How much heat from a heat emitting source (e.g. CPU) is turned into radiation?

[DocSwag]

Do any of you physics people out there know the answer? I'm sort of doing a side project of figuring out a means of turning thermal energy into electricity. One of my friends brought up a good point, which is does a lot of the heat from a CPU, for example, actually get turned into radiation, or is a majority of it actually turned into conduction/convection? Just curious if any of you guys know the percentages or if only a small percentage of heat emitted (on average) is turned into radiation. 

[Djole123]

I don't think much heat is outputted through the cooler, and even less will be turned into electricity. I think so.

 

(unless you get an FX-9590, then you'll have some gr8 results)

[DocSwag]

1 minute ago, Djole123 said:

I don't think much heat is outputted through the cooler, and even less will be turned into electricity. I think so.

 

(unless you get an FX-9590, then you'll have some gr8 results)

No, that's not my question. My question is of the heat that is emitted, how much of it is emitted through radiation vs. conduction/convection? E.G. How much of the heat is actually turned into infrared light?

[DocSwag]

1 minute ago, CUDA_Cores said:

I don't have ananswer for you as I am not at all an expert in physics, but I would guess that a CPU only produces ever so slightly over background radiation if even that (so what we're looking at 0.05 more microsieverts)? Also which radiation are you talking about? Ionizing radiation or thermal radiation?

thermal radiation

[Sunshine1868]

you'd be better served using the temperature differential between ambient and the cpu to power some kind of heat-engine (i.e. stirling engine) or using a thermoelectric generator to generate electricity

[DocSwag]

1 minute ago, Sunshine1868 said:

you'd be better served using the temperature differential between ambient and the cpu to power some kind of heat-engine (i.e. stirling engine) or using a thermoelectric generator to generate electricity

Interesting..... Haven't heard of thermoelectric generators before....

I'll definitely be looking into that now 

[DocSwag]

1 minute ago, CUDA_Cores said:

we'll in that case doesn't all of it get converted into heat? I would imagine the amount of thermal radiation produced would be depending on how the particular chip was binned. Again I just really don't have any answers.

Yes, but only some of it is turned into infrared light. It can also be transferred to the air through conduction or just stored in the CPU itself.

[juretrn]

Quote

turning thermal energy into electricity

Peltier element maybe?

[Sunshine1868]

2 minutes ago, juretrn said:

Peltier element maybe?

its the reverse of a Peltier; a Seebeck element.

[ChrisCross]

the cooler doesnt emit any electromagnetic radiation at all. because to so you need electrons going threw ot which is not happening. so no infrared or something like that.

if your detecting that sort of radiation it comes from cpu, mobo or most likely some leds. 

[Nineshadow]

Most engines use heat to produce work, which can then be turned into electricity.

 

Anyway :

https://en.wikipedia.org/wiki/Thermoelectric_generator

[DocSwag]

1 minute ago, ChrisCross said:

the cooler doesnt emit any electromagnetic radiation at all. because to so you need electrons going threw ot which is not happening. so no infrared or something like that.

if your detecting that sort of radiation it comes from cpu, mobo or most likely some leds. 

It must release some infrared radiation, or you wouldn't be able to detect it with an infrared camera...

(please correct me if wrong)

[juretrn]

2 minutes ago, Sunshine1868 said:

its the reverse of a Peltier; a Seebeck element.

Doesn't a Peltier produce electricity when on a temperature differential?

[Sunshine1868]

Just now, juretrn said:

Doesn't a Peltier produce electricity when on a temperature differential?

yeah, but at that point its not the Peltier effect; it's the Seebeck effect ? 

[agent_x007]

Back to the main topic :
OP : U basicly want to know, how much energy is transfered from CPU into light (infrared to be exact), heat, and radio waves (the famous "spread spectrum") ?

[tomaatvk]

I'm one of the top people at my school when it comes to anything physics related and I'm starting uni (gonna study physics) in a few months so I think I'm allowed to speak.

 

The only radiation that will get formed gets formed by anything at the same temperature. By that I mean, if the CPU is 70°C it will emit the same ammount of radiation as, for example, a rock with the same dimensions at 70°C which is next to none compared to the several hundreds of °C required to make something emit sufficient radiation to be seen, let alone to gather a worthwhile ammount of energy from it.

 

If you are really persistant on getting energy from something hot like a CPU you should have a look at thermoelectric generators as @Sunshine1868 already said. The only thermoelectric generators I've worked with and know of ATM are peltier-elements which are small plates that create a voltage difference based on the temperature difference on both sides. However, they are bad heat conductors themselfs so you shouldn't stick them between a heatsink and a CPU or you will overheat it. They conduct heat through the current that flows through it.

