Jump to content
Search In
  • More options...
Find results that contain...
Find results in...

Does the "W/mK" matter on cooling low powered components, phones, or LED strips?

 Share

Go to solution Solved by AnonymousGuy,
1 hour ago, Newblesse Obblige said:

conductivity is the W/mK of the thermal pad?

temperature gradient is the difference between the temperature of the object to be cooled and the heatsink?

surface area is how big is the surface area I will be cooling?

thickness for conduction is thickness of the thermal pad?

 

So I could get away with a cheap thermal pad on the low power part? I see...

Yep on everything.  You can also easily test this with a cheap pad and a "good" 12W/mK pad.  You likely won't see any difference on the stable temperature of the thing.  If you're choosing pad thickness: as thin as you can go while still getting full contact (thermal paste is better than pads because the paste thins out to almost-nothing)

 

It's not so intuitive but for heat transfer to happen there has to be a temperature difference.  It's why if you place a bowl of water within a pot of boiling water, the water in the bowl won't boil...because the heat transfer goes to zero before the water in the bowl reaching boiling point (one of those trick questions... https://physics.stackexchange.com/questions/649146/cup-of-warm-water-suspended-in-a-pot-of-water-held-at-a-steady-boil)

 

For stuff like watercooling it's completely linear and why you'll hear "delta T" so much.  "a 240 rad can handle 300W at delta T [water temperature vs. air temperature] of 20 degrees" or something like this...which tells you it'll handle 450W if your system can tolerate a ambient + 30C water temperature.

 

 

(this is all like physics 101 level....the 401 level is crazy shit about Fourier equations and by that point in my college career I had no fucks left to give so I don't remember it: https://en.wikipedia.org/wiki/Heat_equation ... nowadays everyone just plugs this shit into CAD software to do the modeling and math for them)

I wanted to buy thermal pads for my LED strips, and for my stupid small project of cooling my old phone using a heatsink that is made for a Raspberry Pi or for NVMe SSDs.

 

I wanted get the cheap ones but my LEDs get so hot after a long time so maybe a better thermal pad might make a difference and I think the best thermal pads are overkill for a phone can go up to around 46°C(114.8°F) which isn't that hot compared to Graphics cards or MOSFETs yet it is the temperature that the phone is overheating.

 

What do you think?

 

Link to comment
Share on other sites

Link to post
Share on other sites

I had to think about this one:

 

The shittiest thermal pad is 100x the thermal conductivity of air.    I think that the thickness of air around a heatsink (convection makes this complicated where adding a fan changes the math) and the surface area limitation of a heatsink (number of fins) makes it so the heatsink itself is more of a limit than the thermal interface (pad).  Look at an extreme: if you had a giant heatsink with tons of surface area, and it was 100x the size of the part you're cooling...then you can start saying "hmmm maybe I need a better thermal interface" [because you can't get energy into the heatsink faster than it can dissipate it].

Workstation:  12900KF @ 5.2Ghz || MSI Pro-A Z690 DDR4|| EVGA FTW3 3090 1000W || G.Skill 3866 4x8GB || Corsair AX1500i || whole-house loop.

LANRig/GuestGamingBox: 9900nonK || Gigabyte Z390 Master || ASUS TUF 3090 650W shunt || Corsair SF600 || CPU+GPU watercooled 280 rad pull only || whole-house loop.

Server Router (Untangle): 8350K @ 4.7Ghz || ASRock Z370 ITX || 2x8GB 3200 || PicoPSU 250W, running on AX1200i || whole-house loop.

Server Compute/Storage: 10850K @ 5.1Ghz || Gigabyte Z490 Ultra || LSI 9280i-24 port || 4TB Samsung 860 Evo, 5x10TB Seagate Enterprise Raid 6, 4x8TB Seagate Archive Backup ||  whole-house loop.

