this is so widely inaccurate that i need to clear up a few things. actually, 96% of the universes energy is of yet unknown nature (dark enery, dark matter) and the rest is mostly rest mass of atomic nuclei (mass is also a form of energy).

now to be more specific: any electronic device such as the pandoras cpu gpu etc. that works in the radio frequency range emitts a fraction of its energy as radiation. and yes, the fraction that gets radiated off vs. the fraction that is dissipated in the chip depends on the fabrication thechnology (the 90nm pentium4 prescott e.g. was known to be very inefficient due to large leakage currents), the core voltage and also the frequency.

now what might really surprise you is the fate of some of the radiation (photons). at the ~1GHz range that the pandora will certainly emmit at, earths atmosphere is almost transparent. so a large fraction of these photons will work up against earths gravity field, loose energy (get redshifted), escape earth and most probably also the solar system, loosing even more enery and the galaxy, loosing even more energy to the gravitational potential they climb. now it so happens, that the whole universe is transparent to such photons. the chances that such a photon will ever interact again are tiny and decrease as the universe expands. so they will never thermalize and become 'heat'. instead, they will gradually loose their energy to the expanding universe (energy is not conserved on these scales).

 

[CandidStan]

Yea....
@CandidStan, You can put it in your a....
But it's good to hear that there's no big risk of overheating the pandora's CPU.
Have pandora on PCB the "thermometer" on the cpu?



HAha, your so witty, really you actually worried about the ARM overheating?

Hundreds of cell phones use an ARM processor, and none have caught on fire or melted because of poor ventilation.

That's why I posted the USB fan, if the Pandora needed a fan or some other sort of cooling method, don't you think Craig and the team would have put one in?

The Nintendo DS (every model) uses two ARM processors. Granted, they are not as powerful as the one going into the Pandora, however not a single model of the DS has a single ventilation hole, nor even a copper heatsink.

If there was a need, it would be on the device

A temperature sensor would be easy enough, however I think it would be useless.

However if you are that paranoid about your pandora frying itself because of lack of cooling, be my guest and install whatever it takes.

As you already know the final cases are about to be whipped out in the next week or so. So if you are suggesting that the team change the design at this stage, your nuts.

 

[lardman]

QUOTE
sorry to nitpick and be ot




QUOTE
so a large fraction of these photons will work up against earths gravity field, loose energy (get redshifted), escape earth and most probably also the solar system, loosing even more enery and the galaxy, loosing even more energy to the gravitational potential they climb. now it so happens, that the whole universe is transparent to such photons. the chances that such a photon will ever interact again are tiny and decrease as the universe expands. so they will never thermalize and become 'heat'. instead, they will gradually loose their energy to the expanding universe (energy is not conserved on these scales).


Photons have no mass, so no work will be done, nor need it be done to overcome the Earth's (or any other body's) gravitational field.

You're confusing redshifting (whereby the wavelength of the photons is increased due, in the case of photons in space, to space being stretched, i.e. the Doppler effect) with simple absorption and re-emission of the photons by some atom/molecule which happens to be in their path. The absorption causes an electron in the atom/molecule to enter a higher energy state, and if/when it drops back down, the photon(s) emitted may have a lower energy (i.e. longer wavelength) than that absorbed.

Of course if the universe continues to expand, these photons will eventually be red shifted back down into the infra-red part of the spectrum, and therefore will appear as heat (though the flux will be pretty low by the time the universe has expanded that much!

 

[hch]

sorry for the wildly ot, but i kind of started it, so let's finish.

Photons have no mass, so no work will be done, nor need it be done to overcome the Earth's (or any other body's) gravitational field.

Photons have no rest mass, but their energy is proportional to their frequency. and yes, they get redshifted (=frequency gets lower=they loose energy) when climbing a gravitational potential. this is an effect of general relativity, most evident in black holes, where the 'potential' is too large for the photons to climb so they can't escape.

You're confusing redshifting (whereby the wavelength of the photons is increased due, in the case of photons in space, to space being stretched, i.e. the Doppler effect) with simple absorption and re-emission of the photons by some atom/molecule which happens to be in their path. The absorption causes an electron in the atom/molecule to enter a higher energy state, and if/when it drops back down, the photon(s) emitted may have a lower energy (i.e. longer wavelength) than that absorbed.

what i am saying is, that a typical photon of about 1GHz frequency will have a rather large chance to never ever again interact with anything at all.

Of course if the universe continues to expand, these photons will eventually be red shifted back down into the infra-red part of the spectrum, and therefore will appear as heat (though the flux will be pretty low by the time the universe has expanded that much!

well, first of all, the infrared part of the spectrum is shorter wavelength than radio wave, so they will get 'redshifted' to longer wavelength radio. and again, since the chances are very low that such a photon will ever interact again, it will not appear as heat. heat means the uncoordinated form of energy that is obtained in a system that had plenty of time to interact and randomize and can therefore be described statistically. a 1GHz photon radiated into space will very probably never interact again, so in no sense will it become heat. its energy will literally disappear in the expanding universe.

well, not that any of this matters, but maybe someone will get a kick out of the fact that some fraction of the energy stored in pandoras battery will ultimately not end up in heat but contribute to the expansion of the universe