 

They can also be used the other way around. In that case you put a voltage difference on it and it will pump heat away from one side to another. However, that by itself costs a lot of power so you will only create more heat for the heatsink to get rid off which is the biggest downside. Peltier-elements are very inefficient when it comes to transfering heat by voltage or when it comes to generating electricity through a temperature difference. I think @Stefan1024 used peltier-elements on his computer to generate electricity for a phone charger if I recall correctly.

 

I hope I helped with this over complicated explanation.

[Sunshine1868]

^watch out, we got a badass over here!

 

1 year out of university, physics and math  

[DocSwag]

1 minute ago, agent_x007 said:

Back to the main topic :
OP : U basicly want to know, how much energy is transfered from CPU into light/photons (infrared), heat, and radio waves (the famous "spread spectrum") ?

mainly how much is turned into infrared.

Though @Sunshine1868 has peaked my interest with Thermoelectric Generators....

[ChrisCross]

3 minutes ago, DocSwag said:

It must release some infrared radiation, or you wouldn't be able to detect it with an infrared camera...

(please correct me if wrong)

yeah but the amount of energy transmitted there is incredibly small. this is why thermal cameras are so expensive.

this makes sense when you take a look at the frequency range of electromagnetic waves.

also the problem with the thermal electronic generator is that you need a material with high electric conductivity and low thermal conductivity. normally those properties go along with each other. the only way to archive that is with rare earth or nanotechnology. both arent exactly cheap... 

[agent_x007]

13 minutes ago, DocSwag said:

mainly how much is turned into infrared.

Though @Sunshine1868 has peaked my interest with Thermoelectric Generators....

That is a REALLY SMALL ammount.
I don't think U need much energy to make photons "move"/"oscilate" at the infrared frequency.
 

Basicly : Glowing in light spectrum is easy to come by (ie. fire) 
Maintaining glow, or to be more exact - energy to glow, is A LOT harder when there is mass envolved (ie. atoms), and photons don't have mass (?).

That's only the theory, based on elementary physics/chemistry - I can be wrong here.

PS. Going BIG :
A nuke (ie. atomic bomb), at the moment of detonation, converts around 5% of it's initial energy into light (and that's what I readed in my school book at least).

[tomaatvk]

2 minutes ago, ChrisCross said:

yeah but the amount of energy transmitted there is incredibly small. this is why thermal cameras are so expensive.

this makes sense when you take a look at the frequency range of electromagnetic waves.

also the problem with the thermal electronic generator is that you need a material with high electric conductivity and low thermal conductivity. normally those properties go along with each other. the only way to archive that is with rare earth or nanotechnology. both arent exactly cheap... 

Well, you can get regular Peltier-elements (thermoelectric generators) from something like aliexpress for like €4 a piece. They're expencive in the western world because the demand is so low but in China they ship everywhere resulting in sufficient demand so it can be done cheap.

[Lord_Karango17]

A relatively small percentage of the chip's TDP will be radiated through thermal radation. This is why heatsink fins are employed so often with these coolers (mainly conduction, and convection).

 

[DocSwag]

Ok, so reading all of these responses, I feel that infrared radiation is a very small percentage of the heat emitted by a heat emitting source.

 

Therefore, I think I may change my direction to finding a way to do the same with conduction (probably via the Thermoelectric effect).

 

Thanks everyone for your responses!

[DebatED Nothing]

8 minutes ago, DocSwag said:

Ok, so reading all of these responses, I feel that infrared radiation is a very small percentage of the heat emitted by a heat emitting source.

 

Therefore, I think I may change my direction to finding a way to do the same with conduction (probably via the Thermoelectric effect).

 

Thanks everyone for your responses!

Generally speaking unless something's in a vacuum it will loose most of it's heat by convection and conduction. Radiation doesn't tend to come into effect until you're muuccchhhhhh hotter.

[Stefan1024]

9 hours ago, DocSwag said:

Do any of you physics people out there know the answer? I'm sort of doing a side project of figuring out a means of turning thermal energy into electricity. One of my friends brought up a good point, which is does a lot of the heat from a CPU, for example, actually get turned into radiation, or is a majority of it actually turned into conduction/convection? Just curious if any of you guys know the percentages or if only a small percentage of heat emitted (on average) is turned into radiation. 

As a rule of thumb, a black annodized heat sink cools about 10% better than one with a blank aluminum surface. But this compared to natural convection that is quite weak. But thermal radiation can make a difference.

 

I used a peltier element as a TEG, but the efficency was way below 1%. The main problem is not the heat source, but keeping the cold side cool. To produce energy you do not need a heat source but a temperature difference.

Even the bid powerstations have this problem. Just remember the hughe cooling tower that every heat based power stations have.