Laptop: HP Elitebook 840 G8 (Intel 1185G7) + 3060 Thunderbolt Dock, Razer Blade Stealth 13" 2017 (Intel 8550U)

Link to comment
Share on other sites

Link to post
Share on other sites

24 minutes ago, AnonymousGuy said:

I had to think about this one:

 

The shittiest thermal pad is 100x the thermal conductivity of air.    I think that the thickness of air around a heatsink (convection makes this complicated where adding a fan changes the math) and the surface area limitation of a heatsink (number of fins) makes it so the heatsink itself is more of a limit than the thermal interface (pad).  Look at an extreme: if you had a giant heatsink with tons of surface area, and it was 100x the size of the part you're cooling...then you can start saying "hmmm maybe I need a better thermal interface" [because you can't get energy into the heatsink faster than it can dissipate it].

My comprehension aint working properly lmao. So in short, it doesnt matter on low heat emitting components and the size of the surface area of the thing you are cooling and the heatsink matters more?

Link to comment
Share on other sites

Link to post
Share on other sites

The equation is: transfer = conductivity * temperature gradient * surface area * thickness for conduction (it's some non-linear thing for convection) . 

 

How much heat your part is putting out is heat that you have to transfer elsewhere.  I think with low power parts it's very easy to make a heatsink that is capable of matching the heat output of the part with an acceptable delta T.   And again...it's linear.  A heatsink that is dissipating 5W at a 10C delta T between the heatsink and the part is capable of dissipating 10W at 20C delta T.

 

To try another way: I think the thermal interface is going to be transparent and not matter.  If the thermal pad is 1" square, you'd need 100 square inches of heatsink surface area to match it.  I don't think you're going to be slapping a heatsink that large on a low power part.  Think of it at an extreme: what if you had infinite thermal transfer into the heatsink...a perfect thermal pad: the heatsink is the limit.  The heatsink only stops being the limit when the rate of transfer from the part into the heatsink through the pad is less than the rate of transfer from the heatsink to the air.  

 

(with watercooling we care a lot about the thermal pad conductivity because copper and water are waaay better than air for conductivity, and with water flow being anything above a tricker there is no non-linear "boundary layer"  like you get with air, so the limit does quickly become how fast you can get the heat into the heatsink)

Workstation:  12900KF @ 5.2Ghz || MSI Pro-A Z690 DDR4|| EVGA FTW3 3090 1000W || G.Skill 3866 4x8GB || Corsair AX1500i || whole-house loop.

LANRig/GuestGamingBox: 9900nonK || Gigabyte Z390 Master || ASUS TUF 3090 650W shunt || Corsair SF600 || CPU+GPU watercooled 280 rad pull only || whole-house loop.

Server Router (Untangle): 8350K @ 4.7Ghz || ASRock Z370 ITX || 2x8GB 3200 || PicoPSU 250W, running on AX1200i || whole-house loop.

Server Compute/Storage: 10850K @ 5.1Ghz || Gigabyte Z490 Ultra || LSI 9280i-24 port || 4TB Samsung 860 Evo, 5x10TB Seagate Enterprise Raid 6, 4x8TB Seagate Archive Backup ||  whole-house loop.

Laptop: HP Elitebook 840 G8 (Intel 1185G7) + 3060 Thunderbolt Dock, Razer Blade Stealth 13" 2017 (Intel 8550U)

Link to comment
Share on other sites

Link to post
Share on other sites

2 hours ago, AnonymousGuy said:

The equation is: transfer = conductivity * temperature gradient * surface area * thickness for conduction (it's some non-linear thing for convection) . 

conductivity is the W/mK of the thermal pad?

temperature gradient is the difference between the temperature of the object to be cooled and the heatsink?

surface area is how big is the surface area I will be cooling?

thickness for conduction is thickness of the thermal pad?

 

2 hours ago, AnonymousGuy said:

I think the thermal interface is going to be transparent and not matter.

So I could get away with a cheap thermal pad on the low power part? I see...

Link to comment
Share on other sites

Link to post
Share on other sites

1 hour ago, Newblesse Obblige said:

conductivity is the W/mK of the thermal pad?

temperature gradient is the difference between the temperature of the object to be cooled and the heatsink?

surface area is how big is the surface area I will be cooling?

thickness for conduction is thickness of the thermal pad?

 

So I could get away with a cheap thermal pad on the low power part? I see...

Yep on everything.  You can also easily test this with a cheap pad and a "good" 12W/mK pad.  You likely won't see any difference on the stable temperature of the thing.  If you're choosing pad thickness: as thin as you can go while still getting full contact (thermal paste is better than pads because the paste thins out to almost-nothing)

 

It's not so intuitive but for heat transfer to happen there has to be a temperature difference.  It's why if you place a bowl of water within a pot of boiling water, the water in the bowl won't boil...because the heat transfer goes to zero before the water in the bowl reaching boiling point (one of those trick questions... https://physics.stackexchange.com/questions/649146/cup-of-warm-water-suspended-in-a-pot-of-water-held-at-a-steady-boil)

 

For stuff like watercooling it's completely linear and why you'll hear "delta T" so much.  "a 240 rad can handle 300W at delta T [water temperature vs. air temperature] of 20 degrees" or something like this...which tells you it'll handle 450W if your system can tolerate a ambient + 30C water temperature.

 

 

(this is all like physics 101 level....the 401 level is crazy shit about Fourier equations and by that point in my college career I had no fucks left to give so I don't remember it: https://en.wikipedia.org/wiki/Heat_equation ... nowadays everyone just plugs this shit into CAD software to do the modeling and math for them)

Workstation:  12900KF @ 5.2Ghz || MSI Pro-A Z690 DDR4|| EVGA FTW3 3090 1000W || G.Skill 3866 4x8GB || Corsair AX1500i || whole-house loop.

LANRig/GuestGamingBox: 9900nonK || Gigabyte Z390 Master || ASUS TUF 3090 650W shunt || Corsair SF600 || CPU+GPU watercooled 280 rad pull only || whole-house loop.

Server Router (Untangle): 8350K @ 4.7Ghz || ASRock Z370 ITX || 2x8GB 3200 || PicoPSU 250W, running on AX1200i || whole-house loop.

Server Compute/Storage: 10850K @ 5.1Ghz || Gigabyte Z490 Ultra || LSI 9280i-24 port || 4TB Samsung 860 Evo, 5x10TB Seagate Enterprise Raid 6, 4x8TB Seagate Archive Backup ||  whole-house loop.

Laptop: HP Elitebook 840 G8 (Intel 1185G7) + 3060 Thunderbolt Dock, Razer Blade Stealth 13" 2017 (Intel 8550U)

Link to comment
Share on other sites

Link to post
Share on other sites

2 hours ago, AnonymousGuy said:

Yep on everything.  You can also easily test this with a cheap pad and a "good" 12W/mK pad.  You likely won't see any difference on the stable temperature of the thing.  If you're choosing pad thickness: as thin as you can go while still getting full contact (thermal paste is better than pads because the paste thins out to almost-nothing)

 

It's not so intuitive but for heat transfer to happen there has to be a temperature difference.  It's why if you place a bowl of water within a pot of boiling water, the water in the bowl won't boil...because the heat transfer goes to zero before the water in the bowl reaching boiling point (one of those trick questions... https://physics.stackexchange.com/questions/649146/cup-of-warm-water-suspended-in-a-pot-of-water-held-at-a-steady-boil)

 

For stuff like watercooling it's completely linear and why you'll hear "delta T" so much.  "a 240 rad can handle 300W at delta T [water temperature vs. air temperature] of 20 degrees" or something like this...which tells you it'll handle 450W if your system can tolerate a ambient + 30C water temperature.

 

 

(this is all like physics 101 level....the 401 level is crazy shit about Fourier equations and by that point in my college career I had no fucks left to give so I don't remember it: https://en.wikipedia.org/wiki/Heat_equation ... nowadays everyone just plugs this shit into CAD software to do the modeling and math for them)

Cool! Although I didn't understand most of the physics part, I will keep it in mind. Thanks!

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share


×